View Full Version : Minimal Trade Mod Idea
"A minimalist mod with little or no trade.... now that might be very interesting indeed and may even make the game harder in itself since trade is the main way we humans can whoop the AI.
Hi all.
Been busy with Civ3 and PTW... Got bored of MTW after i finished my mod. lol. When it gets to the stage of fighting 1 hour long 4000 v 4000 man battles multiple times every turn i get bored of MTW. Especially when you're dealing with hundreds of horse archers, man those guys are annoying to fight.
So, the talk of a minimalistic mod might keep the battles down to a more interesting and less repetitive level.
Any ideas? Remove trade from the sea, or even altogether? Then ships would be useful solely for naval assaults and wouldn't necessarily be so important
Might make it very interesting and the game progression more frought with danger....however, one may have to seriously chop building prices and time to produce them though? Anyway, if any of you are interested, start up another thread and we'll get the ball rolling. Otherwise, see ya in a month or two when i may be interested in another late game hour-long turn lol.
The only trouble i envisage with this is trying to balance the factions when trade is no longer an issue...particularly since the Italians, Sicilians, English and Danes rely so heavily on it.
Doc"
Welcome back DOC, I am taking you up on your suggestion to open a new thread about this idea.
I've thought about this some and I think that it might solve some of the AI's problems, but it might create others, so I'm currently undecided. Anyway, I thought I'd post some suggestions and see what other people thought about them.
First, the Main AI problems which keep it from effectively competing with a serious player:
1. The AI can't seemingly be made to compete effectively at building a trade empire with the player, thus it never develops a competitive income. You can tweak the shipbuilding and merchant preferences as well as lower the build cost/time etc., (we've tried that one) but no matter how many ships the AI builds, it just doesn't keep them in the right places. You need a chain of ships from your home port to whatever province you're trying to trade with and the AI won't reliably keep a ship in it's home port, let alone maintain the chain. Kraellin may have had better success with his modifications on this, I don't know.
2. the AI doesn't safeguard it's most developed provinces sufficiently, thus they get captured and recaptured with distressing regularity. Everytime this happens, the buildings degrade. After 150 years or so you look at the map and every province has a castle or keep (if that) in it - few citadels and no fortresses. I call this problem PROGRESSIVE DESERTIFICATION and it might be the most serious AI shortcoming. (CA ought to have instituted this feature of castle degredation for the player - if he loses a province or captures one by waiting out a seige - but NOT the AI).
SUGGESTED SOLUTIONS AND PROBLEMS:
A. If we can't make the AI trade successfully we can reduce the impact of trade by increasing the merchant build requirements (to castle9 or something). Trade would still exist, but it would be much harder to build a trade empire. The AI would still not do it so it wouldn't be disadvantaged, but the player couldn't gain the usual money advantage by building a master merchant in every seabourne province. An even more radical solution might be to reduce or eliminate trade altogether (just eliminate merchant buildings) and compensate by upping the farm income from all provinces accordingly. All AI build preference for Trader, Port and Shipbuilder would be returned to CA standard and provinces would have CA standard resources.
I like playing as a trade builder myself to simple slash and conquer and wish the AI could be made to compete, but I might have to give up on that one. It would reduce the fun of the game for me though - one of the reasons I've hesitated.
Problem: This would cripple the Italians and make them a virtually non-playable faction. Various solutions suggest themselves.
1. Make the Italians non-major faction and split their provinces up into independent states. This is historical (they only united in the 19th century anyway) and Paladin has already produced a mod that does this. (I'm not sure what this would do to the play balance with Sicily.) Give the various provinces high (3 or 4) rebellion factors so that they wouldn't simply be swallowed up with ease.
Sicily isn't a CA major faction, but I and numerous others have modded it to be. With trade per CA, Sicily is a fun faction to play, and a player can easily win the game by building a trade empire from Sicily. Don't know what would happen to Sicily (as an AI faction)if trade were reduced or removed.(If Sicily started out with a Merchant guild and some ships, they might be able to trade (but not replace their buildings if they were destroyed). Don't know how that would work.
There might be problems with Byzantium also that would require thought.
B. REDUCE the Build Times and building Costs: I took your suggestion to reduce the build times by 1/3 and it seems to help the overall AI balance. The game is quite balanced now between the various AI factions, but the player still has a decided advantage, of course. If there were fewer florins available to the player, then he wouldn't gain as much of an advantage from lowered build times. Reducing the build times to 1/2 might be even better for the AI. Has anyone tried this? What were the results?
Reducing support costs: is another idea that should be explored. What effect would this have on the overall play balance? The Danes might get an advantage, since they are often near their support limit (depite my giving them Norway and doubling their farm income, plus starting them out with 16,000 florins - along with all the other factions).
Has anyone tried just giving the AI factions 100,000 florins to start to see what happens? This would get the AI factions going at least, but they might tail off after a few decades (i.e. they might squander the money). It might lead to better AI play and it might not.
Well, that's enough of a list to start. If anyone has additional ideas, I'd like to hear them.
Ah, Cugel, thanks for posting this.
Sadly, I'm not sure how easy or even worthwhile creating a reduced or no-trade mod might be? You see, the trouble is one of the driving forces behind the interest in the game for the human player lies with the striving to create trade routes as it is the only other thing to do apart from somewhat repetitive and time-consuming tactical battles.
Anyway lets look at this objectively and see if there's anything at all that could be done to ease the woes of the AI, whilst keeping the interest there for the human.
Firstly, the problems:
1. The AI fails to utilise trade routes effectively, while the human understands and can utilise them to their full potential.
2. This might be a personal gripe, but maybe you'll agree with it...? At times I find the late game turns where huge armies battle huge armies in a repetitive and time-consuming matter incredibly tedious. One turn can take anywhere up to 2 hours simply to plough through these multitudes of repetitive battles.
3. Progressive desertification.
Now how to tackle these issues, if it is indeed possible:
1. Unfortunately one cannot customise the AI strategy for utilising its ships, only make the AI produce more in an attempt to prevent the human from gaining uncontested control of naval trade. Success from mods increasing the AI build preferences for trade improvements and ship building i would say have had mixed success.
Ships cost a lot of money to make and support and if the AI only builds ships and leaves 30+ in is home port with only a few reaching outwards to establish trade routes them i'm not sure an improvement has been made. The English, Italians, Sicilians are all very prone to this home port stockpiling. Conversely there seems to be the other phenomenon of the AI failing to utilise its home port effectively, often having ships scattered around the Med without a ship left in the lucrative home province. This is a prime example of the Italians (Venice) and the Byzantines (Constantinople).
Whilst the first problem of stockpiling is completely unmoddable at the moment, it is possible to somewhat alleviate the second problem. I found that if i swapped the province characteristics of Genoa and Venice around this made a massive difference the Italian AI since they appear to derive trade income much better from Genoa (simply by leaving a ship there more often). Additionally the same can be said for the Byzantines with Greece and Constantinople, however, this change is more difficult to conceive as Constantinople is such an important territory in the game in terms of structural improvements and also historically.
2. So that's one option. The other is to remove trade or limit it sufficiently so as it doesn't impact as greatly on the game. How?
a) Well one could cut down the number of tradable items per province, especially the sea provinces.
b) One could delay the arrival of the merchant improvements to later in the tech tree.
c) OR simply remove trade altogether, which may be a bit drastic.
Problems with this are obvious. Not only will the factions in their current setup be imbalanced, the likelihood of the AI going bankrupt grows. But then it also does for the human i guess. There'd also be issues with the now severed borders at Wessex and Morocco. Ships would now no longer be vital for trade, however, we'd still want ot encourage them for sea based attacks and traversing the severed borders. How do you get the AI to get the right balance? Too many ships would be stupid and lead to bankruptcy, too little would also be stupid and would hinder AI travel.
Advantages i envisage would be the game playing at a slower pace and the battles between factions involving much smaller numbers of troops (massive plus IMHO), thereby reducing the repetitive huge battles tedium. One might even see the human losing important territories on a more frequent basis, who knows? One way to combat the AI banruptcy syndrome would be to give it a huge starting income and given that support costs might equal income very quickly without trade involvement, this income might be spent of more useful improvements as opposed to multitudes of troops? Again this is pure speculation.
For me fighting 500 v 500 5 times a turn is far more appealing than 4000 v 4000 5 times a turn. I just don't have the time to complete a full campaign game of MTW when the huge battles start occurring regularly. Furthermore, with more constraints on income as a result of being without a substantial trade income, dominating the globe will be much more problematic for all factions involved. And the mercenary rush won't be nearly as viable as it is now. Mercenaries, whilst being a great idea, never seemed to be properly utilised by the AI like they were by the human. When have you ever seen the AI hire massive merc armies to take a province and then disband them at a later date to save costs?
So in summary, by reducing trade:
1. We'd have to rebalance the factions.
2. We'd get smaller battles (in theory).
3. We'd need to address AI build priorities for ships and shipping inprovements like ports, merchants and shipyards.
4. Consider the effectiveness of keeping severed borders.
5. We'd bring the AI ability closer to the human by making it more difficult to expand?
6. We'd probably need to boost AI starting money (considerably?) to prevent initial bankruptcy.
3. Lowering the amount of money in the game by reducing or eliminating trade would almost certainly mean that building costs and times would need to be reduced. This would possibly also help alleviate the progressive desertification that's all too evident as the AI regularly swaps its most valuable provinces. Having a lower number of troops available will also hinder the human in this respect as they will be less able to stockpile large stacks in valuable provinces and expand at the same time.
Once again this points towards giving the AI a large starting income and hoping that it won't lean towards building dozens of troops with it but rather building improvements. I hope this is what will happen if the AI is close to support limit (i.e. income in = expenses out) even though it has a huge bank balance. I don't know?
So in summary:
1. Smaller income would be better linked with a larger starting balance compared to smaller balance with a potential vast income as it is now
2. Reduction in building times and costs to compensate for smaller troop numbers, less money and progressive desertification.
Right then, that's something to chew on, let me know your thoughts.
Cheers
Doc
PanthaPower
01-10-2003, 15:55
I'm not too much into this kind of editing but it all sounds very good Doc. But it would require an experienced modder in this area to make a campaign whcih uses all your ideas...
"Whilst the first problem of stockpiling is completely unmoddable at the moment, it is possible to somewhat alleviate the second problem. I found that if i swapped the province characteristics of Genoa and Venice around this made a massive difference the Italian AI since they appear to derive trade income much better from Genoa (simply by leaving a ship there more often)."
Well, DOC, that's why I like reading your posts, you always seem to have fascinating ideas I never would have thought of http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif Swap Venice's characteristics with Genoa Hmmmn I think that Kraellin also said that he had gotten the Byzantines building a trade empire, so maybe he'll let us see what he did to achieve that sometime.
It seems we are generally in agreement regarding the AI shortcomings and possible solutions. This narrows down the field of possible approaches and provides a good beginning.
I SHOULD SAY JUST IN CASE ANYONE FROM CA EVER SEES THIS THAT I DON'T LIKE GETTING RID OF TRADE I LIKE TRADE I WANT MORE AND BETTER TRADE FOR TW:3 I JUST CAN'T SEE ANY OTHER WAY TO BALANCE THE AI AT PRESENT. http://www.totalwar.org/forum/non-cgi/emoticons/angry.gif
PROBLEM #2 THE HUGE BATTLES: I should comment, however, that I don't really mind the large 4000 v. 4000 battles, sometimes they can be fun (I just don't want to have to fight 3-4 such battles per turn - when that happens I just hit the "calculate battle outcome" button to get through the turn). However, it's boring to have to fight 4,000 urban militia and light cavalry. http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif
I think that this problem is quite solveable however. The reason there are so many troops in such battles is that the AI builds hordes of peasants, urban militia, archers and light cavalry, etc. Historically, this is accurate. According to Contamine's Warfare in the Middle Ages a typical period army might contain up to 3/4 archers and support troops (knights were terribly expensive and even the most powerful kingdoms couldn't afford more than a limited number of them, nor could they afford to keep them in the field for an extended period as a general rule). However, it's much more FUN to fight battles with just the more advanced troops (i.e. jannisary infantry v. gothic knights). The AI build preferences can be tweaked to favor more expensive troop types (thus there will be far fewer of them). To do this successfully, however, the AI needs more florins, not less. This brings us back to the fundamental problem. THE AI FACTIONS DON'T GENERATE ENOUGH INCOME.
PROBLEM #1: Effects of Limiting Trade: If we limit trade for the player by any of the means we've discussed, then the AI must be given more income in other ways. One solution would be to double or even treble the farm income for each province and reduce the farm improvement build times, while increasing the AI build preferences for farm improvements. If you start them off doubled or trebled, by the time all improvements have been built, the AI should be generating farm income like the Hojo in Shogun An added benefit is that these farming improvements aren't degraded when the province is conquered, so no desertification problem. The AI is quite adept at conquering provinces, so it could amass income quite well this way.
PROBLEM: This might unbalance the game again. Obviously, Italy and Sicily would suffer. England would require attention and the Danes might need looking after. The HRE would benefit, the Byzantines suffer and the Egyptians might need to be cut back to keep them from overruning the map
One solution I've found is to increase the rebellion factors to 2, thus making rebellions more likely. This dramatically increases faction longevity, since they stage loyalist comebacks, plus it slows down the creation of the "HOJO HORDE" phenomena (1 AI superpower at the end of 100 years). I think that I have already gotten a good play balance based on the modifications you sent me some time ago. It might require some re-tweaking, but as it's Very Do-able.
Now shipping: Since we don't want to remove the naval element of the game (like overseas invasions by the AI), whatever we do to trade, we should reduce the support costs for ships (perhaps down to zero) - especially if trade were removed altogether (which I agree is a bit too drastic). AI factions could still be encouraged to produce ships, but it would no longer make them go broke. This wouldn't be historical (maintaining ships costs big money) but we could assume that the support costs were included in the initial build costs and it would improve playbalance. The AI could then do whatever it wanted with it's ships - the main use for them would be to prevent seabourne invasions, especially by other AI factions (which the AI sometimes manages o.k. - in my last campaign the AI Egyptians invaded Scotland and Ireland in 1 turn by sea from Genoa). The player would also have reduced costs, but more ships would no longer automatically mean more trade income, so the incentive to build large numbers of them would be less.
Keeping the number of AI ships relatively high would also make it possible to avoid problems with the severed landbridges at Wessex and the Straights of Gibraltar - hopefully the English and Spanish would be able to cope as they seem to be able to do now (at least in my campaigns).
(Sorry for the length of this post - I have lots of ideas on this theme). http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif
Quote[/b] (+DOC+ @ Jan. 10 2003,08:44)]Ah, Cugel, thanks for posting this.
Sadly, I'm not sure how easy or even worthwhile creating a reduced or no-trade mod might be? You see, the trouble is one of the driving forces behind the interest in the game for the human player lies with the striving to create trade routes as it is the only other thing to do apart from somewhat repetitive and time-consuming tactical battles.
Anyway lets look at this objectively and see if there's anything at all that could be done to ease the woes of the AI, whilst keeping the interest there for the human.
Firstly, the problems:
1. The AI fails to utilise trade routes effectively, while the human understands and can utilise them to their full potential.
2. This might be a personal gripe, but maybe you'll agree with it...? At times I find the late game turns where huge armies battle huge armies in a repetitive and time-consuming matter incredibly tedious. One turn can take anywhere up to 2 hours simply to plough through these multitudes of repetitive battles.
3. Progressive desertification.
Now how to tackle these issues, if it is indeed possible:
1. Unfortunately one cannot customise the AI strategy for utilising its ships, only make the AI produce more in an attempt to prevent the human from gaining uncontested control of naval trade. Success from mods increasing the AI build preferences for trade improvements and ship building i would say have had mixed success.
Ships cost a lot of money to make and support and if the AI only builds ships and leaves 30+ in is home port with only a few reaching outwards to establish trade routes them i'm not sure an improvement has been made. The English, Italians, Sicilians are all very prone to this home port stockpiling. Conversely there seems to be the other phenomenon of the AI failing to utilise its home port effectively, often having ships scattered around the Med without a ship left in the lucrative home province. This is a prime example of the Italians (Venice) and the Byzantines (Constantinople).
Whilst the first problem of stockpiling is completely unmoddable at the moment, it is possible to somewhat alleviate the second problem. I found that if i swapped the province characteristics of Genoa and Venice around this made a massive difference the Italian AI since they appear to derive trade income much better from Genoa (simply by leaving a ship there more often). Additionally the same can be said for the Byzantines with Greece and Constantinople, however, this change is more difficult to conceive as Constantinople is such an important territory in the game in terms of structural improvements and also historically.
2. So that's one option. The other is to remove trade or limit it sufficiently so as it doesn't impact as greatly on the game. How?
a) Well one could cut down the number of tradable items per province, especially the sea provinces.
b) One could delay the arrival of the merchant improvements to later in the tech tree.
c) OR simply remove trade altogether, which may be a bit drastic.
Problems with this are obvious. Not only will the factions in their current setup be imbalanced, the likelihood of the AI going bankrupt grows. But then it also does for the human i guess. There'd also be issues with the now severed borders at Wessex and Morocco. Ships would now no longer be vital for trade, however, we'd still want ot encourage them for sea based attacks and traversing the severed borders. How do you get the AI to get the right balance? Too many ships would be stupid and lead to bankruptcy, too little would also be stupid and would hinder AI travel.
Advantages i envisage would be the game playing at a slower pace and the battles between factions involving much smaller numbers of troops (massive plus IMHO), thereby reducing the repetitive huge battles tedium. One might even see the human losing important territories on a more frequent basis, who knows? One way to combat the AI banruptcy syndrome would be to give it a huge starting income and given that support costs might equal income very quickly without trade involvement, this income might be spent of more useful improvements as opposed to multitudes of troops? Again this is pure speculation.
For me fighting 500 v 500 5 times a turn is far more appealing than 4000 v 4000 5 times a turn. I just don't have the time to complete a full campaign game of MTW when the huge battles start occurring regularly. Furthermore, with more constraints on income as a result of being without a substantial trade income, dominating the globe will be much more problematic for all factions involved. And the mercenary rush won't be nearly as viable as it is now. Mercenaries, whilst being a great idea, never seemed to be properly utilised by the AI like they were by the human. When have you ever seen the AI hire massive merc armies to take a province and then disband them at a later date to save costs?
So in summary, by reducing trade:
1. We'd have to rebalance the factions.
2. We'd get smaller battles (in theory).
3. We'd need to address AI build priorities for ships and shipping inprovements like ports, merchants and shipyards.
4. Consider the effectiveness of keeping severed borders.
5. We'd bring the AI ability closer to the human by making it more difficult to expand?
6. We'd probably need to boost AI starting money (considerably?) to prevent initial bankruptcy.
3. Lowering the amount of money in the game by reducing or eliminating trade would almost certainly mean that building costs and times would need to be reduced. This would possibly also help alleviate the progressive desertification that's all too evident as the AI regularly swaps its most valuable provinces. Having a lower number of troops available will also hinder the human in this respect as they will be less able to stockpile large stacks in valuable provinces and expand at the same time.
Once again this points towards giving the AI a large starting income and hoping that it won't lean towards building dozens of troops with it but rather building improvements. I hope this is what will happen if the AI is close to support limit (i.e. income in = expenses out) even though it has a huge bank balance. I don't know?
So in summary:
1. Smaller income would be better linked with a larger starting balance compared to smaller balance with a potential vast income as it is now
2. Reduction in building times and costs to compensate for smaller troop numbers, less money and progressive desertification.
Right then, that's something to chew on, let me know your thoughts.
Cheers
Doc
We are all agreed on the fact that the naval AI is pretty hopeless, and the chances for another patch are slim. To make matters worse, the trade element of the game is flawed. There is a limit of 5 trade items for provinces, some provinces start with no trade items (Rome for example has none). All trade items produce the same income. Silk and Gems produce the same income as wood. This is absurb.
My approach was to eliminate the naval trade variable and use income producing structures instead. I took the default trade resources and factored them into the base income of each province. I increased the number of income producing structures. Ports, shipyards, inns, cathedrals (pilgrims), etc, produce fixed income. Naval trade is factored in abstractly with ships as combat/transport units.
i'm going to try and create a mod to boost the ai. first limit every provence to one trade item, and one resource. if they dont have one give them one.. that should limit trading for the human player.
secondly gonna increase all provences income by a factor of ten, and then for the units split them in two and create unit for the human player, and the ai.. the ai leave the costs and support costs alone, and the human boost them by a factor of ten for cost and support. that away the ai will benifet from the extra income, more than the human player.
the main problem with that is i'll have to limit the factions the human can play to 2 or three, don't want to, but i can't think of another way to do it.. my main plan is to make a mod where it is virtually impossible for the ai to run out of money and start cranking out crappy units by the gross.. gonna set all provences rebels to 2 such a cugel suggested, but only if it is a faction controled land. all rebel provences, i'll set to 4, try and slow down the ai in the begining.. been testing it for a day now, and it still needs major tweaking, may just have to limit it to 6 or 7 total factions available.. after 100 years, everyone had such huge armies, that it was slowing my system down..
figuring on setting three human controlable factions with 3 provences each, the giving the ai 6 or 7 each. and seeing how that will help make it harder to win..
Good Turbo & Jayrock, I was beginning to think this thread was dead & of no interest to anyone. http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
We are all agreed on the fact that the naval AI is pretty hopeless, and the chances for another patch are slim. To make matters worse, the trade element of the game is flawed. There is a limit of 5 trade items for provinces, some provinces start with no trade items (Rome for example has none). All trade items produce the same income. Silk and Gems produce the same income as wood. This is absurb.
My approach was to eliminate the naval trade variable and use income producing structures instead. I took the default trade resources and factored them into the base income of each province. I increased the number of income producing structures. Ports, shipyards, inns, cathedrals (pilgrims), etc, produce fixed income. Naval trade is factored in abstractly with ships as combat/transport units.
This sounds like you have already produced a working mod. If so, have you uploaded it to the Org? What were the results from playtesting? How exactly did you get the structures to produce extra income (frankly, I don't know how to do this). Abstracting trade in some fashion might be a very good idea if it could be accomplished, it's the "chain of ships" requirement for trade that poses problems for the AI. CA simply installed this feature without programming the AI to follow it - thus, the P*** poor AI trade performance.
One might easily compensate for the lack of AI income by giving the AI, say 100,000 florins to start (but not the player obviously) or even more $ for the AI factions so that they don't run short -- perhaps something radical like 250,000 florins. I was thinking of playtesting this idea when I finish my current campaign to see how the AI did. However, the AI also doesn't seem make the best use of its income to purchase high end units and dissolve old armies of peasants and spearmen when they become available (thus lowering its maintenance costs, enabling it to purchase more high end units). This exacerbates the AI income problems, because the AI is hoarding large armies of peasants and spearmen. Having more $ might only cause it to build even more useless peasants, spear and a few royal knights (this seems to be the standard AI army choice). Because only 16 units can appear at one time, having 6,000 AI spearmen & peasants facing 1,000 Varangian Guardsmen and Kataphratoi is hopeless for the AI. (This is actually fairly accurate historically -- medieval armies weren't all "high tech", but it unbalances play in favor of the player, especially in the high and late periods).
Proposed Solution: Tweaking the AI build preferences to produce more "high end units" when they become available. The problem is "how much is good for the AI". I have not attacked this problem because of the daunting amount of time it would need for playtesting. I wonder if anyone has done anything along this line?
Jayrock
i'm going to try and create a mod to boost the ai. first limit every provence to one trade item, and one resource. if they dont have one give them one.. that should limit trading for the human player.
secondly gonna increase all provences income by a factor of ten, and then for the units split them in two and create unit for the human player, and the ai.. the ai leave the costs and support costs alone, and the human boost them by a factor of ten for cost and support. that away the ai will benifet from the extra income, more than the human player.
It would be easier to simply reduce the support costs for various units or increase the AI income. The problem seems to be that the AI builds too many low tech units and then can't afford the high tech ones because of the high maintenance costs. It may even consider that it doesn't need them. (This was a problem that carried over from Shogun). Of course, this would make it possible for the player to build more units, but he won't have the income to just build anything he wants, while the AI will.
my main plan is to make a mod where it is virtually impossible for the ai to run out of money and start cranking out crappy units by the gross.. gonna set all provences rebels to 2 such a cugel suggested, but only if it is a faction controled land. all rebel provences, i'll set to 4, try and slow down the ai in the begining.. been testing it for a day now, and it still needs major tweaking, may just have to limit it to 6 or 7 total factions available.. after 100 years, everyone had such huge armies, that it was slowing my system down..
My feeling is that you don't have to modify the rebellion factors for provinces with a 3 or 4 already (such as Lithuania), they're rebellious enough. I've achieved excellent play balance simply by raising all the provinces to 2 and leaving the 3 and 4 rebellious provinces alone. This definitely slows down the AI from developing 1 superfaction. If an ai faction expands too fast, it gets quickly cut down to size, plus there are loyalist rebellions that re-establish factions all the time - Few of them seem to die out permanantly in the 1st 100 years or so.
If you achieve success in modding the AI build preferences for high tech units and reducing the AI reliance upon peasants, please let us know I'd like to see the results of that. There's nothing more boring that playing battles with 3-4 waves of peasants & militia sargeants attacking you only to run away again. How do they expect to defeat my Varangian Guards or Janissary Infantry with such rabble? http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
cugel, you were right about the rebelions, in the first 75 years, i had all factions except mine get eliminated, they all reappeared, but the thing i noticed was that most of the eliminations were done by rebelion armies.. so the ai is killing itself.. the ai still cranks out peasants even with the money it's making, i know it has a lot of money, because their is a whole lot of bribing going on. even if they only have half the money i do, i've got over 30 mil florins in the bank, no i haven't reset the players unit yet. trying to get the ai to play nice with each other first.
using th e matteosartori cheat code to track everyone. gonna eliminate the ai's ability to build peasants, and deactivate rabble, and take them out of the possible rebelious cues, i reset the spearmen to the same building level of the peasant. only need a castle to build, and that works pretty good.
next project is gonna reset their building priorities, and set all the lower unit to a lower number, and boost the good units by a couple of hundred, and see if that works any better. unfortunately it's gonna half to wait a few weeks. work dropped a 75 million dollar bid on my desk today, and i have to get it ready by next friday. love those time crunches.. oh well.
"next project is gonna reset their building priorities, and set all the lower unit to a lower number, and boost the good units by a couple of hundred, and see if that works any better."
This is a very good idea. If spearmen replaced peasants it would greatly increase the utility of AI armies. Make them cost the same as peasants did previously and have the same AI build preference as peasants (we know that the AI will then build tons of them). http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif
Here's another idea borrowed from King David's Patrician mod: "Higher Rank Honor and Dread for Most provinces for better AI Play. This is realy good in battle as you will see your AI oponent put its General further from harm and will attempt to outflank you."
I might try that next.
Gregoshi
01-28-2003, 21:01
Ranges in the Entrance Hall has the following to contribute to this discussion:
-----
Quote[/b] ]
2. So that's one option. The other is to remove trade or limit it sufficiently so as it doesn't impact as greatly on the game. How?
a) Well one could cut down the number of tradable items per province, especially the sea provinces.
b) One could delay the arrival of the merchant improvements to later in the tech tree.
c) OR simply remove trade altogether, which may be a bit drastic.
I'm not a modder myself, but if you can alter the value of individual trade goods, you could make certain provinces have extreme value trade goods, while making most other goods very low value. This way the trade dependent factions can have one very good trading province, while all the other provinces give a much more modest income.
Still, i have no idea if this is possible, but it sounded like a worthwhile suggestion to me.
-----
Quote[/b] ]I'm not a modder myself, but if you can alter the value of individual trade goods, you could make certain provinces have extreme value trade goods, while making most other goods very low value. This way the trade dependent factions can have one very good trading province, while all the other provinces give a much more modest income.
it's a great idea, nfortunately all the trade data seems to be hardcoded, whereas i can play around with the txt files fairly easily, i don't know enough about the exe to play with it.
the main problem is that no matter what the good is, you get the same amount of florins for each item, wood gives the same amont as gems, etc. i know it's stupid but that's what i've seen when i pay attention to it. also provences without a port seem to suffer with trade, they can't trade with as many different provinces as a port one so they suffer a penalty of sorts, still tweaking the mod i'm working on made a little progress, but not much. the ai is still either to weak or it's way to strong. wish i knew what the values in all the colums translated to ingame, so it's trial and error right now.
"wish i knew what the values in all the colums translated to ingame, so it's trial and error right now."
You should check out some of the earlier discussion threads with Kraellin. He seems to have mastered how some of the values interrelate. Generally speaking, the AI can change to many different "personalities", even ones from a different religion like muslim_orthodox for a catholic faction. You will see this from the fact that in the crusader_unit file there are entries for, say, Muslim_Peaceful for feudal knights. If only Muslim factions could have the AI personality "muslim_peaceful" there would be no need to have a muslim_peaceful AI build probability setting for feudal knights, since no Muslim faction can build them. As K. pointed out, these are merely string names. Apparantly, the only AI behavior code that is separate is rebel (K. tested the intial behavior sets to see if they would crash the game and only that one did).
The general presumption has been that higher values increase the AI build preference, but check the values for "peasant" in the crusader_unit file (notice the low values): "POVERTY_STRICKEN(31.5), DESPERATE_DEFENCE(1), CATHOLIC_EXPANSIONIST(8.4), CATHOLIC_NAVAL_EXPANSIONIST(8.4), CATHOLIC_TRADER(10.5), CATHOLIC_CRUSADER_TRADER(10.5), CATHOLIC_EXPANSIONIST_CRUSADER(8.4), CATHOLIC_DEFENSIVE_CRUSADER(8.4), POPE(5), CATHOLIC_DEFENSIVE(8.4), CATHOLIC_ISOLATIONIST(5), ORTHODOX_DEFENSIVE(8.4), ORTHODOX_EXPANSIONIST(8.4), ORTHODOX_STAGNANT(10.5), MUSLIM_PEACEFUL(10.5), MUSLIM_EXPANSIONIST(8.4), MUSLIM_DEVOUT(8.4), BARBARIAN_RAIDER(4.2), REBELS(21), CLOSE_TO_SUPPORT_LIMIT(1.05)"
And we all know how many peasant units the AI builds Clearly, the code does some calculation with these numbers we don't understand (at least I don't).
(come to think of it, I may have modded these values, but the point remains). http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif
Cugel, i would argue against some of the points you mentioned there.
Firstly i believe that a faction's personality can only change from its initial starting one to one of the "special circumstances" personalities, i.e.:
Poverty_stricken
Close_to_support_limit
Desperate_defence
I do not believe that a catholic_crusader_trader personality ever changes to another catholic personality or for that matter another religion's personality. It wouldn't make sense and defeats the purpose of each factions' unique strengths and weaknesses. Why would the Italians or English ever relinquish their trader personalities to become musilm peaceful, orthodox stagnant or catholic isolationist, when their main hope relies on them developing a naval/trading empire? It just doesn't fit?
I think those values for knights, etc, that you see for muslim personalities were simply introduced for design reasons and the ease of simply having them there rather than omitting them. There's a lot of code in all the MTW files that is redundant, either resulting from decisions to remove design ideas or code left over from the Shogun days. Look at the morale column of the crusader_prod_units file as an example.
Secondly, there's a reason the AI builds peasants excessively and it has little to do with Ai build settings, instead it's related to the design implementation of MTW itself. It's because the AI will often just build a fort in a newly conquered province and the only units available from a solitary fort are peasants. As a result the AI has no other choice of unit to build in that province If you were to make spearmen available with a fort alone and not also requiring a spearmaker, then you would effectively see no peasants, ever.
Regards
Doc
Quote[/b] ]Secondly, there's a reason the AI builds peasants excessively and it has little to do with Ai build settings, instead it's related to the design implementation of MTW itself. It's because the AI will often just build a fort in a newly conquered province and the only units available from a solitary fort are peasants. As a result the AI has no other choice of unit to build in that province If you were to make spearmen available with a fort alone and not also requiring a spearmaker, then you would effectively see no peasants, ever.
that much is true, i've eliminated peasants, no structure listed in require building, and put urban militia requiring a castle, as well as an archer. reset spearmen to castle3 only, and i notice the ai has no peasants. they still show up in rebellions though. but all i see are urban militia, archers, and spearmen, still early in campaign. plus with all the cash i gave them, they are producing crazy armies. germany is surrounded by english(me) the danish, and the french only one province. but he has at least 20 armies there all full or close to it.
"I do not believe that a catholic_crusader_trader personality ever changes to another catholic personality or for that matter another religion's personality. It wouldn't make sense and defeats the purpose of each factions' unique strengths and weaknesses. Why would the Italians or English ever relinquish their trader personalities to become musilm peaceful, orthodox stagnant or catholic isolationist, when their main hope relies on them developing a naval/trading empire? It just doesn't fit"
+DOC+ you could be right. Unless CA comments on this point I don't see how we will ever really know, since no-one to my knowledge has any idea exactly what functions the string name "Muslim_Peaceful" calls in the game engine, it's all a black box. We also don't know how often the AI personalities change or what other hard code sections affect them.
That said, it seems to me that the probability is against the sections I mentioned merely being due to redundant code. Why would they ever anticipate having a Muslim faction build probability for, say, Feudal knights? At what point in the design process would they have anticipated the need for it? How could this be left over from Shogun where there were no "muslim" factions? I think Kraellin's main point from another thread is still valid: this is just code, these are merely string names. CA could have used "abc_xyz", "123_789" or any other name. Simply because it has the name "muslim" attached to it doesn't mean that Christian factions might not change over to have that AI personality at some point, as Kraellin seems to think (if I'm not mistaken). As for "defeating the purpose of each faction's strengths or weaknesses" that would depend upon how often the AI personalities change. If they change all the time, there might not be any 1 "ai personality" in the way we think there is. How long does the AI keep it's initial personality for each faction?? 1 turn? 10 turns? all the time??? No one knows.
Since we don't know what routines they call within the game engine I don't know how we can tell.
Jayrock: I believe the AI unit preferences for rebellions are separate from the general build probabilities and have to be modded separately, unless you do this you'd still get peasants in rebellions.
I'd be interested to hear what composition the AI armies have in your campaign. http://www.totalwar.org/forum/non-cgi/emoticons/cool.gif Sounds interesting.
Have you increased the build probabilities for more advanced units? With what result?
How much starting money did you give the AI? Or did you increase the income for all provinces?
Does the AI improve it's provinces by building armor4, spearmaker4, etc., now that it has the money?
What effect have you seen on play balance? Do factions last longer? Shorter?
I would think that +DOC+'s persistent complaint about HUGE battles with 5-6,000 troops would be worse under this scenario. Frankly, I'd like to see the ai build smaller, but much tougher armies, say 12 units of Varangian guards with armor 4 and very high morale OUCH
You seem to have done what I've been putting off doing now for some time http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif I'd like to hear the results.
Kraellin
02-01-2003, 06:43
ah, busy minds that dont look are dangerous ;)
doc,
Quote[/b] ]Firstly i believe that a faction's personality can only change from its initial starting one to one of the "special circumstances" personalities, i.e.:
i would invite you to actually test this. there are ways to test it. my experience with this is that you simply ignore the labels. they are ONLY used as names of functions, but dont actually mean that a catholic_trader has to be catholic to use it. they're just names. this much i know for sure. it's a bit trickier to say that any behavior can change to any other, but the evidence in my mods points to that any can change to any, save the rebel behavior. the various parts of the names mean nothing except to call a given function and the function itself doesnt care which faction uses what; or again, at least the data i accumulated seemed to say so. it came as a bit of a shock to me, actually, which is why it stands out in my mind that this seems to be the way it is.
but, no doubt gil or target will come in here and say, nah, doc is right :) so, test it :)
K.
I'd be interested to hear what composition the AI armies have in your campaign.
so far unsure, other than no peasants in armies and they are massive, slowing my pc down.
Have you increased the build probabilities for more advanced units? With what result?
yes, as soon as they get the building i will see advanced units start appearing more regularly.
How much starting money did you give the AI? Or did you increase the income for all provinces?
didn' touch starting money, just boosted income by a factor of ten. starting to question this, massive armies are to much to track sometimes.
Does the AI improve it's provinces by building armor4, spearmaker4, etc., now that it has the money?
unsure just yet, to early only played 30 years in game. work been nuts, and leave it at that.
What effect have you seen on play balance? Do factions last longer? Shorter?
still a toss up, ran three games so far, the first two only lasted 10 turns before i noticed i had forgotten somethings.
it fluctuates, right now. so far no one has been eliminated yet but the germans, and the novgorod are close, one provence left each. one thing that's odd, when i take a provence example flanders, i had loyalist uprising for 8 years in a row before i had to abandon it. i reset all rebelious factors in early.txt to original config.
would think that +DOC+'s persistent complaint about HUGE battles with 5-6,000 troops would be worse under this scenario. Frankly, I'd like to see the ai build smaller, but much tougher armies, say 12 units of Varangian guards with armor 4 and very high morale OUCH
then he doesn't want to touch this one, their are so many armies floating around it's crazy. central europe is a blur of red, blue, black, green, and yellow stacks. put it this way, switzerland currently italian, i cant see his castle unless i cycle the z,x,c key filters. i'm starting to have doubts about my income idea because of this. that last idea is one i never thought of, it would probrably be a hell of alot easier to do, maybe cut all units size in half, and then boost some of their other attributes to compensate. if this one doesn't work out i'll try that next. tried to much on this one to abondan it yet.
side note, next game, i'm thinking of starting all provinces with a castle3 and watch towers. that away the ai can build spearmen, urban militia, and archers right of the bat.
lol, major discussion here
I totally agree that unfortunately it's almost impossible to test faction personality change and my post was purely based on speculation and assumption. http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif
We really need, and have for some time, a developer's final word on this issue. But i don't think it's coming. K, you are the test guru, i have no idea hpow to test this. My assumptions are based on that the English and Danes always seem to build tons ships unless they go bankrupt, thereby suggesting that they never revert to Catholic_Expansionist like the Germans and French who, on their vanilla settings, hardly ever built ships.
There is only so much one can do to make the AI build advanced units over the poorer early units simply because if the AI has only one undeveloped province they'll still spend money building units from it. In contrast, the human player will only build their units from the highly developed provinces and not waste money building crap units from newly conquered undeveloped provinces that could only produce spearman for example.
It was the same in Shogun, the AI kind of likes to develop every province it has and therefore there will always be an assortment of poor units among AI armies. In one of my old STW mods i had each AI clan start with developed provinces in an attempt to help alleviate the problem of fighting shit units. HOwever, the progressive downgrading of provinces following capture meant that as the game went on i actually began fighting poorer quality units (i.e. the old "AI returning to the Stone Age" phenomenon).
So far, in my opinion, the only real way of making the AI substantially tougher is to massively increase their starting money. This means that teching up and building units is never in question, no matter what the size of the faction, and that factions reduced to one province can still churn out units.
One thing i have noticed is that increasing the number of ships massively increases the amount of AI warring. For some reason ship to ship combat is one of the AI's favourite ways of starting conflict. I haven't decieded whether this is a good thing or a bad thing as yet, it certainly leads to more early extinctions (particularly the Byzantines). It does mean that i, as the human player, cannot rely on trade income as much.... but then i suppose neither can the AI.
I wonder waht the AI factions would be like with 1,000,000 starting florins? lol, maybe soemthing worth trying....
Would they get that every time they re-spawned too? That would make those re-spawns much more dangerous, as it seems that sometimes a faction re-spawns with too many units and promptly goes bankrupt and gets stuck. Just like the Byzantines did in my last game when they re-spawned on the MEd islands and then couldn't get their massive armies off them Making a shipyard not requiring a castle4 might also help this.
"i would invite you to actually test this. there are ways to test it. my experience with this is that you simply ignore the labels. they are ONLY used as names of functions, but dont actually mean that a catholic_trader has to be catholic to use it. they're just names. this much i know for sure. it's a bit trickier to say that any behavior can change to any other, but the evidence in my mods points to that any can change to any, save the rebel behavior. the various parts of the names mean nothing except to call a given function and the function itself doesnt care which faction uses what; or again, at least the data i accumulated seemed to say so. it came as a bit of a shock to me, actually, which is why it stands out in my mind that this seems to be the way it is."
Well, that's what I thought K. http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif How, by the way, did you reach this conclusion? It seemed reasonable enough so I just assumed you were probably correct. I guess you never released your mod, so perhaps you could tell us?
I'm going to complete my own next mod sometime soon, if I get the time. It'll have the following features:
1. Peasants eliminated. Spearmen either not to require spearmaker or each province to be provided with one.
2. Initial income set to 250,000 florins for all AI factions. Province income to be 10x increased (something on that order).
3. I've already lowered the build times for most structures by 1/3 - hopefully this will help the AI rebuild its provinces if it gets enough money. (By the way, giving the AI 18,000 florins to start produced no noticable improvement in the desertification problem. The AI, as DOC observes, starts out o.k., but as provinces are conquered everyone but me ends up with at most a castle. Perhaps I ought to decrease the cost of castle building?
4. (This part has been daunting me) Increase the AI build probabilities for higher tech units (including order foot and crusader knights, which I've modded to be buildable by Christian factions (each faction should have at least 1 type available, as per history)). I'm not sure if this last is necessary, if the AI had sufficient income to build up its provinces, it might build them anyway. But, then again, it might just still build hordes of militia sergeants as it does now.
5. Perhaps further modify the build_prod to increase trade. I've already done this somewhat. Interestingly, Wes W. claims that he gets good trade from his mod, which I haven't installed. I might try out his values for shipbuilding and port, etc.
6. Some possibilities I've thought of but not sure to implement (decreasing the maintenance costs for AI faction troops). But, I'm not sure this would be an improvement, and if the AI had enough money, it wouldn't be a problem anyway.
7. Reducing the build time for ships to 1 or 2 turns, so that the AI factions produce more of them. This, in conjunction with reducing the maintenance costs should help the AI maintain its navies. (BTW: DOC, although in my current mod, most nations build ships, I haven't seen much inter AI naval warfare. They mostly try to attack me.) I think I based my mod on the build prod values you sent me earlier too. Perhaps that's just coincidence).
Kraellin
02-03-2003, 07:00
Quote[/b] ]Well, that's what I thought K. How, by the way, did you reach this conclusion? It seemed reasonable enough so I just assumed you were probably correct. I guess you never released your mod, so perhaps you could tell us?
cugel,
the first clue was being able to set any faction's starting behavior to anything, except rebel. that meant that the behavior labels were just that, labels, and that having 'catholic' in a behavior name was just a label and didnt mean that a faction had to be catholic to use it. the same was true with 'muslim' and the others. these were just names of the functions, but didnt actually have anything to do with their faction allegiances.
the second clue was in seeing certain behaviors showing up during play. it became fairly obvious that a faction would change behavior based on the circumstances, but what was it changing to? i suppose, actually, the second clue was the remark in the stat files about changing factions and that they could and that the parameters were alterable and obviously could change just by what was in the files.
and, frankly, i forget the exact circumstances that arose to convince, at least myself, that any faction could take on any behavior (again, except rebel), but i do remember going, 'wow. i didnt think they could do that'.
the simplest way of doing this would be to simply make some rather blatant and exaggerated stats for the behaviors in various stats and see if this stuff shows up when the ai isnt given these behaviors to start and see if they take them on at any point. use cheat codes so you can see the stuff and just watch what the ai does. it's really just trial and error. tweak this, watch for it, if it shows up, bingo... if not, tweak it again or tweak something else...repeat as necessary.
i really have nothing i can show you at this point in time. when my harddrive failed, i fussed with it for a long while and finally just went out and bought a new computer. i've only recently transferred stuff from the old drive to this new system and still havent installed mtw on it yet, so, dont look for anything in the near future. i'm still playing around with XP, this being the first time i've had a computer with it on it.
so, if you really want to know for yourself how it works, why just go start banging on the files and convince yourself one way or the other :)
K.
No thanks K., I think I'll take your word for it http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
Actually, I think you're right because there should be game engine conflicts if the Christian factions and Muslim factions couldn't assume each other's faction behavior (you know, as when you attempt to build (make_unit in early.txt) a muslim unit with a christian faction and get an error that terminates the game).
Well, I playtested the idea of unlimited funds for the AI and let's just say it needs some more work. http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
I gave each faction, except my Byzantines, 160,000 florins to start, plus increased all province income by 10x (essentially unlimited funds). Well, the AI immediately started an enormous BRIBING spree It was hilarious and annoying at the same time The Hungarians kept trying to bribe my generals every turn until I got pissed off and elminated them. Swarms of agents all over the map bribing everything under the sun. Sometimes agents for different powers were bribing different generals in the same province at the same time, then they fought it out. I supose that the playbalance was o.k., but THIS IS NOT WHAT I HAD IN MIND http://www.totalwar.org/forum/non-cgi/emoticons/angry.gif http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
All factions are now much more aggressive and wars break out immediately. Oh, well Next, I'll cut them back by about 1/2 to 3/4 and see if that makes any difference.
In addition, if the rebellion factors are raised to 2 for each province, then the governors (add_title) for each province must be given loyalty boosting honors (eg: steward, 10+ happiness)in order to compensate. Otherwise, the rebellions become too prevelant. The idea behind this is to make it much tougher for the AI to simply take over the map and increase playbalance by increasing faction longevity through loyalist rebellions, but massive rebellions all the time are a bit ridiculous. The AI still seems to move it's now massive armies in great swarms, denuding one province (in which rebellions now break out) in order to invade another.
Perhaps if I set all the factions to Trader??
Positive notes: 1. The AI does build ports much more than it did. Even the rebel factions build them, which they never had the money to do before.
2. The rebel AI is much more aggressive and counter-attacks to regain provinces it has lost. I thought this was interesting, since the rebels now play much more like major factions -- which they should historically. The only diffence between "rebel" and "legitimate" factions is whether you win. A faction that wins, establishes itself and becomes "legitimate". All dynasties started in this fashion. The game doesn't ordinarily reflect this however.
BTW: I've added Lord K.'s units and the AI seems to prefer them to the original ones (at least in the 1st 50 years).
Kraellin: Sorry to hear that you had to buy a new computer. I know how that goes. Since I'm kind of an old fart, I just don't enjoy spending a week figuring out how to reconfigure the file system. Learning how to set up XP doesn't really appeal to me either http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
Just say after me: "I am not a man who is happy with change" http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
Quote[/b] ]In addition, if the rebellion factors are raised to 2 for each province, then the governors (add_title) for each province must be given loyalty boosting honors (eg: steward, 10+ happiness)in order to compensate
i had the same prob, what i did was took the watch towers, and border forts and increased their happiness bonus for compensation, then i increased the number for odds of building. that helped a little. that saved me from having to change every province in the text. hopefully within a week or two i can get back to modding. work is nuts right now.
still tweaking it as i can, but it's hard to find the time right now.
Ahh bribing. Now i must admit, i had overlooked that aspect of the game as the AI rarely has enough money to perform bribrs on another faction's army?
Well, that unfortunately rules out giving the AI massive amounts of starting money... shame. Back to the drawing board..
+DOC+, have you tried WesW's mod? Apparently, he has solved the trade problem. I'm going to give it a shot, as getting the AI to build up trade routes is better than denying them to the human.
gonna have to give up on the 10x income per provence, their are just to many armies, and the bribing going on is insane. gonna start anew. only certain of the following now.
1) gonna limit each provence to one trade, and one resource item. still give some trade ability, but nothing overpering.
2) gonna set spearmen, and archers to build immediately prob gonna give each provence a spearmaker, and a bowyer(sp)
gonna eliminate peasants, and rabbel from game entirely.
reset ship priorities, and reduce build time, and support.
3) gonna set building faster to build, gonna cut times in half, and remove iron prerequisite from blacksmith(or armorer, late not thinking clear) so everyone can build it.
gonna increase loyalty bonus for a couple of structures, prob watch tower and border forts. try and reduce spy rush tactic. probrably gonna boost improve farm money a little to give ai extra income.
4)try and figure which personality to set ai to, want cautious ai, don't send everything after one objective and leave self open for rebelions.. gonna re read kraellins russian q thread, they were talking about it there.
4) gonna try and implement cugel idea, cut all unit size in half, cost and support the same, but increase units abilities accordingly. any suggestions plop them in this thread, i check it daily. might prevent huge battles make doc smile for a change http://www.totalwar.org/forum/non-cgi/emoticons/biggrin.gif kidding basically. smaller more powerful units, really like that idea the more i think about it. problem gonna be balancing it.
anyway that's all i can think of right now, really tired and starting to ramble. anyway any suggestions let me know. should be able to get cracking again by friday.
simon haven't tried wes's mod yet, but will look at before i start any serious efforts anew. jay
Jayrock -- those all sound like VERY good suggestions. If you need some help with playtesting send me your early.txt file & build_prod files and I'll help: sjv1@qwest.net.
I don't want to criticize Wes W. at all since he's done a nice job with his mod (especially as I haven't even installed his mod yet), but from examination of his crusader_unit and build_prod13 files, I don't believe he has "solved" the trade problem. His values for shipyard and port in medmodv.1.5 are quite similar to the ones I had already adopted (mine were slightly higher in some instances). Reducing the build times for ships, etc. -- already tried these out (based on earlier suggestions from Kraellin + DOC). The result I saw from my own experience: Yes, the AI does build more ships and factions that never built ships before now build them. Yes, trade is improved as a result, but the AI still tends to clump it's ships in odd places and doesn't really spread out too well. The Italians still put lots of ships in the Straits of Sicily and few or none in the Adriatic. The Byzantines won't keep a ship in the Marmarra. The Egyptians still want to attack anyone who puts a ship into the Nile coast, whether this makes sense for them or not (taking overall naval superiority into account). The English still put most of their ships in the English channel (even if there is no real threat to this zone since no-one else has shiping there). They also like to put a ship at Gibraltar, but it generally isn't connected to anything. Both factions that build deep water ships like to keep them in the central med. for no reason that I can see, etc. . .
All this is hardcoded and we can't do anything to change it.
"2) gonna set spearmen, and archers to build immediately prob gonna give each provence a spearmaker, and a bowyer(sp)
gonna eliminate peasants, and rabbel from game entirely.
reset ship priorities, and reduce build time, and support."
This is very good See to eliminating them from rebellions as well. If you eliminate the unit from the unit_prod file, they should not be built in rebellions (since there's no AI build preference for peasants), but if that isn't the case, I don't know why.
Crusades: It is important to modify crusades to include many more order foot and crusader knights and many fewer fanatics and spearmen. I haven't figured out how to do this. Does anyone know how to affect the AI build probabilities for rebellions and crusades?
Weak-assed AI crusades are just sad. I beat 4 crusade armies in one battle, because they had such few good units left. They suffer from attrition because of crossing many provinces and by the time they reach a position to attack they have only a few good troops left(if that). Compare this with the armies the Pope has when he respawns (large numbers of teutonic knights and order foot, etc.). Crusades should be much more formidable. There's no worry about making them too strong so that the player has an advantage -- the player just shouldn't use them. I've modded all crusader knights (including foot knights & order foot) to be buildable anyway without crusades (if you have the right buildings), so there's no need for the player to use crusades. When playing as the muslims or Byzantines crusades are necessary to keep one honest, but aside from that, they just detract from the game because the AI can't handle them well and they are too weak. They cause the AI factions to get into ill-advised wars they would be better served avoiding.
All knights should be dismountable as well in open battle, and I've done that too (historically them sometimes did and if it's available as a feature, then it should be up to the player's discretion). If you want, I can e-mail you the unit_prod file so you can copy this, if you haven't already done it. Plus, I've integrated Lord K's new units so that adds to the mix.
Quote[/b] (Gregoshi @ Jan. 28 2003,14:01)]Ranges in the Entrance Hall has the following to contribute to this discussion:
-----
Quote[/b] ]
2. So that's one option. The other is to remove trade or limit it sufficiently so as it doesn't impact as greatly on the game. How?
a) Well one could cut down the number of tradable items per province, especially the sea provinces.
b) One could delay the arrival of the merchant improvements to later in the tech tree.
c) OR simply remove trade altogether, which may be a bit drastic.
I'm not a modder myself, but if you can alter the value of individual trade goods, you could make certain provinces have extreme value trade goods, while making most other goods very low value. This way the trade dependent factions can have one very good trading province, while all the other provinces give a much more modest income.
Still, i have no idea if this is possible, but it sounded like a worthwhile suggestion to me.
-----
This is a bad idea in my opinion. The AI does not look at province resources when making decisions on where to build ports and shipyards. It frequently builds them in provinces that have no resources
Quote[/b] (cugel @ Feb. 06 2003,17:58)]Jayrock -- those all sound like VERY good suggestions. If you need some help with playtesting send me your early.txt file & build_prod files and I'll help: sjv1@qwest.net.
I don't want to criticize Wes W. at all since he's done a nice job with his mod (especially as I haven't even installed his mod yet), but from examination of his crusader_unit and build_prod13 files, I don't believe he has "solved" the trade problem. His values for shipyard and port in medmodv.1.5 are quite similar to the ones I had already adopted (mine were slightly higher in some instances). Reducing the build times for ships, etc. -- already tried these out (based on earlier suggestions from Kraellin + DOC). The result I saw from my own experience: Yes, the AI does build more ships and factions that never built ships before now build them. Yes, trade is improved as a result, but the AI still tends to clump it's ships in odd places and doesn't really spread out too well. The Italians still put lots of ships in the Straits of Sicily and few or none in the Adriatic. The Byzantines won't keep a ship in the Marmarra. The Egyptians still want to attack anyone who puts a ship into the Nile coast, whether this makes sense for them or not (taking overall naval superiority into account). The English still put most of their ships in the English channel (even if there is no real threat to this zone since no-one else has shiping there). They also like to put a ship at Gibraltar, but it generally isn't connected to anything. Both factions that build deep water ships like to keep them in the central med. for no reason that I can see, etc. . .
All this is hardcoded and we can't do anything to change it.
"2) gonna set spearmen, and archers to build immediately prob gonna give each provence a spearmaker, and a bowyer(sp)
gonna eliminate peasants, and rabbel from game entirely.
reset ship priorities, and reduce build time, and support."
This is very good See to eliminating them from rebellions as well. If you eliminate the unit from the unit_prod file, they should not be built in rebellions (since there's no AI build preference for peasants), but if that isn't the case, I don't know why.
Crusades: It is important to modify crusades to include many more order foot and crusader knights and many fewer fanatics and spearmen. I haven't figured out how to do this. Does anyone know how to affect the AI build probabilities for rebellions and crusades?
Weak-assed AI crusades are just sad. I beat 4 crusade armies in one battle, because they had such few good units left. They suffer from attrition because of crossing many provinces and by the time they reach a position to attack they have only a few good troops left(if that). Compare this with the armies the Pope has when he respawns (large numbers of teutonic knights and order foot, etc.). Crusades should be much more formidable. There's no worry about making them too strong so that the player has an advantage -- the player just shouldn't use them. I've modded all crusader knights (including foot knights & order foot) to be buildable anyway without crusades (if you have the right buildings), so there's no need for the player to use crusades. When playing as the muslims or Byzantines crusades are necessary to keep one honest, but aside from that, they just detract from the game because the AI can't handle them well and they are too weak. They cause the AI factions to get into ill-advised wars they would be better served avoiding.
All knights should be dismountable as well in open battle, and I've done that too (historically them sometimes did and if it's available as a feature, then it should be up to the player's discretion). If you want, I can e-mail you the unit_prod file so you can copy this, if you haven't already done it. Plus, I've integrated Lord K's new units so that adds to the mix.
How to add units to crusades
You can modify the number of units and the types by editing the Unit_build_prod file in your MTW root directly. Add Crusade(xx) in the AI build priorities (use the GNOME editor available for download here). The 'XX' number should be a negative number for highest probability. -30 is usually enough to ensure at least 1 unit appears.
Quote[/b] (cugel @ Jan. 23 2003,22:45)]Good Turbo & Jayrock, I was beginning to think this thread was dead & of no interest to anyone. http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
We are all agreed on the fact that the naval AI is pretty hopeless, and the chances for another patch are slim. To make matters worse, the trade element of the game is flawed. There is a limit of 5 trade items for provinces, some provinces start with no trade items (Rome for example has none). All trade items produce the same income. Silk and Gems produce the same income as wood. This is absurb.
My approach was to eliminate the naval trade variable and use income producing structures instead. I took the default trade resources and factored them into the base income of each province. I increased the number of income producing structures. Ports, shipyards, inns, cathedrals (pilgrims), etc, produce fixed income. Naval trade is factored in abstractly with ships as combat/transport units.
This sounds like you have already produced a working mod. If so, have you uploaded it to the Org? What were the results from playtesting? How exactly did you get the structures to produce extra income (frankly, I don't know how to do this). Abstracting trade in some fashion might be a very good idea if it could be accomplished, it's the "chain of ships" requirement for trade that poses problems for the AI. CA simply installed this feature without programming the AI to follow it - thus, the P*** poor AI trade performance.
One might easily compensate for the lack of AI income by giving the AI, say 100,000 florins to start (but not the player obviously) or even more $ for the AI factions so that they don't run short -- perhaps something radical like 250,000 florins. I was thinking of playtesting this idea when I finish my current campaign to see how the AI did. However, the AI also doesn't seem make the best use of its income to purchase high end units and dissolve old armies of peasants and spearmen when they become available (thus lowering its maintenance costs, enabling it to purchase more high end units). This exacerbates the AI income problems, because the AI is hoarding large armies of peasants and spearmen. Having more $ might only cause it to build even more useless peasants, spear and a few royal knights (this seems to be the standard AI army choice). Because only 16 units can appear at one time, having 6,000 AI spearmen & peasants facing 1,000 Varangian Guardsmen and Kataphratoi is hopeless for the AI. (This is actually fairly accurate historically -- medieval armies weren't all "high tech", but it unbalances play in favor of the player, especially in the high and late periods).
Proposed Solution: Tweaking the AI build preferences to produce more "high end units" when they become available. The problem is "how much is good for the AI". I have not attacked this problem because of the daunting amount of time it would need for playtesting. I wonder if anyone has done anything along this line?
Jayrock
i'm going to try and create a mod to boost the ai. first limit every provence to one trade item, and one resource. if they dont have one give them one.. that should limit trading for the human player.
secondly gonna increase all provences income by a factor of ten, and then for the units split them in two and create unit for the human player, and the ai.. the ai leave the costs and support costs alone, and the human boost them by a factor of ten for cost and support. that away the ai will benifet from the extra income, more than the human player.
It would be easier to simply reduce the support costs for various units or increase the AI income. The problem seems to be that the AI builds too many low tech units and then can't afford the high tech ones because of the high maintenance costs. It may even consider that it doesn't need them. (This was a problem that carried over from Shogun). Of course, this would make it possible for the player to build more units, but he won't have the income to just build anything he wants, while the AI will.
my main plan is to make a mod where it is virtually impossible for the ai to run out of money and start cranking out crappy units by the gross.. gonna set all provences rebels to 2 such a cugel suggested, but only if it is a faction controled land. all rebel provences, i'll set to 4, try and slow down the ai in the begining.. been testing it for a day now, and it still needs major tweaking, may just have to limit it to 6 or 7 total factions available.. after 100 years, everyone had such huge armies, that it was slowing my system down..
My feeling is that you don't have to modify the rebellion factors for provinces with a 3 or 4 already (such as Lithuania), they're rebellious enough. I've achieved excellent play balance simply by raising all the provinces to 2 and leaving the 3 and 4 rebellious provinces alone. This definitely slows down the AI from developing 1 superfaction. If an ai faction expands too fast, it gets quickly cut down to size, plus there are loyalist rebellions that re-establish factions all the time - Few of them seem to die out permanantly in the 1st 100 years or so.
If you achieve success in modding the AI build preferences for high tech units and reducing the AI reliance upon peasants, please let us know I'd like to see the results of that. There's nothing more boring that playing battles with 3-4 waves of peasants & militia sargeants attacking you only to run away again. How do they expect to defeat my Varangian Guards or Janissary Infantry with such rabble? http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif
I don't really have one 'MOD'. I have a separate MOD for each faction I like to play. I use faction exclusion as a way of tweaking the AI behavior. If I play the French, the French are excluded from building taverns and brothels for example.
Increasing the rebellion rate on its own hurts the AI as it doesn't garrison its provinces well. A trick here is to give the AI unique buildings that the human player is restricted from building, to pump up happiness. I use the tavern and brothel. The tavern and brothel pump up hapiness and add income. The Tavern acts as a higher level mercenary magnet. The brothel adds to the taverns mercenary magnet, has a higher hapiness and income factor.
The problem with desertification (stone age) is easily fixable. The problem is in the AI building priority with the level 1 castle. The priority is too low and later in the game the AI concentrates its resources on high priority buildings. Try increasing the build priority for all factions to 500 for level 1 castles. Also include level 1 castles in your companion building priority field. You will even see Rebels and the Mongols building level 1 castles and upgrading them.
Abstract income works well. The AI typically is able to keep pace with the player, factions expand a bit slower, and warfare seems less. The AI is allowed to build woodsmen instead of peasants (the AI needs a cheap unit to flesh out its armies). The woodsmen look cool and put up a little bit of a fight before routing.
"You can modify the number of units and the types by editing the Unit_build_prod file in your MTW root directly. Add Crusade(xx) in the AI build priorities (use the GNOME editor available for download here). The 'XX' number should be a negative number for highest probability. -30 is usually enough to ensure at least 1 unit appears."
TURBO: I always use the Gnome editor. Great tool. I'm afraid I don't understand what you're saying, however.
Edited post: I went back and looked at Crusader_Build_Prod11 and now I see what you're talking about. IT's column 11 "rebelling troop mixes" not column 15 AI build prob. D'oh http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif http://www.totalwar.org/forum/non-cgi/emoticons/joker.gif This wasn't one of my brighter days.
How do we know that negative integers increase the build probability? I believe you're probably right, because I remember hearing something about this somewhere before, but I don't understand how. Do we know how these values work? Why do negative integers increase the build probability? I always assumed, and playtesting seemed to me to indicate, that positive integers increased the build probability, the higher the number, the higher the build probability. (This also seemed indicated from the fact that, for example, if the AI had already built a port and shipyard plus merchant the probability for building merchant2 is increased, i.e. a higher integer (as would be logical for the AI).
Edited post: I now notice that order foot are "CATH_HERETICS(1), CATH_ZEALOTS(5), CRUSADE(-20)" (-20 for crusades, which makes sense since that's the only way to build them without modding the game as I have done. Have you tried -30? Are there two sets of probability, completely unrelated to each other? I.e. the higher integers in column 15 increase the AI build probability, but LOWER integers in column 11 (rebelling troop mix) increase it?
Thanks for the input Turbo, a lot of very useful stuff there. http://www.totalwar.org/forum/non-cgi/emoticons/smile.gif
Back to rebellion mixes. Hmmm, i'm not overly sure how all this works really, i did ask a question in these fora a long while back hoping to get a developer's reasoning into how all this works, but as yet i heard nothing to help enlighten the issue. Basically it's not as simple as it would seem, which by a glance at the files would suggest that a higher -ve integer would result in more of that unit type in a rebel mix.
The way i tried to test this a few months ago was with the Mongol uprising in 1230. I saved a game in 1229 and then modded, reloaded, modded, reloaded, etc, to gain an idea as to what the various number setups translated to. And believe me it wasn't very clear What i did kind of see was that a higher an integer, whether +ve or -ve, translated to more of that unit type. However, this seemed to greatly depend upon what the other integers of the other potential troops mixes were set at (i.e. lots of high +ve integers didn't mean a much greater number of troops all round). What i was trying to do was eliminate or at least lessen the amount of those irritating Mongol Horse Archers whiolst bolstering teh amount of foot and heavy cavalry... and as you can see i failed somewhat in my attempt. All i seemed to do was lessen the total amount present or nerf the whole troop mix.
It just doesn't really make sense.
Therefore, unless i have a good wroking example of what does what i'd be hesitant in modding too greatly the relative troop mixes in rebellions/uprisings.
Turbo's ideas on the Brothel and Tavern sound interesting as a way for the Ai to combat internal rebellions. However, it's the instant rebellions (upon province capture) and the bandit rebellions (having less than 100 troops in any province) that really nails the AI and no buildings can help these as loyalist rebellions happen in the first couple of years and bandit rebellions only occur with lack of troops.
The idea of increasing the build priority of the level 1 castle is intriguing Maybe i will try this out and see. Once the castle1 is built though, does the AI upgrade the province further as the other build priorities associated with the level 1 castle are also relatively low compared with the higher level structures? Won't the AI simply then remain with lots of level 1 castles and pump out tons of pishy peasants?
Regards
Doc http://www.totalwar.org/forum/non-cgi/emoticons/smile.gif
"The idea of increasing the build priority of the level 1 castle is intriguing Maybe i will try this out and see. Once the castle1 is built though, does the AI upgrade the province further as the other build priorities associated with the level 1 castle are also relatively low compared with the higher level structures? Won't the AI simply then remain with lots of level 1 castles and pump out tons of pishy peasants?"
I agree. I must try to implement this idea as well I think you're right about the peasants though. That's a good reason to either eliminate them and replace them with spearmen or mod their stats to be identical with spearmen. I'm unsure which to do. What we really want, I think, is smaller AI armies that are high tech. The problem there is that the AI will then have difficulty holding on to provinces. I'm beginning to think that going back to +1 rebellion factor might be necessary. Loyalist rebellions aren't a problem for the player, but the AI can't seem to manage to avoid them by leaving a few troops in all its provinces. I wonder why CA programmed this? http://www.totalwar.org/forum/non-cgi/emoticons/confused.gif I can only scratch my head and wonder what they were thinking. Perhaps it's just an oversight. Too bad the code is all encrypted. I bet some bright org. member could modify a few routines and fix all these problems
BTW: DOC, did you note from the article posted in the main hall that trade and diplomacy are going to be fixed in RTW? http://www.totalwar.org/forum/non-cgi/emoticons/cool.gif I can hardly wait (I wonder if there'll be a retroactive patch for MTW)?
Quote[/b] (+DOC+ @ Feb. 07 2003,04:35)]Thanks for the input Turbo, a lot of very useful stuff there. http://www.totalwar.org/forum/non-cgi/emoticons/smile.gif
Back to rebellion mixes. Hmmm, i'm not overly sure how all this works really, i did ask a question in these fora a long while back hoping to get a developer's reasoning into how all this works, but as yet i heard nothing to help enlighten the issue. Basically it's not as simple as it would seem, which by a glance at the files would suggest that a higher -ve integer would result in more of that unit type in a rebel mix.
The way i tried to test this a few months ago was with the Mongol uprising in 1230. I saved a game in 1229 and then modded, reloaded, modded, reloaded, etc, to gain an idea as to what the various number setups translated to. And believe me it wasn't very clear What i did kind of see was that a higher an integer, whether +ve or -ve, translated to more of that unit type. However, this seemed to greatly depend upon what the other integers of the other potential troops mixes were set at (i.e. lots of high +ve integers didn't mean a much greater number of troops all round). What i was trying to do was eliminate or at least lessen the amount of those irritating Mongol Horse Archers whiolst bolstering teh amount of foot and heavy cavalry... and as you can see i failed somewhat in my attempt. All i seemed to do was lessen the total amount present or nerf the whole troop mix.
It just doesn't really make sense.
Therefore, unless i have a good wroking example of what does what i'd be hesitant in modding too greatly the relative troop mixes in rebellions/uprisings.
Turbo's ideas on the Brothel and Tavern sound interesting as a way for the Ai to combat internal rebellions. However, it's the instant rebellions (upon province capture) and the bandit rebellions (having less than 100 troops in any province) that really nails the AI and no buildings can help these as loyalist rebellions happen in the first couple of years and bandit rebellions only occur with lack of troops.
The idea of increasing the build priority of the level 1 castle is intriguing Maybe i will try this out and see. Once the castle1 is built though, does the AI upgrade the province further as the other build priorities associated with the level 1 castle are also relatively low compared with the higher level structures? Won't the AI simply then remain with lots of level 1 castles and pump out tons of pishy peasants?
Regards
Doc http://www.totalwar.org/forum/non-cgi/emoticons/smile.gif
For the rebellion mixes, a high negative number increases the probability of the unit appearing but not necessarily increasing the number of appearances of the unit.
For AI building priorities a high number increases the probability of building it. The highest setting here appears to be 699.
The AI does a fair job of developing once you give it a level one castle, however, you can get as detailed as you want. You can use the building companion field to chain building structures one after another. After the level 1 castle is built, have a Town_watch(500) in the castle level 1 companion field. When the town_watch is built, have a Spearmaker(500) in the Town_watch companion field. In my build_prod file, I have structure building development up until level 7 castle. Even the rebels upgrade their castles (but slower).
I developed a workaround the instant rebellions. I noticed the Novgorods were disappearing early in the game. To help them out, I gave the Boyar unit a hapiness factor similar to the priests. You can apply the same trick to other units to tone down loyalist rebellions. You can also determine what types of units appear in bandit and loyalist rebellions. The bandit incursions are you point out only hurt the AI, so I eliminated most of the labels.
For the rebellion mixes, a high negative number increases the probability of the unit appearing but not necessarily increasing the number of appearances of the unit.
For AI building priorities a high number increases the probability of building it. The highest setting here appears to be 699.
The AI does a fair job of developing once you give it a level one castle, however, you can get as detailed as you want. You can use the building companion field to chain building structures one after another. After the level 1 castle is built, have a Town_watch(500) in the castle level 1 companion field. When the town_watch is built, have a Spearmaker(500) in the Town_watch companion field. In my build_prod file, I have structure building development up until level 7 castle. Even the rebels upgrade their castles (but slower).
I developed a workaround the instant rebellions. I noticed the Novgorods were disappearing early in the game. To help them out, I gave the Boyar unit a hapiness factor similar to the priests. You can apply the same trick to other units to tone down loyalist rebellions. You can also determine what types of units appear in bandit and loyalist rebellions. The bandit incursions are you point out only hurt the AI, so I eliminated most of the labels.
You know Turbo, if you've done all this it's really quite interesting. You probably should consider uploading your mod to the org. so that others can see it. http://www.totalwar.org/forum/non-cgi/emoticons/cool.gif
BTW: I assume that you're referring to field 11 "AI building influences for building combos labels"? What values did you use?
I presume that the empty brackets signify add nothing to the probability of the 1st level building (which you would need since the AI build prob. for building castle1 can't add 100 for building castle1.
This is bloody brilliant I never thought of that http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif http://www.totalwar.org/forum/non-cgi/emoticons/idea.gif
vBulletin® v3.7.1, Copyright ©2000-2025, Jelsoft Enterprises Ltd.