The AI cheats financially in Kingdoms

[Jambo]

Sorry for going OT but BtS does exactly the same on the difficulty settings where it becomes challenging (for argument's sake let's say Prince and above, although it depends on the player's skill level), AI builds cheaper, researches faster, has less maintenance and pays a lot less for upgrading units.
The AI is actually so easily exploitable by gifting them, giving them tributes and paying them to attack other people that I can win about 2/3 of the games with a decent starting position on Emperor difficulty. Most through domination, admittedly, where it's very hard for the AI to keep even (on the tactical level I mean - a lot of times I have K:D ratios of 3:1 and more) but still.

All in all it's as I said on a thread on civfanatics: The AI is very much shorthanded in the brains department so consider handicaps as a way to partially make up for the unfair advantage you have over the AI.

True, but the actual AI which controls the units, build decisions, etc, is a vast improvement over what it was in vanilla Civ 4. The AI can actually contend in the wars, rather than just cheating its way to a space victory.

[Monsieur Alphonse]

@Jambo

At least that explains probably why I don't get the extra money when I am promoted during my New Spain campaign. According to the script and the pop up New Spain should 5000 at the first promotion and 10000, 15000, 20000, 25000 at the next promotions but I have never experienced that I get that extra money.

The line in the campaign script is:

add_money spain 5000

Which may bee should according Jambo with a comma after spain.

Can someone confirm that the player doesn't get the extra money after a promotion in a New Spain campaign?

[Jambo]

Well, it would seem it is bugged.

Does anyone know if the "increment_kings_purse" command also requires a comma?

[Zenicetus]

That's not possible. It might be possible if PC's came equipped with a specific AI CPU.

Sure it's possible. There are good games out there where the AI doesn't cheat, like GalCiv2. It's programmed to recognize and counter common player strategies, and it's very tough to beat at the harder levels, with no cheats like economic advantage, lifting fog of war, etc. GalCiv2 does have the advantage that it has no separate tactical AI (all battles are auto-calc'd), so the programmers can focus on turn-based strategy without any distractions.

I probably won't buy Kingdoms for other reasons (see the SecureRom discussion), but just as a personal opinion... since they weren't going to rebuild the AI from scratch, I don't mind this way of making the game more challenging. I do think it's unfortunate that it's tied to the campaign difficulty setting. It would be nice if that were separate, so you could use a medium difficulty where diplomacy isn't broken by the constant normalizing towards aggression that happens at H and VH settings. Maybe the same thing could be done by modding in a King's Purse boost for medium difficulty.

[Rumpole]

Yup, i've won with New Spain, though i did lose 39 battles in the process.



It's not a bug, it's designed that way so to provide more of a challenge on harder difficulty settings.

The bug (in my opinion) is in the part of the script that gives 10000 florins per turn to the Tarascans and the Chichimecs, independent of the difficulty level.

As an experiment, I set up a game at Easy/Easy and ran through 10 quick turns.
Almost everyone was broke, but the Tarascans and the Chichimecs had about 40K.

[alpaca]

This is true. I think Civ has always had one of the best AIs in the strategy genre, but it also has cheated outrageously at higher difficulty levels. The cheating is what makes it challenging in a way that most TW games have not been.

I think vanilla TW could cheat more in favour of the AI - a money top up is the obvious route; mods like EB use it successfully. Over in the Throne Room, we've been making sure AI factions have 50k or so in the bank in our M2TW HRE PBM. This, together with limits on player expansion, has meant that that AI has gone from fielding virtually no significant armies to fielding forces that can match our own (and indeed sometime out-tech us).

However, Civ does have some better AI behaviour than TW (as well as some limitations). Having played Civ4 recently, the two obvious ones to me are:
(a) In Civ, when an AI goes to war, it usually does so very seriously - it may mass several scarey stacks and slowly march halfway across the world to get at you. It can be quite frightening to see those looming out of the fog of war reading to take down a border town (and keep going). By contrast, TW still tends to go for penny packets, not just ruthlessly compiling a stack of doom and not going for the jugular, being content to just snatch the odd settlement.
(b) The Civ diplomatic programming is great fun - dealing with the varied personalities of the different factions (Ghandhi vs Shaka etc) makes that side of the game rather rewarding. In TW, AI diplomatic behaviour is often hard to make sense out of and almost never pulls off the trick of feeling "human" (e.g. you don't feel you have annoyed a faction or that they are your friends, etc).

Yeah the problem lies with the "default state" being a kind of war in TW states.
Not completely unrealistic for a Medieval setting, mind. Most wars back then didn't go to the end in a sense like wars today.

In Civ, most AIs don't go into full-scale war mode either though, you often see them having large stacks in their own cities, far from the front. To be honest, war in Civ4 isn't much harder than in TW.
It got a bit better with BtS but not by a huge margin, at least the AI more often creates larger stacks now, but this obviously also opens up collateral damage possibilites

Edit: Jambo, "As a long term modder of campaign scripts in Med II" you should've realized that add_money is a new command. You're referring to the console_command of the same name which however doesn't necessarily have the same syntax. I'd say it probably works (can't imagine their scripters screw so big time) but can't test it since I don't have the game.
Why don't you just try giving yourself money with it? It's a two-minutes' test.

[Jambo]

As a long term modder of campaign scripts in Med II, I'm rather perplexed at the code entries for the CA-inserted AI money scripts in Kingdoms...

	monitor_event FactionTurnStart FactionType england
		and IsFactionAIControlled
		and Treasury < -5000
		
			add_money england 10000

During our testing of money scripts in vanilla we discovered that there had to be a comma after the faction's name but before the money award. Well, looking at the various new money scripts in the Britannia campaign, they look like they're not going to work! Based on my knowledge the one above should be:

	monitor_event FactionTurnStart FactionType england
		and IsFactionAIControlled
		and Treasury < -5000
		
			add_money england, 10000

Anyone confirm to the contrary?

So, unless anyone can prove otherwise, it would seem all this debating over the merits of the AI getting financial cheats is inconsequential...

Personally, I have found the AI getting tougher towards the latter half of the Britannia campaign, so my guess is that the increment_kings_purse might be working... even though I couldn't get the actual command to work from the console.

[Lusted]

Personally, I have found the AI getting tougher towards the latter half of the Britannia campaign, so my guess is that the increment_kings_purse might be working... even though I couldn't get the actual command to work from the console.

Jambo, as alpaca pointed out, there not console commands, the new add_money command appears to be different to the console command version. From my playing of the campaigns i would guess they are working as the ai does produce a lot more troops on VH, and generally better quality troops.

[Didz]

That's not possible. It might be possible if PC's came equipped with a specific AI CPU.

I disagree, I've played games where the AI provides a challenge without cheating. (Combat Mission being one example that springs to mind)

What we are looking at in this case is a simple example of lazy programming and game design. CA can't be bothered to go through the effort of making the AI more of a challenge as its far easier just to have it cheat.

[Jambo]

Didz, have you played Kingdoms?

I've found the AI to be far better than it is in Med II 1.2. I should say I also use a few personalised AI tweaks to help the AI armies reform better when on attack and defence, but on the whole, Kingdoms is the best TW to date, and by a long way. So, I think you're probably being a little harsh here.

I also can't actually think of any strategy game that doesn't impart bonuses (read "cheats" for some here) to the AI at tougher difficulty levels...

[Jambo]

Lusted, Alpaca,

I'll do a test later, things may be different in Kingdoms. If I recall correctly I did this under Med II and despite giving large amounts (40000 max) using the add_money command in a script (without the comma) the financial graphs of the AI factions never showed any evidence they were working. With the comma, they did....


Edit: There's also Monsieur Alphonse's post to consider too. He's not received any of the monetary awards in the Spanish campaign.

[Mad Mac]

I seem to remember getting truckloads of money every time I earned a title as New Spain, for what it's worth...

[Jambo]

As a long term modder of campaign scripts in Med II, I'm rather perplexed at the code entries for the CA-inserted AI money scripts in Kingdoms...

	monitor_event FactionTurnStart FactionType england
		and IsFactionAIControlled
		and Treasury < -5000
		
			add_money england 10000

During our testing of money scripts in vanilla we discovered that there had to be a comma after the faction's name but before the money award. Well, looking at the various new money scripts in the Britannia campaign, they look like they're not going to work! Based on my knowledge the one above should be:

	monitor_event FactionTurnStart FactionType england
		and IsFactionAIControlled
		and Treasury < -5000
		
			add_money england, 10000

Anyone confirm to the contrary?

Whoever wrote this is an idiot.

Code in Kingdoms works fine.

[TB666]

After playing kingdoms for a while I can say that the AI builds alot better troops now and is generally tougher too.

[Didz]

Didz, have you played Kingdoms?

Nope! I bought two copies and then discovered they were trojan horses for the SecuRom hack, so I haven't installed them on either of my PC's yet. Still dithering about whether I want to knowingly install a piece of invasive software on my systems that might bite me when I least expect it.

The fact that this thread is adding so many other issues, like AI cheats, into the mix is merely making me even less inclined to take the risk.

[Wessex Rob]

I've always been interested in the AI aspects of strategy games. When I increase the difficulty of a game, I don't want the enemy to have falsely inflated morale bonuses, their troops to be individually harder to kill, or their treasuries gifted so that they can buy their way to victory. What I'd like to see is the enemy playing smarter, using more innovative and/or complicated tactics. The same could go for better generals (ie as more command stars are gained, more developed tactics are used) at the same difficulty level. I'm always disappointed when a 7 star general has (seemingly) the same level of tactical ability as a mere captain.

Taking it further, perhaps human players could be limited in terms of what they can do by the level of the commander of a force in a battle ie if your force is only commanded by a captain you can only complete basic tactics, like hold a line in front of a brigde, or advance in a line. Attempts to complete more complicated manouvres could have a chance of ending in a confused disaster, unless you have a general with some command ability.

[Jambo]

Didz, the AI has been cheating its way to losing for decades.

[alpaca]

I've always been interested in the AI aspects of strategy games. When I increase the difficulty of a game, I don't want the enemy to have falsely inflated morale bonuses, their troops to be individually harder to kill, or their treasuries gifted so that they can buy their way to victory. What I'd like to see is the enemy playing smarter, using more innovative and/or complicated tactics. The same could go for better generals (ie as more command stars are gained, more developed tactics are used) at the same difficulty level. I'm always disappointed when a 7 star general has (seemingly) the same level of tactical ability as a mere captain.

Taking it further, perhaps human players could be limited in terms of what they can do by the level of the commander of a force in a battle ie if your force is only commanded by a captain you can only complete basic tactics, like hold a line in front of a brigde, or advance in a line. Attempts to complete more complicated manouvres could have a chance of ending in a confused disaster, unless you have a general with some command ability.

How would you implement this? I mean in TW you have total control over your units so would you want to introduce a kind of random factor to check whether the units do what you want or something completely strange?
Could be a nice idea for some players (including me ) because it's actually pretty realistic but I guess most people would scream for vengeance

[Wessex Rob]

I admit it would be difficult to do, and some players would doubtless get annoyed if their brilliant tactics degenerated into chaos (I include myself here!). I'd envisage a sort of sliding scale so that a simple manouver by a well trained, well-led unit would be flawless, but more complicated marches and counter-marches lead to a loss of cohesion. I'd even like a bit more "fog of war" on the battlefield, not just in terms of what can be seen, but in terms of what can be done and by whom. I can't imagine that these things would be easy or even (universally) popular.

Personally, I wouldn't want units doing something completely random, but maybe misunderstanding slightly; so advancing too far/not far enough, moving too fast/slow to where they're supposed to go. A classic would be the disappearing cavalry charge...you know the one, where the elite charge through the enemy and keep going...and going...and going, maybe if the commander lacks command stars it could increase the chance of them leaving the battlefield whilst in pursuit and (maybe) not even coming back before the battle ends.

Alternatively, you could limit the number of orders per minute according to the skill of the commander. So a mere captain can pass down one order a minute, eg Everyone (select all units) Move Forward, or shift the position of one unit of the army precisely. Meanwhile, a 4 star general can issue 4 commands in the same time eg units 1+2 move forward, unit 3 move back, unit 4 form schiltrom, unit 5 change to wedge formation etc. A timer could show how long before the next order is available. This would make "simpler" tactics the only thing achievable with low level commanders, and more complicated plans on the player's part would require a general capable of carrying them out. It would also make grouping of units vital, as each order to a group would only count as 1 order.

Probably all very difficult to do, but I think it would be interesting to fight with slightly more realistic limits to what you can do...and probably very frustrating at times!

[_Tristan_]

I'd even like a bit more "fog of war" on the battlefield, not just in terms of what can be seen, but in terms of what can be done and by whom.

You can get a feeling of that by selecting the "General camera" option... It restricts your vision to what your general can see of the battlefield... Makes it much harder to see what is happening...

[The historian]

Upon reading this thread i realized why the stalin tactic failed in my teutonic campaign against the lithuanians guess you can;t beat the ai economically and without fighting many battles, I'm a tad disapointed economic warfare is just as important i am disapointed at CA.

[IvarrWolfsong]

What is even more annoying than 10k armies spammed at you is taking Jerusalem with your Ghulam/Mamluke horde and still having the AI launch armies with Marshals and Constables every turn. It has been 16 turns since I took the city and the AI still keeps cranking these REQUIRES JERUSALEM units out.

I am all for AI boosts but it really takes the fun out of capturing key cities if the AI just ignores the rules.

[TevashSzat]

Over in the Throne Room, we've been making sure AI factions have 50k or so in the bank in our M2TW HRE PBM. This, together with limits on player expansion, has meant that that AI has gone from fielding virtually no significant armies to fielding forces that can match our own (and indeed sometime out-tech us)

Yeah, I am actually grateful that we started giving the AI money. We were just steamrolling through everything before it and now with the AI having so much money, they pump out full stacks like crazy. It actually makes fighting battles alot more challenging, not as much in whether you win or lose, but on how many troops you lose as the deficent AI makes winning battles pretty easy, but large stack sizes destroys your armies through attrition.

I think of giving the AI money as offsetting their innate inferiority in fighting battles as well as their economic management

[Moah]

I think of giving the AI money as offsetting their innate inferiority in fighting battles as well as their economic management

Absolutely. As Wales it wasn't could I beat the English full stack coming at me, it was can I beat it with almsot no losses because there was another one behind it and then two more coming next turn...

The AI is still easily beatable unless you have a far inferior stack. The challenge is surviving the war of attrition.

And for those complaining it's just the AI trying to do to us what we did to the Mongols...

I presently control all of England an dwales in britannia and was thinking of quittign because I figured now I was 1 in all categories the challenge weas over. Figured the scots would be too easy if we're even stevens. but maybe it won't be if they've got some cheats up their sleeves...

[Swoosh So]

Scots cheating? never! Btw most of the good mods give the ai cash and i believe al totalwwar games have done to keep the ai out of debt its nothing new to kingdoms.

[CeltiberoMordred]

Is there any cheat for the AI on battlefield or the only one is more money in campaign?

[GFX707]

There's some fairly moronic logic being deployed in this thread. I won't name names.

Why don't you guys try playing the game facing the other way? I'm sure that will make it more challenging. How about breaking your clicking finger with a hammer? How about playing after a hefty dose of tranquilisers? All of those things would make it more challenging.

You're as bad as the "Oh, so what if they left (insert feature) out? Someone will eventually mod it in" crowd.

If you keep giving CA and Sega breaks like this you'll keep getting "Graphics - Total War" forever.

[Moah]

There's some fairly moronic logic being deployed in this thread. I won't name names.

I totally agree with the first part.

But I suspect the names we're not naming are very different...

[Didz]

Well for my part I don't find beating a steady stream of full stack armies at all challenging. My word for it would be BORING.

What we needed was a AI that could field a decent army, give it a sensible goal and use it effectively. What we seem to have got is an AI, thats just as dumb as before but now has the ability to bore you into submission.

[GFX707]

Well for my part I don't find beating a steady stream of full stack armies at all challenging. My word for it would be BORING.

What we needed was a AI that could field a decent army, give it a sensible goal and use it effectively. What we seem to have got is an AI, thats just as dumb as before but now has the ability to bore you into submission.

You're right. There's far too much spamming in the last few CA offerings. Spamming of annoying voice messages on the battlefield, spamming of full stacks on VH, spamming of agents....the list goes on.

I hope they start from scratch with the campaign and interface on ETW.