Printable View
There is big difference in custom battles when AI faction is acting like attacker or defender.
If you don't set anything in custom battle they act as attacker, and are playing agressively.
If set to defender, if weaker they will camp on nearby hill, and wait your forces.
P.S.
Yes there is an option, just near team markings (I, II, etc...), the red thing.
Thanks player1. Yes it does say clicking the shield so it lights up red makes the team defender. I clicked the shield next to my faction in the last test, and that's why there was no difference in the AI behavior.Quote:
Originally Posted by player1
I made the AI archers the defender in another 3 archer vs 3 hastati test, and the archers didn't move. They stood with 2 in front side by side and the general centered behind. When I advanced the hastati, at a certain distance the 2 front archers moved forward to what looked like max range and started shooting. I stopped the hastati and the 2 archers fired all their arrows at which point they moved back to their original position. I moved the hastati forward somewhat and the AI general moved forward and fired all their arrows and then moved back to its original position. The hastati each had about 40 men left.
I advanced to about 50 meters away and the front 2 achers charged. This attack was repelled with the archers having 47 and 45 men left which rallied and returned to approximately their original positions. After that, the archers didn't want to attack and gradually shifted to the sides with the general moving back when I advanced. Eventually, the 47 man archer made a flank attack on my left hastati, and it was repulsed. I sent one hastati forward, and the AI general's unit ran around behind my hastati line near the 45 man archer. Using the hastati on the right, I moved toward the general, and the 45 man archer attacked and was repused. It rallied and came back and attacked the hastati again, and this time the AI general charged in support and routed the hastati. The AI general pursued and ran into my other hastai and lost after some fighting.
So, the AI's behavior is quite different in this testr when it is the defender. I've observed this in my campaign battles as well, but I have had campaign battles where I was the attacker and the AI defender acted aggressively and attacked me.
THat's what I thought would happen. From your tests custom battles do appear to be different from campaign battles - in campaign battles missile units will retreat without necessarily resorting to melee, as they can fight another day. Like I said, in custom there isn't this option and it would be frustrating to play custom battles where the AI always retreated.
16. Wardogs still pose absurd difficulties, because they still can't be targeted directly. So on VH, where they have plenty of morale they just keep killing and killing in AI hands. They are non-sensical anyway, so it looks like I need to mod them out once again. GRRRR!!!!
Interesing.
I thought that first time in BI factions that have them make sense.
I have actually found wardogs to be less of an issue now.
They take 2 turns to train (didn't it use to be 1 turn?), the handlers need to get a whole lot close to release their little pets and the dogs aren't as powerful or numerous as before. I thought I was bright to bring some with me on rebelhunting against Steppe Raiders... Well, that turned out to be a resounding failure of the dogs.
They couldn't even deplete a single unit of enemy archers despite suffering no losses to archery.
And the AI doesn't use them in important capacity as far as I have noticed.
Can anyone tell me if 1.3 fixed the coastline bug? Where in a battle-map the water close to the coast was turned into a big flat dirt plain? This wasn't there in vanilla, but was introduced with 1.2. I would think 1.3 would fix it, after all the complaining about it, but I haven't had time to try 1.3 out yet and was really curious about this one point in particular. Can anyone answer that?
I've been watching for this problem, and haven't seen it in RTW v1.3. However, I've only fought a few battles near the coast.Quote:
Originally Posted by Teleklos Archelaou
The impact is primarily morale/flanking/enemy in the rally area. While they didn't seem to kill very rapidly, their ability to insta rout units already in combat and then enter the rear (insta routing a bunch of others) was decisive. If I could target them, it would be manageable. With them causing havoc and running wild in the rear without the opportunity to engage them or keep them out of the rear, I'm crying "foul." The morale hit is too great on VH. Having one unit of wardogs effectively cause the rout of an entire army is BS.Quote:
Originally Posted by Kraxis
I think the coastline bug is gone. I haven't seen it yet. The Roman AI hasn't been as likely to attack my coasts in 1.3 so I haven't had as many battles along there.Quote:
Originally Posted by Puzz3D
I've fought a number of battles on the coast with no glitches with the coastline.Quote:
Originally Posted by Red Harvest
Well it can't be right at medium difficulty and also right at very hard. Dogs are a nuisance at medium, but haven't been a decisive factor in any battle I've had.Quote:
Originally Posted by Red Harvest
I don't doubt that, VH makes them more of an exploit for the AI. Similar for the flaming arrows. On VH, the AI loves those stupid flaming arrows. It can see the morale and knows when to go after units with them. While I applaud it for making a good decision with its archers for a change, it is in essence making super weapons out of a non-historical use of flaming arrows. Immersion killer. Might as well have Zeus throwing thunderbolts at my army, or maybe Sponge Bob Square Pants annoying them to death.Quote:
Originally Posted by Puzz3D
I wouldn't mind the dogs so much if they could be directly confronted, or there weren't 3 or 4 sets of them in each AI stack at times--eliminating the counter of shooting them to death at range before they release. Druids and Screeching women don't bother me as much.
Wardogs have been 2 turns to train since 1.0.Quote:
Originally Posted by Kraxis
Heh... Ok. I never used them much, and prolly never will.~:)Quote:
Originally Posted by gardibolt
Per the lack of missile unit/skirmish unit effectiveness by the AI: it is DEFINITELY a huge problem in the campaign. The AI can't use its ranged units as much more than a nuisance.
Example:
In my Carthaginian campaign on VH/VH, I've had a string of consecutive battles with archers where the archers tried to attack my melee units in melee. My standard mode of operation whether sallying or fighting open field is to put my Balearic slingers in front, backed by spears, then swords, with cav on both wings and general in the rear.
How does the AI handle this? Well, in two back to back battles vs. the Julii, the Romans have had the reforms and are pumped up with early cohorts, archer auxilia, light auxilia, and other auxilia, as well as onagers, scorpions etc. I'm stuck fielding Libyan spearmen, spanish/barb mercs, Balearics, and round/long shield cav in this section of Gaul. In the 1st battle the force ratio was 3:1 in favor of the AI according to the slider (similar number of men, but their stats were WAY better.) The AI actually outnumbered me in the archer auxilia vs. balearics category in several battles. Yet in 3 battles it killed only about 20 Balearics, and most of those were from the 1st battle where I got caught unprepared by an onager.
So what do you think the AI did? Did it: A) Put its archers in front and engage my ranged units with the superior range (170) of the auxilia vs. balearics (120.) B) Hold its archers in the rear cutting the balearics down out of range. C) Rush its melee units forward en masse avoiding the ranged combat altogether. D) Backup the army for a while, then advance 1 archer unit at a time against my 3 partial units of Balearics, allowing the archers to be cut to shreads, sometimes sending a few melee/cav units forward in a piecemeal attack supported by more archers attacking as melee infantry.
If you guessed D you get the prize. ~:eek:
Of course, with its overwhelming force, the AI withdraws a distance when I sally. In doing so, it lets me pepper it with Balearics shooting into the rear of its moving men. It also leaves the slow moving onagers/scorpions/ballistae unprotected vs. my cav. Fortunate, since I don't have a counter to the onager at the moment.
I'm not doing anything gamey. I'm just moving out the front gate and using a logical deployment: missiles/spears/swords with cav on flanks and general in the rear/center.
The openfield battles are a bit different in that a mass rush by the AI is the norm. Again, the AI fails to directly engage my long ranged units with its own long range missile units before charging my lines. The archers have a tendency to engage in melee as they are moving forward in melee support role, rather than seeking good missile engagement range.
The AI is giving up one area where it should be on close to even ground: head to head missile exchanges. I'm not doing anything gamey, simply putting the ranged units in front so that they make 1st conatct.
This is as broken as the AI's failure to use pila.
I'm sure it could be improved, but unlike the pila problem the ranged unit AI works the way it was designed to work. It's not going to seek out an even exchange. It's going to attempt to shoot at units more valuable than enemy ranged units if they are present. The AI definitely rushes exposed enemy skirmishers. I think it uses cavalry for this if available, but it does use skirmishers as well possibly because they are fast moving. Maybe if skirmishers didn't run so fast the AI wouldn't use them to rush. I know in MP that shooting enemy melee units is what players try to do, and they try to neutralize enemy ranged units by charging them with cavalry.Quote:
Originally Posted by Red Harvest
The change in v1.2 to reduce friendly fire casualties causes the AI to stop shooting if its target engages in melee which can lead to ranged units not shooing at anything. I've seen this with my own ranged units which are on fire-at-will. I think that's good otherwise ranged units would be shooting into the backs of their own men.
Since no one in a ranged unit can shoot until all the men stop moving, an AI ranged unit can take a lot of casualties as it marches forward to a shooting location. I noticed in archer vs archer tests in custom battle, that this causes the AI to loose every shootout. So, if you set the AI up to engage enemy ranged units it would loose almost every time.
One thing I notice about the AI is that it doesn't take into consideration possible losses due to ranged fire. It's decision to charge is apparently being made only on a comparison of melee combat stats and possibly unit speed.
Puzz3D,
There is a far larger problem here. Even with its archer auxilia out in front and stationary, I'm noticing that they are not firing on my approaching slingers (who are within their range.) They are just standing there, until they either start taking fire, or the AI does a formation change. Something is seriously borked.
Doesn't seem to effect jav cavalry as much, they always charge forward then halt. That part is the the AI's way of chasing off other skirmishers. Fairly one dimensional, but not a bad approach.
RTW's AI is fairly well porked when it throws its archers into melee behind other units it charged forward, rather than using two lines of cohorts directly behind it.
Coupled with the lack of pila use, the Roman AI is seriously defanged. It's best ranged units are wardogs. ~:rolleyes:
I just set up a custom battle test with the AI as attacker having 3 hastati and 2 archer auxilia, and me having 3 libyan spearmen and 2 slingers. The AI had 2 hastati in front backed by 2 archers with the hastati general behind. My formation started with 2 slingers out front backed by 2 spearmen and my spearman general behind. The AI played perfectly. I remained stationary and the AI formation moved into archer range and started shooting my slingers. They fired continuously without stopping. I then advanced so that my slingers were in range of the hastati, and started shooting. They actually killed more archers which were close behind the hastati. The AI contined to shoot my slingers even though my spearmen were in archer range. Eventually, the hastati I was shooting stepped back out of range, and the slingers didn't have another unit in range so they stopped shooting. At that point, the AI archers targetted my spearmen. I advanced back into slinger range, and the AI hastati charged with the archers firing at the slingers in support. The archers didn't advance. Two hastati charged my spearmen while the 3rd chased a slinger off to the side for a short distance. That hastati turned back to help in the melee against my spearmen, and the archers resumed firing at the slinger off to the side while that slinger fired into the back of the 3rd hastati.Quote:
Originally Posted by Red Harvest
The AI played very well in this test. I wonder if it's getting confused in larger battles where there are lots of different unit types operating.
.
In a recent campaign battle I (Carthaga) fought against SPQR, unless I was having an illusion, the AI had the Hastati trow the pila before charging. Their archers were firing, presumably well protected beneath the infantry.
self_plug
{
It didn't help against my heroic victory and destroying the Scipii a couple turns later.
https://img469.imageshack.us/img469/743/00006qs.th.jpg
}
.
Had an interesting non-siege battle that reveals the depths of the problem. It was trees and plains, with rolling hills in lower Gaul. The AI was on defense, and I had about 50% more men. I had three partial units of Balearics in front, with spearmen, cav, and some merc spanish infantry, and general. The AI had some legionary cohorts, some auxilia (spears) a heavy cav unit, and an archer auxilia.
The AI had the archer auxilia in the center of its formation in a clearing and I wanted to keep its army out of the woods. I advanced my army through the trees. stopping near the tree line as my balearics reached range. At this point, the archers should have gone to town on the slingers, but they sat there not firing. I let auto targeting hit the enemy infantry. The archers standing just behind took a worse beating than the heavily armoured/shielded infantry. They never fired. Finally the AI turned tail and ran, never firing a shot. I chased down much of the force.
Then again, I had one siege battle where the AI actually engaged with its force of archer auxilia. I started counter targeting its two full units with my 3 partials. At that point it started the traditional charge forward. Not much exchanging going on before it tried to melee. That is unfortunate, because it had the opportunity to give my missile units a pounding.
I don't know why the AI archers didn't want to shoot, but it's pretty clear that the AI doesn't realize its going to take casualties from ranged fire. So, it stands there when under long range fire until it's weakened to the point where it runs away. In my custom battle test, the AI hastati charged when I moved my slingers inside their own maximum range. Apparently, my slingers moved within the necessary proximity distance to trigger the charge. Baleric slingers are longer range, and can shoot effectively from outside this proximity distance therefore never triggering a charge. Maybe those slingers are outside the distance necessary to trigger a shooting response as well. STW didn't have different ranges for different types of shooters. All shooters had the same range which meant shooters engaged shooters because melee units would be out of range, and melee units engaged shooters when they moved up within a certain distance. I wonder if the RTW AI is still designed for that situation, and the introduction of vairable ranges is beyond what the AI can handle intelligently.Quote:
Originally Posted by Red Harvest
It was the other way around in the slinger vs archers tests that I ran. The slingers beat the archers easily despite their shorter range and being under archer fire constantly. I didn't even use loose formation with the slingers and walked into firing position while the AI did switch its archers to loose formation once they started taking losses.Quote:
Originally Posted by Red Harvest
One thing I noticed with archers is that, if there are friendly units in front of them and the trajectory is too low, no arrow is released. However, the archers are still doing their reload and firing animation. I guess this is the result of the high velocity being used for the arrow. In my custom battle test on flat ground, the AI had its archers too close behind the hastati, so when I moved my slingers closer only a few archers who extended beyond the sides of the hastati were actually releasing arrows. It's possible the high velocity was introduced to help ranged units hit moving targets because, in STW, archers could not hit moving targets unless the targets were moving directly away. This helped cavalry beat archers because the archers would overshoot cavalry which was charging at them, and was purposefully designed this way to help the cav > ranged > spear > cav RPS. Some MP players picked up on this and would try to make you waste arrows by moving the units at which you were shooting.
While I didn't mod any STW/MI stats I don't think that is correct. The troopstats file shows quite a few different weapons: the ninja's had very short range, as did the Korean javelinemen and thunderbombers. The two firearms had ranges longer than the bows. The projectiles files show differing ranges and speeds.Quote:
Originally Posted by Puzz3D
In addition to the motion effects, the velocities are problematic in that they produce unrealistically flat trajectories and give the missiles too much reach vs. elevated targets. This also allows them to "chase" targets well out of range who are retreating when their firing sequence starts. Accuracy against stationary targets is greater with flat trajectories (imagine looking at a unit's site profile from above vs. head on.) This is the only distance attenuation effect I have noticed by the way, so reducing velocities does reduce hit percentage at distance.Quote:
One thing I noticed with archers is that, if there are friendly units in front of them and the trajectory is too low, no arrow is released. However, the archers are still doing their reload and firing animation. I guess this is the result of the high velocity being used for the arrow. In my custom battle test on flat ground, the AI had its archers too close behind the hastati, so when I moved my slingers closer only a few archers who extended beyond the sides of the hastati were actually releasing arrows. It's possible the high velocity was introduced to help ranged units hit moving targets because, in STW, archers could not hit moving targets unless the targets were moving directly away.
(I think most of what changed with respect to movement vs. archery was actually automatic "lead" targeting in RTW. This is an improvement.)
The "range" stat is actually a "targeting range." It won't allow intentional targeting out of that range, but the extra velocity can be used to hit things out much, much farther when the unit is running away, or if some part of it can be targeted while the center is far away. (You can hit war dogs trainers this way from well over 300 meters using slingers...)
Another issue is the max angle allowed is probably too great (70 degrees). I suspect it should be limited to 45 degrees or so. Otherwise you get some strange "indirect fire" effects as if they were shooting mortar rounds. This probably allows too many units to actually shoot men on walls.
Back in 1.2, I reduced velocities for most rounds and set them to values that would only let them reach a smaller X % above their stated range. I created several new archery rounds with velocities to match. (This way I didn't foul up anything already used for stationary defenses, etc.) Going from memory I had long range archer rounds (composite bows), vanilla archer rounds, and an intermediate for mounted archers with composite bows. I'm going to dust this off and reapply it soon.
Alternately, you can also knock down the range of flaming arrows this way, just reduce the velocity of the flaming round (you can't change its range directly.)
Another fellow was using this trick to put in flaming pila as used by the Spaniards.
I was refering to STW. There are no ninjas, javelins or thunder bombers in STW. Foot archers, cav archers and muskets were range = 5000. The arquebusiers were 4000. All those new ranged units in STW/MI were problematic and had to be completely reworked in the v1.02 rebalance patch. The javelins would never fire because their range was shorter than their skirmish distance. I don't remember if the AI was able to handle thunderbombers or ninja. The xbows were a lot like guns without rotating ranks, and I think it handled them ok.Quote:
Originally Posted by Red Harvest
The 70 degree max angle is so the archers can choose a second trajectory if the low trajectory is blocked. Since the velocity is so high, 70 degrees is going place the arrow so far away I would guess it effectively makes the second trajectory useless for distances inside the normal targetting range. Are you sure archers lead their targets? I can't imagine why the high velocity is being used if not for increasing the chance of hitting moving targets.
There is no reduction a projectile's power with distance in the previous engine and I doubt that has changed in RTW, so there's no attenuation in that sense. Accuracy is a small random error added to a projectile's initial path, so the further away the target the more likely the projectile will miss. The flatter the trajectory, the more likely the projectile will hit a secondary target, and the cross section of the targets will also be greater as you say.
17. In some instances, the AI fails to use its "warcry." It seems to do so on defense, but not when it decides to attack. I had wondered if I was seeing this right in the campaign. In custom battle the difference was clear.
Example: Scutarii vs. British swordsmen. Fairly even match up. Try playing as the scutarii, toggle Brits to defense at times. When not on defense they advance without using warcry. On defense, they will actually use their warcry, as they stand waiting for you to attack. Unfortunately, they do a lot of other things wrong...like trying to withdraw from melee only to run in circles around your formation. (As either side this is a very easy win for the human 1vs1.)
I think 1.3 *might* be doing a better job of positioning its army altogether, and is less apt to commit suicide with the general, but when it comes to employing the units it actually has taken a step backward. :ahh: It doesn't seem to use key strengths of individual units consistently or well.
Ok, thought that might be the case, but didn't have the original in standalone install anywhere.Quote:
Originally Posted by Puzz3D
I've seen some really wild "bullets" whizzing by at super steep angles after those wardog handlers. ~D (I swear that once I thought my men were firing straight up as I was doing a "flyby.")Quote:
The 70 degree max angle is so the archers can choose a second trajectory if the low trajectory is blocked. Since the velocity is so high, 70 degrees is going place the arrow so far away I would guess it effectively makes the second trajectory useless for distances inside the normal targetting range.
Reasonably certain. I was checking a bit with my scutarii's pila tonight as the British swordsmen ran in a circle around my formation. ~:rolleyes: I went ahead and threw with the camera right behind my men and noticed that the pila were being launched with a few degrees of lead. Maybe I'm only fooling myself, but those 25 meters/sec pila appear to be leading the target.Quote:
Are you sure archers lead their targets? I can't imagine why the high velocity is being used if not for increasing the chance of hitting moving targets.
One thing I noticed about RTW early on was the "intercept course" approach that was being used for regular attack (which differed from previous TW lag intercept approach IIRC.) I think the same method is used for missile targeting vs. a moving target.
Is 1.3 RTW even worth playing. Reading all these issues discourages me.
In BI this seems also to be an issue right? I have been decimating hordes with just 4 units of archers covered by com's. Even horse archers don't seem to help.
I have a challenging RTW v1.3 Julii campaign (M/M difficulty) going in 142 BC. I made some mistakes managing my cities which has made it harder, and I might not be able to win now. I find the battles to be much better in v1.3 than in v1.2 where I played a Julii campaign out to 166 BC, but never finished it. My last battle in this RTW v1.3 campaign was with 1487 men under a 3 star general against 1404 Brutii under a captain. It lasted 15 to 20 minutes, and I was just able to win with each side suffering over 1200 casualties. I only had about 50 men on the field at the end. I never had any battles like that in RTW v1.2 except when using Mordred's Community mod, but the broken sieges when reloading and the suicide general killed RTW v1.2 for me. Overall, most of the battles in this campaign have been good with some exceptions. In my next campaign, I will play with fatigue off because the AI doesn't manage its fatigue and is therefore at a big disadvantage. I'll also have to play as Scipii because that faction's strategic AI is not working.
Tough question. I'm asking myself the same thing. There are some aspects that are better, but CA has broken quite a few bits as well. Another very mixed bag.Quote:
Originally Posted by Veresov
Some of the good stuff:
A number of the unit stats and misc. mistakes are corrected, command stars work better (you don't end up with max stars after 6 or 7 battles), and the AI does a better job of combining its forces. Traits seem to work better overall. (They were so badly bugged in 1.0, 1.1, and 1.2 that it would have been better had they not existed at all.) The save bug seems to be fixed. Coastline bug is gone. Speeches are more varied and interesting (wonder if fixing the traits had some impact there.)
Some of the bad stuff:
The memory leak is a killer. Perhaps it would be acceptable for a beta release before the game is published, but not up to snuff for anything else. If I wasn't running XP I would expect the machine to lock up during shut down. As it is, it takes several minutes just to get out of the game and shutdown after playing for a few hours. So RTW 1.3 fails on technical grounds.
The AI still has trouble expanding. It seems to have lost the ability to conduct naval invasions more than once or twice. (Carthage is much easier as a result.)
Britain and Egypt still are powerhouses. Carthage/Spain still get the short end of the stick infantry wise until fully upgraded.
The AI combines its forces, but still attacks/sieges with undersized forces far too often, when additional troops could be added.
AI can't use many of its units' abilities properly: pila, archers, slingers, javelins, even war cry at times. Ranged combat is about half broken.
Conclusion: Neither good nor bad, just different. Too badly bugged to really feel it is complete. There are some fundamental improvements, but new bugs take the luster off of this patch.
See this is strange for me.Quote:
Originally Posted by Red Harvest
I experience the leak once in a while, but I can leave the game for hours (or play for that matter) and it is still good when I return. So for me it is at worst a limited nuisance as I have yet to only experience a single battle becoming affected by this (in fact it CTDed).
How much RAM are you running with?