Over many years, I have had many tough, challenging games against Athena. After reading about some of the horrendous AI problems experienced by others, I realized that I rarely run into them, the main exception to this being the Stack of Doom. I looked at how I play, and I thought it might be helpful to share some things.
I use the AI settings recommended by the developers, and I usually give the AI more time.
The most important thing I do is this. I act as if I am facing a human being, and only make moves that make sense according to the situation and what armies actually did during that period.
The second most important thing is that I only give orders that make sense according to the military realities of the time, protecting key areas, trying to stay in supply, etc.
I do not study Athena to see what her weak points are, I just try to play the strategic situation as if I was facing a human general. for example, I protect every key point that might be vulnerable to a human general, without even trying to find out if Athena might ignore it.
I do not engage in exploits that will drive her crazy, as they involve maneuvers that no real general would ever make.
The stacks of doom are still a problem, but not in every game, and I have hope that the solution proposed by Kensai will be adopted.
My final point is this - If you treat Athena with respect, and pretend she is a real general, she will give you a challenging game where most of what she does makes sense. And while I usually win wars, I have suffered many defeats, and lost some wars and games to her.
Just my two cents, but I hope it helps someone enjoy these games as much as I do.
As I've seen some of that odd ai behviour I'll add a quick reply :-) ...
I also always try not to exploit known weaknesses of the ai. One problem I might have encountered is that I thought upping some ai behaviour would make for a more challenging game (I did not even notice there were recommended settings until after my latest game ;-) ). That also means I always give added processor time to the ai, one problem was probably giving it better recon...
One area which might indeed play a role is that I probably play both more aggressively and less cautious than I would against a human player. Two examples for that would be my move into Belgium as CP (though I did it on a wide enough front that even had there been some unlocked units to move against me I would have swept them away easily) (one stack ending the first turn at Lille, another that should have gotten into France failed it's force march). Were that not the first turn I would certainly not have made such a bold (deep penetration) move against a human player, even being the first turn I would probably have planned for a stronger left flank to counter unexpected WE moves in turn two. Against a human player I would also most likely have evacuated Silesia an co. after a strong Russian stack had moved deep behind my lines cutting off two of my three rail lines. Against the ai I decided to pull my front back into the mountains only and draw some forces from the front as well as reserves from the rear to encircle and destroy the stack that had peirced my lines. Then again I'm not sure my actions in Austro-Hungary adversely affected the ai (it lost a 500+ power stack, but occupied the more important Silesian cities, was otherwise inactive), certainly far less than my destroying the two ai armies in East Prussia (in a rather historic fashion with countermarches).
The most devastating effect to the ai were probably it's repeated assaults against Metz (I at first considered it akin to the historic French August campaign in Alsace Lorraine combined with a Verdun effect, but in the end the bleeding of the french army was much worse than anything a historic general would have tried even in WWI.
And I really can't take any blame for odd ai fleet deployments. My High Seas Fleet was only in very limited operations (7 day patrol up to the German Bight and no further), the East Sea Fleet played no role as the Russians never entered my patrol areas and the few raiders I had in the Atlantic posed no threat to the ai, to the contrary they should have spurred the ai to send forces there and not elsewhere (my combined Asiatic and African squadrons were destroyed on turn 3 when I made a sucessful sortie against a British squadron only to be caught on the way back to an African port). Yet I found that were was no actual WE naval force in the North Sea and odd concentrations off Britanny and in Asia...
But otherwise I largely agree with you. There's no point in knowingly exploiting Athena's weaknesses. Players also have give the ai settings that will enable it to work correctly (just that what the best settings are for EAW doesn't seem to be clear cut yet). I also know the existing ai issues will be solved eventually. But I must say I'm not enjoying the game currently with the results I had in three games against the ai (once WE, twice CP)...
I also always try not to exploit known weaknesses of the ai. One problem I might have encountered is that I thought upping some ai behaviour would make for a more challenging game (I did not even notice there were recommended settings until after my latest game ;-) ). That also means I always give added processor time to the ai, one problem was probably giving it better recon...
One area which might indeed play a role is that I probably play both more aggressively and less cautious than I would against a human player. Two examples for that would be my move into Belgium as CP (though I did it on a wide enough front that even had there been some unlocked units to move against me I would have swept them away easily) (one stack ending the first turn at Lille, another that should have gotten into France failed it's force march). Were that not the first turn I would certainly not have made such a bold (deep penetration) move against a human player, even being the first turn I would probably have planned for a stronger left flank to counter unexpected WE moves in turn two. Against a human player I would also most likely have evacuated Silesia an co. after a strong Russian stack had moved deep behind my lines cutting off two of my three rail lines. Against the ai I decided to pull my front back into the mountains only and draw some forces from the front as well as reserves from the rear to encircle and destroy the stack that had peirced my lines. Then again I'm not sure my actions in Austro-Hungary adversely affected the ai (it lost a 500+ power stack, but occupied the more important Silesian cities, was otherwise inactive), certainly far less than my destroying the two ai armies in East Prussia (in a rather historic fashion with countermarches).
The most devastating effect to the ai were probably it's repeated assaults against Metz (I at first considered it akin to the historic French August campaign in Alsace Lorraine combined with a Verdun effect, but in the end the bleeding of the french army was much worse than anything a historic general would have tried even in WWI.
And I really can't take any blame for odd ai fleet deployments. My High Seas Fleet was only in very limited operations (7 day patrol up to the German Bight and no further), the East Sea Fleet played no role as the Russians never entered my patrol areas and the few raiders I had in the Atlantic posed no threat to the ai, to the contrary they should have spurred the ai to send forces there and not elsewhere (my combined Asiatic and African squadrons were destroyed on turn 3 when I made a sucessful sortie against a British squadron only to be caught on the way back to an African port). Yet I found that were was no actual WE naval force in the North Sea and odd concentrations off Britanny and in Asia...
But otherwise I largely agree with you. There's no point in knowingly exploiting Athena's weaknesses. Players also have give the ai settings that will enable it to work correctly (just that what the best settings are for EAW doesn't seem to be clear cut yet). I also know the existing ai issues will be solved eventually. But I must say I'm not enjoying the game currently with the results I had in three games against the ai (once WE, twice CP)...
Marc aka Caran...
I concur with these, as well. It's almost always easy to "game" an AI, but then I never considered that fun.
Extending a bit on what you two have said already, I also do not attack with forces that are weakened (morale significantly reduced), as this is something a general would not typically do.
Interesting thing about the Metz note: the French did this very thing (attacking in that general area over and over and over) during the early and middle stages of the war. The French "cult of the offensive" ran amok in this war, much to their own undoing.
Extending a bit on what you two have said already, I also do not attack with forces that are weakened (morale significantly reduced), as this is something a general would not typically do.
Interesting thing about the Metz note: the French did this very thing (attacking in that general area over and over and over) during the early and middle stages of the war. The French "cult of the offensive" ran amok in this war, much to their own undoing.
A very good topic, vaalen. Well said! 
I have only one advance myself: play as you would have played ιf it was a real war, with real humans, fronts, regions. Would you do crazy stuff and take unnecessary risks there? Chances are that you wouldn't. Then don't in your game.
I have only one advance myself: play as you would have played ιf it was a real war, with real humans, fronts, regions. Would you do crazy stuff and take unnecessary risks there? Chances are that you wouldn't. Then don't in your game.
Care to unify Germany as Austria? Recreate the Holy Roman Empire of the 20th Century:
Großdeutschland Mod
Are you tough enough to impersonate the Shogun and defy the Westerners? Prove it:
Shogun Defiance Mod (completed AAR)
Großdeutschland Mod
Are you tough enough to impersonate the Shogun and defy the Westerners? Prove it:
Shogun Defiance Mod (completed AAR)
Kensai wrote:A very good topic, vaalen. Well said!
I have only one advance myself: play as you would have played ιf it was a real war, with real humans, fronts, regions. Would you do crazy stuff and take unnecessary risks there? Chances are that you wouldn't. Then don't in your game.
Kensai, that is excellent advice, and I will happily take it.
-
TrenchFoot
- Private
- Posts: 22
- Joined: Tue Aug 12, 2014 4:50 pm
Yes, agreed generally on the points made. However I would quibble on this point:
Surely studying the weak points of your enemy generals is actually one of the things any good general would do. The Battle of Tannanberg is a good example of this, the Germans knowing the personal animosity of the two generals certainly was a positive factor in their plans. So exploiting a weakness in your enemy actually makes sense whether it is an AI or human. However the difference between the two (for now, maybe not in commercial games in 10-20 years time) is that humans can learn whereas the AI cannot (not organically in game, though certainly can be improved in patches). In that respect you have to fight with one hand behind your back.
So the best way is to treat the AI (Athena) as a child, playing a fair game, without taking advantage of her inexperience or weak points. Then point out how she could improve her tactics & strategy as you play along with her.
vaalen wrote:I do not study Athena to see what her weak points are, I just try to play the strategic situation as if I was facing a human general. for example, I protect every key point that might be vulnerable to a human general, without even trying to find out if Athena might ignore it.
Surely studying the weak points of your enemy generals is actually one of the things any good general would do. The Battle of Tannanberg is a good example of this, the Germans knowing the personal animosity of the two generals certainly was a positive factor in their plans. So exploiting a weakness in your enemy actually makes sense whether it is an AI or human. However the difference between the two (for now, maybe not in commercial games in 10-20 years time) is that humans can learn whereas the AI cannot (not organically in game, though certainly can be improved in patches). In that respect you have to fight with one hand behind your back.
So the best way is to treat the AI (Athena) as a child, playing a fair game, without taking advantage of her inexperience or weak points. Then point out how she could improve her tactics & strategy as you play along with her.
Nico - Icon
'From without a thousand cycles
A thousand cycles to come
A thousand times to win
A thousand ways to run the world'
- Nico, 'Frozen Warnings'
'From without a thousand cycles
A thousand cycles to come
A thousand times to win
A thousand ways to run the world'
- Nico, 'Frozen Warnings'
PJL wrote:Yes, agreed generally on the points made. However I would quibble on this point:
Surely studying the weak points of your enemy generals is actually one of the things any good general would do. The Battle of Tannanberg is a good example of this, the Germans knowing the personal animosity of the two generals certainly was a positive factor in their plans. So exploiting a weakness in your enemy actually makes sense whether it is an AI or human. However the difference between the two (for now, maybe not in commercial games in 10-20 years time) is that humans can learn whereas the AI cannot (not organically in game, though certainly can be improved in patches). In that respect you have to fight with one hand behind your back.
So the best way is to treat the AI (Athena) as a child, playing a fair game, without taking advantage of her inexperience or weak points. Then point out how she could improve her tactics & strategy as you play along with her.
MAX Hoffman overstated some of the 'animosity' issues between Von Rennenkampf and Samsonov; but the victory came due to some other things like-
Max Hoffman himself was one of the key factors, one of the most outstanding staff officers ever.
The Prussian/German Railways being the benchmark for efficiency transported tens of thousands of troops to their points in a short time
Staff work at divisional and corps level was also very important, as Mackensen's corps had badly lost at Gumbinnen but won strongly at Tannenberg.
Aerial Reconnaissance helped as also cracking of the Russian Code (Russians were stupid here and used a variant of MORSE code - easy to crack)
TrenchFoot wrote:AWESOME POST!
By the way what are the current recommended settings for the AI mentioned in the first post??
Thanks, Trenchfoot! If you go to the options section, you can go to a screen where you can set the AI. There are recommended settings, that you will find if you mouse over the various potential settings. In EAW, there is also a default icon you can click on, located in the lower right corner of the screen, I believe.
PJL wrote:Yes, agreed generally on the points made. However I would quibble on this point:
Surely studying the weak points of your enemy generals is actually one of the things any good general would do. The Battle of Tannanberg is a good example of this, the Germans knowing the personal animosity of the two generals certainly was a positive factor in their plans. So exploiting a weakness in your enemy actually makes sense whether it is an AI or human. However the difference between the two (for now, maybe not in commercial games in 10-20 years time) is that humans can learn whereas the AI cannot (not organically in game, though certainly can be improved in patches). In that respect you have to fight with one hand behind your back.
So the best way is to treat the AI (Athena) as a child, playing a fair game, without taking advantage of her inexperience or weak points. Then point out how she could improve her tactics & strategy as you play along with her.
PJJ, I was referring more to finding the weakness of the AI to gamey tactics that no real general can use. It is possible to drive Athena crazy, by undertaking movements that no real general would ever use.
That is different from noticing the type of things that Athena does that are what real generals will do. For example, in a PON campaign as France, against Tonkin, I noticed that the Athena would cut my supply lines, so I heavily protected them in response. I also learned that if I tried to move into an area, and the move would take more than one turn, due to bad terrain and weather, Athena would attack my troops near the end of the march, when my troops had very low cohesion, and defeat them decisively. So I adopted my tactics to wait for good weather, so I could complete the move in one turn, which was far less damaging to cohesion.
Or, in the first civil war game, Athena would raid behind my lines and cut off my supply, very effectively, so I learned from the forum how to protect against this, after some of the things I tried failed. This does not drive Athena crazy.
The difference is difficult to explain, but the main point I was trying to make was to not look for the weaknesses Athena has as an AI, rather than a general.
vaalen wrote:It is possible to drive Athena crazy, by undertaking movements that no real general would ever use.
We should add to that what we learned from Hubu's game. Abandoning entire theaters/fronts is a sure way to break Athena's planning. Athena will seek the opportunity to push in the weak front, potentially taking troops away from where important defenses are against the main thrust of the human player. The AI cannot understand the human's intentions of trying to troll it... I consider this kind of playing self-cheating, apart from totally unhistorical*.
*obviously, there ARE some instances in real life that politicians and generals decided in abandoning certain theaters: in WW1 Hindenburg's predecessor wanted to abandon East Prussia when he became aware of the unexpectedly swift mobilizing of the Russian Imperial Army; nonetheless, if the player abandons something, he has to be certain it is not blatantly unhistorical and inexplicable (like when he is winning and/or can totally defend it!)
Care to unify Germany as Austria? Recreate the Holy Roman Empire of the 20th Century:
Großdeutschland Mod
Are you tough enough to impersonate the Shogun and defy the Westerners? Prove it:
Shogun Defiance Mod (completed AAR)
Großdeutschland Mod
Are you tough enough to impersonate the Shogun and defy the Westerners? Prove it:
Shogun Defiance Mod (completed AAR)
-
Ebbingford
- Posts: 6162
- Joined: Sun Oct 14, 2007 5:22 pm
- Location: England
I think the difficulty level that you play on might also contribute to how "good" a game that you have. On any thing over sergeant level the AI gets bonuses on several things, movement and command penalties for stacking being two of them. I would recommend playing on sergeant level. 
"Umbrellas will not be opened in the presence of the enemy." Duke of Wellington before the Battle of Waterloo, 1815.
"Top hats will not be worn in the Eighth Army" Field-Marshal Viscount Montgomery of Alamein K.G.
"Top hats will not be worn in the Eighth Army" Field-Marshal Viscount Montgomery of Alamein K.G.
Well Ataris has an interesting MP beta MOD which allows you to play as a single entity and changes a lot of the ways the game works, such as the Germans having to bombard the Belgian forts and take them to get further, rather than what happens which is the germans just skip past them and as above make it to Lille in one turn.
However I did notice a queer thing in my game I started last night as the CP, I went Moltke/Conrad and the Russians went all in Austria and French went Defensive. First the French have not doom stacked Metz, yet, however they did push from Nancy to in front of my lines in force, then one unit moved into my lines. I lost the battle due to MTSG, but caused more losses to them. Then the supporting French Troops marched back and the unit in my lines continued to try to march down my lines.
Secondly Silesia is fairly open to the Russians, but the first troops to arrive at Posen assualted and all 23k Russians apparently ran directly at the troops inside the fort and were wiped out to a man. Which is quite crazy.
However I did notice a queer thing in my game I started last night as the CP, I went Moltke/Conrad and the Russians went all in Austria and French went Defensive. First the French have not doom stacked Metz, yet, however they did push from Nancy to in front of my lines in force, then one unit moved into my lines. I lost the battle due to MTSG, but caused more losses to them. Then the supporting French Troops marched back and the unit in my lines continued to try to march down my lines.
Secondly Silesia is fairly open to the Russians, but the first troops to arrive at Posen assualted and all 23k Russians apparently ran directly at the troops inside the fort and were wiped out to a man. Which is quite crazy.
Who is online
Users browsing this forum: No registered users and 0 guests