Matrix Games Forums

Question is quite simple isn't it. Did anyone start the GC in '41 and either lost Moscow and Leningrad as the Soviets or was pushed back into Germany as the Axis. Thus, anybody lost from the AI or came close to doing so?

In AIvAI testing I don't think an Axis decisive victory was achieved. The SU got a minor victory and was 3 hexes from Berlin in one of the last tests that was done before the game shipped. Settings were adjusted up and down from challenging during the course of the tests.

In human v AI games, I think you would have to play on Impossible settings to stand any chance of the AI getting a decisive victory. I think RedMarkus ran at impossible setings in one of his AARs and Moscow fell pretty quickly..

I restarted my very first GC when I realized Leningrad and Moscow were encircled and I could do nothing to avoid their capture.

I restarted my very first GC when I realized Leningrad and Moscow were encircled and I could do nothing to avoid their capture.

Was this before the blizzard? Insufficient forces to recapture during the winter?

ORIGINAL: BigAnorak
Was this before the blizzard? Insufficient forces to recapture during the winter?

Yes, it was before the blizzard. I have also to add that I didn't try too hard... I was more focused on learning the game mechanics than to devise an effective defense!

(LOL I just imagined the poor Pavlov saying such a thing to the NKVD! [:'(][:@][:-])

In my second game I won a decisive victory (normal AI).

Now I'll try with a different AI setup. But I do need Tank Armies! I hope the next patch will appear before summer '42! [;)]

I lost a '41 GC - Soviet versus hard Axis AI. That said, the challenging Axis AI got to within 30 miles of both Moscow & Leningrad and was 20 miles from Kharkov in Autumn '41, so I'm clearly not a top player .

ORIGINAL: amatteucci


In my second game I won a decisive victory (normal AI).

That's disappointing news. Not the fact that you won, but you won a "decisive" victory, on your second try. AI must be very bad, but that is not surprising considering no one ever finished a game before it was released. (that still amazes me). Such a "monster" game should not be "decisively" beaten in just 2 attempts. I know, "play a live person" and all that crap! The more I read the forum, the more this game loses its luster.

I don't blame the AI, as such - the AI fights a VERY effective battle. However, there are several design/coding/scripting issues that I find puzzling and I believe that these have a lot to do with why the AI can be beaten in 1943 by new players on their second or third attempt. I expect these will be fixed now that it's obviously an issue.

I am switching to PBEM to see how that goes and will probably post an AAR.

ORIGINAL: vonRocko

ORIGINAL: amatteucci


In my second game I won a decisive victory (normal AI).

... that is not surprising considering no one ever finished a game before it was released. (that still amazes me).

And yes, that amazes me too... [&:]

We have said that if you are playing against the German AI, the German AI in 1941-42 needs a bump up to give a fair fight. So a normal game against the German AI is like an easy game against the Soviet AI. It is not surprising that a decent player with some experience would be able to decisively beat the normal German AI. I don't think this means the AI is worthless, and in fact by upping the play levels you can find plenty of challenge and hours of entertainment.

ORIGINAL: redmarkus4


And yes, that amazes me too... [&:]

I'm amazed that you are amazed. Your expectations are impractical and would have added possibly years to the time of release.

We had to kill many player games in progress due to changes and tweaks in code to properly test the new code. During that testing, new changes would appear, and the process start over from scratch. Only the pure AI vs. AI games could be played from start to finish rapidly enough to keep up with this process. (And we did in fact run many such tests to full completion.)

Human beings just aren't as fast, especially human vs. human. Even in a solo game versus the AI the best I ever managed was 80 turns in a week's time...and that was only possible due to me being on vacation.

A player versus player game is lucky if they can get in one turn a day given all our other real life commitments: we aren't being paid for this. A full GC played over 200ish turns would take months to complete and somewhere along the line be hopelessly compromised along the way by code changes anyways and be mostly useless for testing purposes as a result.

In comparison to other monster games, and especially the boardgame predecessors of this game, WitE has been intensively tested. But there are still practical limitations on how much can be done and at some point you start running into diminishing returns with any kind of internal development team.

This team is absolutely committed to continued improvement of the game, and necessarily there's many things that show up once the pool of people playing the game expands.

ORIGINAL: vonRocko

ORIGINAL: amatteucci


In my second game I won a decisive victory (normal AI).

That's disappointing news. Not the fact that you won, but you won a "decisive" victory, on your second try. AI must be very bad...

Compared to what? I've never played an operational or strategic war game in my life that I couldn't crush on the first attempt against normal AI once I got the game mechanics down.

TOAW, Accross the Dnepr X, SC X, HOI X, etc. All easily destroyed on normal difficulty settings playing either side.

This is easily the best AI I've faced in a wargame. I actually had to restart a couple of the Road To X Scenarios as the Axis which almost never happens to me.

And I only expect it to get better. And there are 3 difficulty levels above normal.

It may not be Deep Blue but if you can name a tougher AI (in an operational war game), I'd like to hear it.

ORIGINAL: vonRocko

ORIGINAL: amatteucci


In my second game I won a decisive victory (normal AI).

That's disappointing news. Not the fact that you won, but you won a "decisive" victory, on your second try. AI must be very bad, but that is not surprising considering no one ever finished a game before it was released. (that still amazes me). Such a "monster" game should not be "decisively" beaten in just 2 attempts. I know, "play a live person" and all that crap! The more I read the forum, the more this game loses its luster.

As redmarkus4 said, I think it's not a problem of the AI per se but of some scripts that influence strategic decisions. The AI fight reasonably well on the tactical and operational scale (consider that after rushing to take Berlin as fast as possible, I had to complete the conquest of Germany/Czechoslovakia/Austria to score a win and it was a pain in the **** to do this in reasonable time (90% of my cursing was acted before December 1941 and after August 1943 [:D]). But while in WitP/AE you can script some operations and implement them practically the same way in 1942 or 1944 (the difference being the quantity and quality of aircrfatt/ships/troops committed), in WitE (I presume) you can do something like this only for the first stages of a given campaign/scenario.

Moreover, I don't know it it's actually possible to have a realistic WW2 strategic game in wich the Axis has a good chance of obtaining even a stalemate. I mean, after Barbarossa is halted and Germany didn't manage to kill/capture at least four or five million Soviet soldiers and seriosly disrupt USSR production, is there a way, even for a human player, to avoid defeat? And, even if it's the case, how is the AI supposed to devise a brilliant strategic plan (or even one not so brilliant but comparable to Fall Blau or Zitadelle)?
(see the scripting issues above)

Perhaps to give a fair judgment on the AI capabilities one should consider mainly the GC played vs the Soviet computer. Well, if now everyone playing as the Axis says that routinely reaches the Urals in 1942, that would be a problem. Suffering a defeat in 1943/44 after not being able to put Barbarossa to fruition would be simply too difficult for any AI. Well, it would be challenging even for an experienced human player, I presume.

ORIGINAL: Flaviusx

ORIGINAL: redmarkus4


And yes, that amazes me too... [&:]

I'm amazed that you are amazed. Your expectations are impractical and would have added possibly years to the time of release.

We had to kill many player games in progress due to changes and tweaks in code to properly test the new code. During that testing, new changes would appear, and the process start over from scratch. Only the pure AI vs. AI games could be played from start to finish rapidly enough to keep up with this process. (And we did in fact run many such tests to full completion.)

Human beings just aren't as fast, especially human vs. human. Even in a solo game versus the AI the best I ever managed was 80 turns in a week's time...and that was only possible due to me being on vacation.

A player versus player game is lucky if they can get in one turn a day given all our other real life commitments: we aren't being paid for this. A full GC played over 200ish turns would take months to complete and somewhere along the line be hopelessly compromised along the way by code changes anyways and be mostly useless for testing purposes as a result.

In comparison to other monster games, and especially the boardgame predecessors of this game, WitE has been intensively tested. But there are still practical limitations on how much can be done and at some point you start running into diminishing returns with any kind of internal development team.

This team is absolutely committed to continued improvement of the game, and necessarily there's many things that show up once the pool of people playing the game expands.

You're amazed that Redmarkus4 is 'amazed', but you're not amazed that vonRocko or anyone else is 'amazed'? LOL Very amusing

ORIGINAL: amatteucci

ORIGINAL: vonRocko

ORIGINAL: amatteucci


In my second game I won a decisive victory (normal AI).

That's disappointing news. Not the fact that you won, but you won a "decisive" victory, on your second try. AI must be very bad, but that is not surprising considering no one ever finished a game before it was released. (that still amazes me). Such a "monster" game should not be "decisively" beaten in just 2 attempts. I know, "play a live person" and all that crap! The more I read the forum, the more this game loses its luster.

As redmarkus4 said, I think it's not a problem of the AI per se but of some scripts that influence strategic decisions. The AI fight reasonably well on the tactical and operational scale (consider that after rushing to take Berlin as fast as possible, I had to complete the conquest of Germany/Czechoslovakia/Austria to score a win and it was a pain in the **** to do this in reasonable time (90% of my cursing was acted before December 1941 and after August 1943 [:D]). But while in WitP/AE you can script some operations and implement them practically the same way in 1942 or 1944 (the difference being the quantity and quality of aircrfatt/ships/troops committed), in WitE (I presume) you can do something like this only for the first stages of a given campaign/scenario.

Moreover, I don't know it it's actually possible to have a realistic WW2 strategic game in wich the Axis has a good chance of obtaining even a stalemate. I mean, after Barbarossa is halted and Germany didn't manage to kill/capture at least four or five million Soviet soldiers and seriosly disrupt USSR production, is there a way, even for a human player, to avoid defeat? And, even if it's the case, how is the AI supposed to devise a brilliant strategic plan (or even one not so brilliant but comparable to Fall Blau or Zitadelle)?
(see the scripting issues above)

Perhaps to give a fair judgment on the AI capabilities one should consider mainly the GC played vs the Soviet computer. Well, if now everyone playing as the Axis says that routinely reaches the Urals in 1942, that would be a problem. Suffering a defeat in 1943/44 after not being able to put Barbarossa to fruition would be simply too difficult for any AI. Well, it would be challenging even for an experienced human player, I presume.

I agree completely - on the tactical and operational scale the AI is very good indeed. Although it's true that there will probably never be a perfect strategic AI, there are still some fundamental things that can be fixed to make it better:

- We have already seen the 'running away in the south' bug fixed. (Flavi-what's it said it wasn't a bug and then Gary said he'd solved it LOL)

- Next we need to see the AI holding onto cities more effectively - the manual says that they are used to calculate VPs, right? Should be important to capture and hold them then.

- Finally, we need to see the AI make some random (and stupid) offensives, to add the human error that is an essential part of military conflict. At the moment the AI plays it too safe strategically, IMO.

And then there's the air war, of course, and the AI's clustering of huge parks of airfields all in one place. It can't take a magic trick to fix that!

von Rocko isn't playing the game, Red.

You are. You should have some practical appreciation here of what a beast it is and how time consuming a GC will be anytime humans are introduced into the equation. I'm trying to give you some insight here on the development process. So, yes, I'm very irritated by your flippant remarks about lack of testing. Actually, you and I simply have never hit it off from the getgo. Possibly we should agree to ignore one another.

And to clarify: your game didn't really demonstrate conclusively there was a bug in the Ukraine. Another one did. We followed up on that and made adjustments.

Thanks for all the explanations guys. I realize an ai will be beat, but I wondered about getting a "Decisive" victory on just the second try. Are the mechanics of the game easy enough to master after one game? I guess Joel explained it best to me by saying that the german ai on normal, is more like easy. No offense to Ammateuci who, i'm sure is a skilled player![;)]

ORIGINAL: redmarkus4

- Finally, we need to see the AI make some random (and stupid) offensives, to add the human error that is an essential part of military conflict. At the moment the AI plays it too safe strategically, IMO.

That's a splendid suggestion.

People seriously want to play against an AI deliberately programmed to do dumb things?

I'm gobsmacked by this. Do we also make Hitler standfast rules a la Stalingrad?

I would get no pleasure at all from playing against such an AI.

In the end, it looks like there is no pleasing some people. No matter what we do here, somebody is going to be unhappy.

ORIGINAL: vonRocko

Thanks for all the explanations guys. I realize an ai will be beat, but I wondered about getting a "Decisive" victory on just the second try. Are the mechanics of the game easy enough to master after one game? I guess Joel explained it best to me by saying that the german ai on normal, is more like easy. No offense to Ammateuci who, i'm sure is a skilled player![;)]

Well, consider that (IMHO) "decisive" is just a label. I mean the only difference between minor, major and decisive (for what concernes Soviet victory) is only a question of time. Maybe it's realistic, because of the actual concerns of the Soviet leadership (as soon as it's clear that Germany is going to loose the war).
But, speaking from the point of player's difficulty, I presume that a "minor" victory with 20% less casualties would have been much more difficult to achieve than my "decisive" victory with more than 6 millions of military irrecoverable losses and an overall AFV kill ratio of 4:1+ in favour of the Axis.

In addition, I don't think that, in a wargame, the point is simply (or mainly) mastering the game mechanics.
A simulation would be a very poor rimulation if learning the tricks-of-the-trade is more important than sound strategical thinking and historical knowledge. Actually, I think that the fact that WitE was relatively easy to dive in, is a merit of the game and not a flaw (although I admit I'm still not knowledgeable about all its nuances). A game in which I experience a difficulty because I know exactly what has to be done but I don't know how it's done in the game (or the game don't allow me to do it) would be a very challenging game but not a good simulation.

I'm happy with my recent campaign because, although in some respect it was easier than I presume, it forced me to think in historical and realistic terms and not (only) in game terms. I had the occasion to apply all I learned in these years reading upon the eastern front and Soviet doctrinal evolution. Historical playing is rewarded. Something I dislike in a wargame is the possibility to come up with some gamey/munchkin combo that outperforms by far everything that was actually done in practice just because the rules are bugged, or the possibility for a player to apply 20/20 hindsight not only in strategical decisions but also in tactics and doctrinal organization. Well, in WitE it seems this is not possible (in other games is!) and this is very important for me and my enjoyment of the game.