Axis Struggles

[Erik Rutins]

I'll say that the campaign has been tested many times before release and I expect as time goes on you'll find that it's possible for the Axis to exceed its historical performance even with equal player skill, just comes down to the choices each player makes. I'm assuming both players have learned how to play their side though and from what I've seen there's still a lot of climbing the learning curve going on. However, the historical challenges are all there, which tend to lead to some of the same decisions that historical commanders had to make. If that's what you're saying feels "scripted" then I'd suggest it's more like "realistic" as there is no script for the player (or the AI after Turn 1).

1.07 is no longer beta, it became an official update. As far as I know, it did not make any changes to combat resolution.

[GloriousRuse]

In one of the later beta waves where I came in we witnessed a certain pattern.

At first the axis do poorly. People are learning that while their combat formations are powerful, they aren’t the invincible legions of victory that so many east front games portray them as. The Soviets aren’t playing any better, but the onus is on the axis at first to deliver the big wins so they pay more heavily for being unfamiliar with the increased realism.

Then people start learning to pay attention to the nuance in the engine - roads, logistics, rest and refit, operational pauses, taking more efficient fights rather than fighting everywhere and all the time, understanding that some terrain can’t support offensives while other corridors are high speed approaches. The Axis players typically learn faster by Darwinian default, and soon everybody is nodding that the HWMs are pretty easy to make, and that it’s a fight.

As that proficiency starts really growing, the inherent German advantages - dynamism, mobility, what boils down to a faster OODA loop for everything but the employment of reserves, big explosive potential - really start seeing the Germans impose their will on the Soviets, even between good players.

I like to think of the two sides as a bit of a spectrum of return on capability, with the Soviets having a more centered and narrow curve with tighter limits - easier to play in the beginning, harder to get every last drop of as people get better - while the Germans have a much wider distribution, easier to mess up tragically at first, but also with a very high upper limit for what an experienced player can do.

[jlbhung]

The German ground and air forces were the first class in the World at that time. In the game, the players are assuming the role of strategic planning and command for the German military. If the AI performance is acceptable and the game is balanced, then I think it would be normal for the average players to perform poorer than the history in their first few attempts. At least for me, I don’t think I have higher IQ, wisdom, training and experience compared to the German military planners in 1941.

However, this is not to say that the average players cannot perform better than history. Players will eventually benefits from knowing the history, from the fact that the game’s AI is less competent than a human, from experience gained from playing the game several times as well as from experience learnt from fellow players.

For me, it is in fact a good news that the game is challenging when playing against AI. It is also very good that the game allows the players to change the difficulty settings in the middle of a campaign, so that players could fine-tune the settings to a level that they can manage comfortably.

[carlkay58]

It is also easier to gain wide experience as the Axis versus the Soviets. An Axis player can get well based experience in both tactics and strategy playing against the AI. The AI does very well on defense and does a good job putting up a fight and challenge for the Axis. The Axis AI, on the other hand, can cause the same level of casualties to the Soviets as a human player, but it is not as dynamic or fluid in deep penetrations as a human. So it is much harder for a Soviet player to gain more than tactical experience playing against the AI than a human. Also the game has to go much longer to see if the Soviet's play is any good while the Axis can evaluate the first 16 turns in many cases.

[Zemke]

ORIGINAL: Erik Rutins

I'll say that the campaign has been tested many times before release and I expect as time goes on you'll find that it's possible for the Axis to exceed its historical performance even with equal player skill, just comes down to the choices each player makes. I'm assuming both players have learned how to play their side though and from what I've seen there's still a lot of climbing the learning curve going on. However, the historical challenges are all there, which tend to lead to some of the same decisions that historical commanders had to make. If that's what you're saying feels "scripted" then I'd suggest it's more like "realistic" as there is no script for the player (or the AI after Turn 1).

1.07 is no longer beta, it became an official update. As far as I know, it did not make any changes to combat resolution.

In reference to the above and the following posts, has the game even been tested between two players trying to do exactly what happened historically? In other words, match the same moves and attacks as happened in the war. Perhaps this is a bit much for the entire operation, but at least on one front like the drive to Leningrad. And if so, did the casualties figures match what happened, did the advance rates match what happened? Because the one and only true measure we have is history, and to make a game based on history, that bills itself as the most accurate, the "game" or "simulation" should roughly match what happened historically. Also, I would say this should be done across the board for all scenarios, do advance rates match history, casualty rates, prisoners were taken, and so on. If this has not been done, then I feel like we the public are the testers, trying to see if the programming matches. Because frankly, I feel like this is "adventure" learning" so far, even with the manual and charts as aids. This is a simple question, has the game been tested using history as the benchmark?

I find it very difficult (impossible) to execute a historical maneuver, and I will use the following example. The 3rd Panzer Group was sent north to aid Army Group North in surrounding Leningrad, after Smolensk falls on 17 July 41. To get the 39th Pz Corps in position to attack and take Novogorod by 25 August 41 rested and with some CCP is well difficult while at the same time maintaining the gains made around Smolensk. The same would apply to the 57 Pz Corps attacking towards Demyansk, which fell on 8 Sept 41. The combination of moving, resting, CCP accumulation, and last the terrible terrain in both axis of advance all combine in this game to make this (for me) an impossible move to replicate. This is why I ask the above question, has the game been tested in this way to see if the programming matches the historical results?

Last I am on my fourth game, to try and replicate this very maneuver in the game, trying to stick as close to historical moves and routes as possible.

Also, let me add, I am loving the game, just trying to learn it, get better and understand what I am missing.

[loki100]

what you are asking for is pretty much impossible to be honest.

Did a lot of the testers stick pretty close to the historic German plan in 1941 - yes. Early on this produced some pretty wild outcomes as the game came together, over time we started to see relatively predictable variations around the historical outcome. Metrics in this respect mostly being front line location, army size and loss rates.

Remember the nature of beta (& even more alpha) testing - weekly or more ,exes, every now and then save breaking.

I had one HtH, we reran the first winter with a different exe to test that, we found a huge problem with level 1 depots giving railyards that meant we reran the 1942 summer offensive (I had week after week of 45-50 MP pzrs). In the end the game was abandoned in Sept 42 when the AOG functionality came into the .exe (save breaking).

We had to rely on AI-AI and then individuals pulling apart the mid-late game saves for most testing. We had a few completed vs AI games that came out reasonably historical. I actually had two complete games as the Soviets, one with a win in May 45, the other in March 45 (the AAR that has been posted). There was another in late testing that ended in late 44.

We also substantively used StB and VtB to test the mid/late game and in particular if the big set piece Soviet offensive could be modelled.

So yes, tested with a close eye on the historical record. Tested by two players emulating every historical move? No, of course not, first battle is going to set a new game on a unique track. Tested by players sticking fairly close to the historical choices - yes; tested by players constructing their own strategy - yes of course

[Karri]

ORIGINAL: Zemke

In reference to the above and the following posts, has the game even been tested between two players trying to do exactly what happened historically? In other words, match the same moves and attacks as happened in the war. Perhaps this is a bit much for the entire operation, but at least on one front like the drive to Leningrad. And if so, did the casualties figures match what happened, did the advance rates match what happened? Because the one and only true measure we have is history, and to make a game based on history, that bills itself as the most accurate, the "game" or "simulation" should roughly match what happened historically. Also, I would say this should be done across the board for all scenarios, do advance rates match history, casualty rates, prisoners were taken, and so on. If this has not been done, then I feel like we the public are the testers, trying to see if the programming matches. Because frankly, I feel like this is "adventure" learning" so far, even with the manual and charts as aids. This is a simple question, has the game been tested using history as the benchmark?

The problem is that the game is never going to play historically. No Axis player would let the 6th army be destroyed, and no Soviet player would allow the massive encirclements. Thus the aim is not to re-create history as it happened, but to create results that mimick history. So things will play out so that the Axis player should be able to reach the outskirts of Leningrad, Moscow and Rostov.

[Zemke]

Well, it cannot be a real historical game if it cannot match what happened in history. I am not talking about what real players will or will not do. I am asking IF the game has been tested this way, and do advance rates match as well as the other factors like casualties.

[loki100]

ORIGINAL: Zemke

Well, it cannot be a real historical game if it cannot match what happened in history. I am not talking about what real players will or will not do. I am asking IF the game has been tested this way, and do advance rates match as well as the other factors like casualties.

for the metrics we used - yes its a good match, above I told you the ones we mostly relied on.

Has anyone sat with a divisional history and modelled every move exactly as it happened - not as far as I am aware.

But I don't think that is a valid criticism of the game?

[Joel Billings]

We had some testers decide that going for Leningrad wasn't worth it. Only to have the next AAR have the German player take Leningrad. There is no doubt that the game system makes advancing in clear terrain much easier than advancing in bad terrain. It's very easy to be tempted into heading south because of this. But so much of the game, as in the war, is about doing the unexpected and making your main drive where the enemy isn't strong and doesn't expect it (see Battle of the Bulge for a West Front example of both surprise allowing for big advances, and importance of terrain in countering them). I'm sure we will learn a lot about the game as we watch AARs and get to see more games. So far we feel good about the ability of players to make a difference, but within historical constraints, but we'll keep an eye on this and are open to adjusting in the future.

[Speedysteve]

Also to add my 2p. From the times I’ve been involved in testing games before there’d be so many things to test - accuracy, data, glitches, bugs, trying to stress test added to the fact or frequent updates (I know Loki alludes to this above) that to have the time to plot every single realistic move week by week would, frankly, be impossible to achieve. The amount of time it would take would be inordinate and not possible bearing in mind the testing factors I mentioned.

[Zemke]

ORIGINAL: loki100

ORIGINAL: Zemke

Well, it cannot be a real historical game if it cannot match what happened in history. I am not talking about what real players will or will not do. I am asking IF the game has been tested this way, and do advance rates match as well as the other factors like casualties.

for the metrics we used - yes its a good match, above I told you the ones we mostly relied on.

Has anyone sat with a divisional history and modelled every move exactly as it happened - not as far as I am aware.

But I don't think that is a valid criticism of the game?

I am not interested in what two different players could or could not do against each other. To me, that really does not test anything unless the baseline test of a historical cross-reference of historical advance rates combined with casualties and prisoners taken. In other words, you have designed a program to do what? Is it a historically accurate simulation or an approximate facsimile of history? The answer is the latter, as the game has not been tested using a historical baseline as the measure. Which hey I get, if there was not the time to do the research, I do understand, that is a lot of historical research, which is why I suggested it should be done on a small area to even see how close the game is. As I stated previously, I have NOT been able to replicate the shifting of 3rd Pz Grp after taking Smolensk, mostly due to the time it takes to gain back enough CCP for those units to be effective in that terrain.

Also, I am saying this DOES NOT need to be done for every single unit or division, only a small area to see how the game stacks up. I think this could be done in a reasonable time frame.

[MechFO]

ORIGINAL: Zemke

Well, it cannot be a real historical game if it cannot match what happened in history. I am not talking about what real players will or will not do. I am asking IF the game has been tested this way, and do advance rates match as well as the other factors like casualties.

That's a very deterministic view of warfare. Given repeated runs of even a division scale fight, so much depends on the outcome of the aggregation of individual decisions and circumstances that's anything past the broadstrokes are impossible to predict weeks and months out and pointless to try and replicate. Friction will mess things up in different ways each time.

Have some fights take a bit longer, some events happen slightly differently, and the "time plan" is obsolete.

[carlkay58]

Zemke - the problems you are encountering in your attempt to duplicate the historical shifts of the Panzer Corps are to be expected. Why? Because I have yet to see any human play the Soviets and do out right attacks starting on turn 1 and continuing to attack with just about every unit within a few hexes of the front. Even the AI is too smart to do that. When the Battle of Smolensk started around July 10, the Soviets had five armies that stretched from Polotsk to Mogilev with no troops from Polostk to Idrista or Mogilev to Gomel (as per Glantz in his Battle for Smolensk). What Soviet player would do that AND attack with every unit possible? The Soviets lost 6M men from June 22 to December 31, 1941. This is according to three different experts (Glantz, Sharp, and Nasfinzer something that I can't remember off the top of my head). If the Soviets lose 6M men in this game, the Soviets will have roughly 1.8M men at December 31. This is quite possible as all three experts agree that the Soviets lost over 1M men just in December while attacking.

What you are asking is just too hard to test. What was extensively tested was the strategies and the outcomes. They came out pretty close to historical in territory and numbers.

[Joel Billings]

You could try playing Road to Leningrad in a historical way. However, even there the exact reinforcements for both sides are just the scenario designers best guess as to what was available. We can get unit names, but once you drill down to exactly when each man and gun showed up on the front, you're just approximating what actually happened. We do our best to get it as correct as reasonably possible.

[Bamilus]

ORIGINAL: carlkay58

Zemke - the problems you are encountering in your attempt to duplicate the historical shifts of the Panzer Corps are to be expected. Why? Because I have yet to see any human play the Soviets and do out right attacks starting on turn 1 and continuing to attack with just about every unit within a few hexes of the front. Even the AI is too smart to do that. When the Battle of Smolensk started around July 10, the Soviets had five armies that stretched from Polotsk to Mogilev with no troops from Polostk to Idrista or Mogilev to Gomel (as per Glantz in his Battle for Smolensk). What Soviet player would do that AND attack with every unit possible? The Soviets lost 6M men from June 22 to December 31, 1941. This is according to three different experts (Glantz, Sharp, and Nasfinzer something that I can't remember off the top of my head). If the Soviets lose 6M men in this game, the Soviets will have roughly 1.8M men at December 31. This is quite possible as all three experts agree that the Soviets lost over 1M men just in December while attacking.

What you are asking is just too hard to test. What was extensively tested was the strategies and the outcomes. They came out pretty close to historical in territory and numbers.

To be fair I'd probably be dumb enough to do that in my first game as Soviets (which I haven't tried yet) [:D]

[Zemke]

This idea of testing could even be done on a smaller scale, of a few divisions, and do several runs to see what the average results are, and answer some basic questions. Do the combat results match, can units match the historical advance rate, and so on. I think the computer program can handle several runs to get an average and let's see what that average is, not impossible at all.

I would be interested to know if anyone has been able to shift 3rd Pz Group to the North and match the historical attacks.

Anyway, I can see I am swimming upstream on this one.

[MakeeLearn]

ORIGINAL: carlkay58
...
Why? Because I have yet to see any human play the Soviets and do out right attacks starting on turn 1 and continuing to attack with just about every unit within a few hexes of the front. Even the AI is too smart to do that.
...

The "spoiling attacks", potential defender's loss of MP and CPP, would seem to make this a rewarding strategy. I play as Germans so I have not done this.

[Zemke]

How about we contact David Glantz (COL) retired and ask him to test this for a small fee...lol.

[Speedysteve]

ORIGINAL: Zemke

This idea of testing could even be done on a smaller scale, of a few divisions, and do several runs to see what the average results are, and answer some basic questions. Do the combat results match, can units match the historical advance rate, and so on. I think the computer program can handle several runs to get an average and let's see what that average is, not impossible at all.

I would be interested to know if anyone has been able to shift 3rd Pz Group to the North and match the historical attacks.

Anyway, I can see I am swimming upstream on this one.

[:)] I do understand what you mean but how can we possibly guarantee what specific experience or morale level would be for each unit? What about the replacements and supply provided that is done via the AI. We can't know that level of detail for most units in RL so how to model etc. I know it's details but those things also add into the combat results and also the RL advance.