Does the Ai Army size scale with you?

[A. P. Hill]

On ‎11‎/‎23‎/‎2016 at 4:23 PM, Koro said:

... So take your pick :). If you've got any suggestions on how the game could otherwise handle a large variance in the army numbers of players at the same stage, then I'd love to hear it.

 

I would suggest limiting the AI to the same restrictions as players have to go through.  Granted the AI will always have the upper hand on being able to calculate, but it should also be limited.  None of it's calculations should extract information from it's opponent. It's calculations should be separate and independent.

[Lincolns Mullet]

1 hour ago, A. P. Hill said:

I would suggest limiting the AI to the same restrictions as players have to go through.  Granted the AI will always have the upper hand on being able to calculate, but it should also be limited.  None of it's calculations should extract information from it's opponent. It's calculations should be separate and independent.

Well, the restrictions the player has to go through involves being awarded men and gold for battles and then building an army with those resources.  The men and gold are abstractions because the greater war isn't simulated.  Why award 4500 men for winning a battle?  That number only makes sense from a design perspective (Ie, the campaign is balanced when providing 4500 men for this victory and not 10000 men or 1500).  If the AI is similarly restricted to building an army with limited gold/men, you'd essentially have the same effect you currently have now: a strong or weak AI. How it gets strong or weak is irrelevant except through a players own perception on how the AI is constructing its army.

From what I see being posted, what players are really asking for is to see a direct correlation between a battle and the state of the AI army in subsequent battles.  If I kill 30k Union soldiers in this battle, why are they now STRONGER in the next battle?  The perception here is that the AI is cheating, and nobody likes a cheat. lol

We want greater purpose to our achievements in each battle.  We want to see the enemy suffer because, through our sheer military genius, we've been destroying them and they deserve to have only 10,000 troops at Antietam.  However, because of the structure of the campaign in that we are essentially playing a highlight reel of Civil War battles in chronological order, how do the devs create a malleable AI without sacrificing the core design of the campaign?  The core design being: you're in control of your army, but we want you to be able to fight these historical battles as they were and see how you perform vs. history.  In this design, the AI must always have a functioning army regardless of how well you do against them in previous battles. Whether you agree with that or not...that's just how it's currently designed.

So, you either allow the enemy to get destroyed to a pulp and you never have any need to make it past 2nd Bull run or Antietam because the AI no longer has an army to fight with...OR...you allow the AI extra flexibility to keep their forces strong enough to oppose you all the way to Richmond or Washington, regardless of the losses they suffer in previous battles.

Let's say the AI army has to play by the same exact rules as you do: they receive men/gold depending on whether they win, lose, or draw in the battles you play against them. They also have reputation.  This seems like a fair set up, no?  This is balanced. Maybe the rewards are different from what you receive had you won the battle (more or less), but the effect is the same: the AI has to play fair.

This would essentially be a multiplayer experience vs the AI.  Assuming the devs could create an AI that could properly form new brigades, sell weapons, upgrade skill points, etc. then you will have an AI that is fairly constructed but incapable of fighting historical battles.  Antietam won't be Antietam, Gaines Mill won't be Gaines Mill, etc.  The force composition will be all over the place, and once they suffer a few heavy losses...game over, just like it would be for you. Unless, of course, they get an extra boost to stay competitive...which is what they currently receive.

I believe there's some middle ground solutions the devs can try, but because of the design of the campaign as it stands now, the AI has to remain competitive all the way to the end. Unless the devs want to change their vision for the campaign, I don't see a way around this except for some minor tweaks/additions to simulate carryover effects between battles.  In a dynamic campaign, the sky is the limit on how the AI constructs its army as the battles aren't meant to be historical.  But in a fixed campaign like this, Antietam must always be Antietam...or at least, a close approximation.

Just my .45 cents!

 

 

[bonsainut]

1 hour ago, Lincolns Mullet said:

From what I see being posted, what players are really asking for is to see a direct correlation between a battle and the state of the AI army in subsequent battles.  If I kill 30k Union soldiers in this battle, why are they now STRONGER in the next battle?  The perception here is that the AI is cheating, and nobody likes a cheat. lol

Not from my perspective.  What I am asking to see is NO correlation between one battle and the state of the AI army in subsequent battles.  The AI battles should get progressively harder in a fixed balanced fashion (regardless of how well you fight or don't) with the outcome being that if your army doesn't get continuously stronger at the same pace, you will lose the game.

What I object to is that I can save the game, and start a battle with 20,000 troops and the AI has 40,000.  I reload to the save, sell some equipment, and buy 10,000 more troops, so I start the battle with 30,000 troops, and suddenly the AI has 60,000.  So it doesn't matter how many troops I buy or don't buy, because the AI is giving me the same odds regardless.  Therefore it doesn't matter how well I play, because the AI will always scale up (though it appears it won't scale down, not completely).  So instead of enjoying the game, and playing to win, I find myself trying to figure out how the game system will cheat to deny me a hard-won advantage.  Won't scale for premium gear?  Ok so now I won't buy troops, but will spend all my money on equipment...

Contrast that to a scenario where you know that the CSA will always have 80,000 troops at Shiloh.  Play poorly in the earlier battles, and you might only have 40,000 troops and might not be able to win.  Play WELL in the earlier battles, and you might have 60,000 troops and might win more easily.  But you know the AI isn't suddenly getting extra troops because you just purchased some yourself.  And if you find the gameplay too easy, you can just slide the difficulty up one notch, and the AI has the same number of troops, but they suddenly have 125% morale and do 125% the damage (for example).

AI scaling is just a lazy way to say "I don't know how to game balance in my game design."  They said as much when they said their concern was that a player would win big in an early battle and "steamroll" the rest of the content.  I can think of a million ways to balance so this doesn't happen, so I hope that they will figure it out as well.

The funny thing is... we can't even comment on the difficulty or ease of the different battles (as game testers) because we are all playing different games!  No one can say "I won Shiloh on uber difficulty" because one person's version of that battle is currently different from everyone elses' versions - based on an arbitrary algorithm that doesn't take player skill into account.  Not only are we bringing our own different armies to the battle - the AI is bringing a different army as well. It makes no sense.

Edited November 29, 2016 by bonsainut

[Flef]

I think the player should not be able to use its manpower during a campaign (like 1st Mannassas or the Peninsula). It would require to  tone down the autoscaling,  but It would be a bit more a game of "do I really want to pay the high price for this crossroad?"

 

And I also think that the scaling should high-capped relatively to the casualty rate. If the player makes a campaign with high casualties but not by inflicting himself enough casualties, it should be in a difficult position at the grand battle. But if the player is able to spare its army, and to inflict a lot of casualties to the AI, then the grand battle should be quite easy.

 

 

 

 

 

 

 

 

 

Edited November 29, 2016 by Flef

[A. P. Hill]

15 hours ago, Lincolns Mullet said:

Well, the restrictions the player has to go through involves being awarded men and gold for battles and then building an army with those resources.  ...

 

 

Then the AI should also have these restrictions and not an unlimited amount of men or gold to work with. But based on your thought process, maybe the AI can get a 2 percent increase in these items to sustain itself.  It should be forced to calculate how best to use it's limited resources as well.

But then I suppose a resource limitations setting like this is better suited for PvP than PvAI.

Edited November 29, 2016 by A. P. Hill

[Koro]

31 minutes ago, A. P. Hill said:

Then the AI should also have these restrictions and not an unlimited amount of men or gold to work with. But based on your thought process, maybe the AI can get a 2 percent increase in these items to sustain itself.  It should be forced to calculate how best to use it's limited resources as well.

But then I suppose a resource limitations setting like this is better suited for PvP than PvAI.

If this is the case, the AI would run out of men against me and many others after 2 or 3 battles and I'd never experience the full campaign.

In almost all strategy games, the AI "cheats" somehow. In Total War they get extra resources and don't pay upkeep for huge amounts of troops f.x. This is all generally accepted as an AI will always have a handicap. While this solution isn't elegant the only alternative I can think of this:

Have each stage of the campaign set for a certain amount of men, that can perhaps vary a bit. If you're behind as a player you get more men, same terms as the AI. You'll not have veterans though if you throw away your army but always be replenished. I'm not sure this would be more fun though.

[Grognard_JC]

16 minutes ago, Koro said:

If this is the case, the AI would run out of men against me and many others after 2 or 3 battles and I'd never experience the full campaign.

In almost all strategy games, the AI "cheats" somehow. In Total War they get extra resources and don't pay upkeep for huge amounts of troops f.x. This is all generally accepted as an AI will always have a handicap. While this solution isn't elegant the only alternative I can think of this:

Have each stage of the campaign set for a certain amount of men, that can perhaps vary a bit. If you're behind as a player you get more men, same terms as the AI. You'll not have veterans though if you throw away your army but always be replenished. I'm not sure this would be more fun though.

As an alternative to making AI stronger or smarter... We could maybe think of going MULTIPLAYER for the campaign. There would be much less trouble, and this would be most exciting.  No more scaling problems and you really have to manage your ressources in the long run this time.

Edited November 29, 2016 by Grognard_JC

[Koro]

Just now, Grognard_JC said:

As an alternative to make AI stronger or smarter... We could maybe think of going MULTIPLAYER for the campaign. There would be much less trouble, and this would be most exciting.  No more scaling problems and you really have to manage your ressources in the long run this time.

Then what happens when player 1 wipes player 2 off the map? Same problem really, at least if you're going for a campaign :). I think you'd just see lots of people quitting and very few actually making it through the campaign.

I'm all in for the MP battles though and at some point I'll write a thread about all the good and bad reasons I see for it.

[Grognard_JC]

3 minutes ago, Koro said:

Then what happens when player 1 wipes player 2 off the map? Same problem really, at least if you're going for a campaign :). I think you'd just see lots of people quitting and very few actually making it through the campaign.

I'm all in for the MP battles though and at some point I'll write a thread about all the good and bad reasons I see for it.

Aye, this is true. However, with a more dynamic approach to the battle, like in UG:G (victory points counter-weighted by your own losses), players would be "dissuaded" from all-out strategies.

What's more, you Koro especially know this well, that in UG series, it takes 2 players to have very different levels of gaming to finish up wining with a complete wipe. I hardly won ever of my game in 1v1 in UG : G, nor was I defeated that way either. So I'd be very confident in an MP campaign if I were you.

Edited November 29, 2016 by Grognard_JC

[Koro]

11 minutes ago, Grognard_JC said:

Aye, this is true. However, with a more dynamic approach to the battle, like in UG:G (victory points counter-weighted by your own losses), players would be "dissuaded" from all-out strategies.

What's more, you Koro especially know this well, that in UG series, it takes 2 players to have very different levels of gaming to finish up wining with a complete wipe. I hardly won ever of my game in 1v1 in UG : G, nor was I defeated that way either. So I'd be very confident in an MP campaign if I were you.

Yes, that's true actually. I've rarely wiped the floor with anyone, even with my good grasp of UGG.

 

[Lincolns Mullet]

28 minutes ago, A. P. Hill said:

Then the AI should also have these restrictions and not an unlimited amount of men or gold to work with. But based on your thought process, maybe the AI can get a 2 percent increase in these items to sustain itself.  It should be forced to calculate how best to use it's limited resources as well.

But then I suppose a resource limitations setting like this is better suited for PvP than PvAI.

Right. This would work great in PvP or skirmish PvAI where its more about army balance than historical numbers. The way the campaign works now, your force vs the AI force should be closer to the historical ratio of forces during a specific battle so you experience the battle on similar footing. Thats really the crux of the matter. You're able to greatly vary your army size and strength vs a fixed force opponent which can either make the battle impossible to beat, a rollover, or anywhere inbetween (if you removed scaling).

AI "boosts" to morale, damage, and/or numbers all have the same given effect to make the game more difficult. These are all easy to implement in some degree or another. And without creating a brilliant AI, this is how you make the enemy stronger.

The alternative is that you scale down the amount of resources you receive for winning battles, so that you have to be ultra careful with your men to be able to realistically face a 63k strong Union army at Shiloh for example. Then you put the burden of difficulty on the player rather than the AI, which would reward skill.

The key, then, is making hard realistically "hard" and then figuring out what normal and easy would be.

 

 

[Grognard_JC]

11 minutes ago, Lincolns Mullet said:

Right. This would work great in PvP or skirmish PvAI where its more about army balance than historical numbers. The way the campaign works now, your force vs the AI force should be closer to the historical ratio of forces during a specific battle so you experience the battle on similar footing. Thats really the crux of the matter. You're able to greatly vary your army size and strength vs a fixed force opponent which can either make the battle impossible to beat, a rollover, or anywhere inbetween (if you removed scaling).

AI "boosts" to morale, damage, and/or numbers all have the same given effect to make the game more difficult. These are all easy to implement in some degree or another. And without creating a brilliant AI, this is how you make the enemy stronger.

The alternative is that you scale down the amount of resources you receive for winning battles, so that you have to be ultra careful with your men to be able to realistically face a 63k strong Union army at Shiloh for example. Then you put the burden of difficulty on the player rather than the AI, which would reward skill.

The key, then, is making hard realistically "hard" and then figuring out what normal and easy would be.

 

 

Agreed. And AI Boost already was in UGG by the way.

What's more, we already need a very hard mode. Hard is not hard enough. I want to lose the campaign while trying my best.

[Lincolns Mullet]

I just thought of resource restriction in my last post. I'm thinking this may be an easy way to make it more difficult for the player while keeping AI forces intact and without bonuses. The harder the difficulty, the less resources you get from battles.

As one poster pointed out..they felt cheated by successfully managing their army and reducing casualties and maximizing kills, only to find the AI magically gets more men anyway.

So if resources tight, managing your army effectively will be your reward vs a fixed AI army. This still leaves the AI army unaffected by your actions in previous battles, but the devs may have something up their sleeve to address this. At least in some small ways.

[Caesar15]

6 hours ago, Lincolns Mullet said:

I just thought of resource restriction in my last post. I'm thinking this may be an easy way to make it more difficult for the player while keeping AI forces intact and without bonuses. The harder the difficulty, the less resources you get from battles.

As one poster pointed out..they felt cheated by successfully managing their army and reducing casualties and maximizing kills, only to find the AI magically gets more men anyway.

So if resources tight, managing your army effectively will be your reward vs a fixed AI army. This still leaves the AI army unaffected by your actions in previous battles, but the devs may have something up their sleeve to address this. At least in some small ways.

Well if it scaled and you won a blowout or a very good win, wouldn't this make every battle after it easy/boring? And why does it matter if you wreck Grant's western army when the next battle is facing someone completely different. In some cases not scaling would be bad and in others it would be good (Peninsula campaign maybe?).

[Karri]

One way to balance things out a bit is to make the smaller battles before the grand battle have some effect on the actual battle. Ie. if historically it was a loss and win should give you some benefit like less enemy forces present, less quality forces present or something like this. I would also like to see what I am about to face BEFORE I reinforce and build up my units.

Played one game(easiest level) where I managed to constantly cause 3 to 1 casualties, rout enemy troops and keep my own troops in high strenght. This only meant that in the grand battle I was facing a fresh and superb US army scaled to mine. If I beat them at every corner they should be weaker, or at the very least my forces should be stronger. Right now there is little to no advantage of actually winning.

[Grognard_JC]

I just started my 1st Union campaign on hard. I get the feeling playing the Union is actually harder than the CSA. Am I alone to think so ? It feelks like your troops have very low morale.

And.. I actually put all my carreer points in army organization as I had promissed here (kind of scaling-challenge...). Although my army is still tiny, it does feel like the enemy got widely boosted.

 

I won 1st Bull Run, but I got completely crushed at crossroad, with a full brigade surrenderng what's more.

[Hitorishizuka]

7 hours ago, Grognard_JC said:

I just started my 1st Union campaign on hard. I get the feeling playing the Union is actually harder than the CSA. Am I alone to think so ? It feelks like your troops have very low morale.

And.. I actually put all my carreer points in army organization as I had promissed here (kind of scaling-challenge...). Although my army is still tiny, it does feel like the enemy got widely boosted.

 

I won 1st Bull Run, but I got completely crushed at crossroad, with a full brigade surrenderng what's more.

Union troops are garbage against equal tier if both are in cover IME. I can have 2 star with more men and what should be decent stats shooting against CSA 2 star in equal heavy cover and they'll slowly lose that fight. (50 casualties to 15 inflicted, that sort of thing.) You pretty much have to rely on artillery, cavalry, and doing whatever you can to take forest edges and force the AI to place troops in poor cover.

[TortugaPower]

Hi gang,

I find myself making an account and coming on the forum just to talk about this issue. Something feels off with the game as is. In a Union campaign, I was just obliterated (at Crossroads) despite having playing almost perfectly up to that point.  I'm okay with that if it's intentional, like maybe I should have put points in Recon in order to see that the Rebel patrol was a full army.  But it does feel funny to being doing really well and realize the AI is just going to double my forces no matter what.

This is a fantastic game tactically and the economy system they've thrown on top of it is really satisfying.  However, once you realize that the AI is scaling, I have to say that I almost lose interest in the game.  There is no point in having an economic system if it has no impact on gameplay.  If I do well and buy more men, but the AI just scales, why do I waste my time with the economy?  Just give me fixed number of troops like in UG Gettysburg.  In fact, the Gettysburg system feels better because of losses carrying over.

[Koro]

3 hours ago, TortugaPower said:

Hi gang,

I find myself making an account and coming on the forum just to talk about this issue. Something feels off with the game as is. In a Union campaign, I was just obliterated (at Crossroads) despite having playing almost perfectly up to that point.  I'm okay with that if it's intentional, like maybe I should have put points in Recon in order to see that the Rebel patrol was a full army.  But it does feel funny to being doing really well and realize the AI is just going to double my forces no matter what.

This is a fantastic game tactically and the economy system they've thrown on top of it is really satisfying.  However, once you realize that the AI is scaling, I have to say that I almost lose interest in the game.  There is no point in having an economic system if it has no impact on gameplay.  If I do well and buy more men, but the AI just scales, why do I waste my time with the economy?  Just give me fixed number of troops like in UG Gettysburg.  In fact, the Gettysburg system feels better because of losses carrying over.

Hi Tortuga. The AI does scale with you and there are various reasons for this I won't get too much in to here - I've written about it earlier in the post.

What I think happened at Crossroads last patch is that more AI units were added so that they could have smaller units instead of the massive 2.900 sized ones so the player wasn't overwhelmed What I see could have happened is the scenario scales wrong perhaps. Can you report it in-game, the deployment where you can see the numbers or in the scenario when all of the forces have arrived on the field? That will give the developers a chance to look at it. You shouldn't be that outnumbered.

[TortugaPower]

Hi Koro,

Yes, I've read through the three pages of this discussion. It was good to see confirmation of the AI scaling issue, and a good civil discussion about it overall, including your comments.

Thanks for the info about Crossroads. Nick Thomadis informed me (in this topic here) that my troop numbers for Crossroads were too small. I'll have to figure out what that means for me.

Anyway, it looks like the AI will have some reaction to battles (in terms of strength) in the campaign as of the next patch, so perhaps this will address the issue that many have presented here.

 

 

[Koro]

I responded in the other thread :).

[Alavaria]

AI wars does escalation in a nice manner, Building a small system around war progress might work.

 

Eg: Based on your performance (not too fine-grained so say Victory/Draw/Defeat) you build up the other side's Desperation meter. Since this is a fairly large war, the issue of resources isn't just your guys taking on the entire enemy country, but parts. Handwavy perhaps, but if you get the meter high, the enemy diverts resources to you which perhaps leads you to a better ending (eg: some sort of Legendary Commander ending would mean your allies managed to sweep the country while you broke the enemy head on, as it were). So you might be able to have an "easier" big fight if you accept a Draw rather than a costly Victory...

 

Also, changing things based on Desperation would be amusing eg: you are defending. At higher desperation the enemy brings a larger force, but it's recruits perhaps with a morale penalty. But it would be the case be it if you had a large army or a smaller one (it scales based on your victories, not your army number or composition). A high desperation attack mission might have the enemy with better fortifications, making better use of skirmishers but reduced numbers or supplies.

 

Your own losses affect your army (more or better vs less/worse). As for the enemy side, more losses you inflict on them reduces what they have to bring, but if you win it increases their desperation.

 

 

Explicitly showing the Desperation meter might not be very useful, but something like "War Progress updates" showing how you are (or are not) drawing all the enemy's attention and resources, thus allowing the other NPC generals to make gains elsewhere would be nice.

Edited December 4, 2016 by Alavaria

[dowdpride]

Regardless of the eventual outcome regarding AI scaling, at the very least I would like to see captured brigades not show up in the set piece battle. If I capture sherman in the peninsular campaign, he should not be available to the union until the next scenario begins, some sort of prisoner exchange etc. As it is targeting elite brigades is really only situationally useful, as they will always be forming the core of the Union battle line no matter what you do.

[TortugaPower]

@Alavaria I think the desperation system makes sense, especially having the Union as AI (as mostly likely the Union would have rallied many more individuals and increased war production if their lands had been under serious threat). However, I wonder if it is too difficult to implement at this stage of the development.

[Grognard_JC]

Hello guys.

I just found time to finish my CSA campaign at Antitentam with Full Army Organization focus. The goal was to max the number of units I could get. Hard difficulty.

 

So I was able to bring together 4 corps into battle. 1 veteran, 1 average, 2 recruits.

I don't know what happened (I got wipped too fast maybe ?) but the reserve corp NEVER entered the battle. Maybe it's a bug ?

 

The Union was SO NUMEROUS (minimum brigade strength is around 2600 while I oculd not bring more than 1000 men/brigade) and SO STRONG (2-3 stars everywhere).

 

It was complete disaster. I got completely wiped, and I really tried y best.

 

The AI deployed more than 130 000 men, while I deployed 48 000 (according to the battle result) of which I did not see one full corps of 13 000 men.

I totally got wipped.

 

Think of it : 130 000 against 35 000 !!!

 

 

This is it, maxing Army Organization is the way to make the game MUCH harder. I you want less challenge, just play soft and max out Veteran and politic/economy.

 

Hope the next patch will  help things balance.

Edited December 11, 2016 by Grognard_JC