AI adjustments

[rocketman]

In what sense does the adjustment to AI (+1,+2,+3) change the game? How will you tell that the AI got "smarter"? What aspect of AI decisions will improve most significantly?

[JasonC]

It doesn't get any smarter. It issues the same dumb orders as before.

The effect of + AI settings is to raise the quality of AI side units. They fight better because they get artificially inflated morale. This makes them more resilient under fire - particularly the higher settings (really, +2 or more and you wind up having to KIA most of them).

I consider it a lousy balance mechanism, because it falsifies tactics. You wind up relying more on things that don't care about it, like mega HE or flame.

The mechanism of increasing enemy force size is preferable in my opinion. That makes things harder too, without changing the relationships between the various weapons.

Better still, play a human. (I know, I know, it takes a lot longer...)

[rocketman]

I often see scenario designers suggest an increase in AI setting, rather than increasing the enemy force size, as you suggest JasonC. Might increasing the force size throw their scenario design off and disrupt the intended structure/flow of the scenario instead of giving you a tougher fight. Or both...

[JasonC]

Designers usually suggest + AI settings because they balanced the scenario for human vs. human play, and think they are supposed to provide some change to even chances for an AI game. Few of them have any real idea what level of AI setting would be appropriate. A few do - those who frequently fight the AI themselves and like using + AI settings. I suspect many of the others are just trying to make their fight more challenging, on the principle "the harder it is, the more fun it will be". Which needless to say is completely false.

But do whatever you like. Personally, I don't use + AI settings.

[GreenAsJade]

You should ask MadRussian about this - he hangs out more at WeBoB or TSD.

What I understand from his posts is that the AI fights smarter (contrary I know to what the learned JC tells us above).

GaJ

[Kingfish]

As Jason stated, you'll still get the same behavior out of the AI regardless of the raised level, only it now has better preforming troops in which to execute those orders.

[rocketman]

OK, so the AI will still be as "stupíd" as before the adjustment. What do you consider the major flaw in the AI behaviour? The thing that annoys me the most, since it is so obvious, is how troops are sent into the open despite overwhelming presence of enemy units.

[Mad Russian]

Originally posted by rocketman:

In what sense does the adjustment to AI (+1,+2,+3) change the game? How will you tell that the AI got "smarter"? What aspect of AI decisions will improve most significantly?

Well you could try playing hundreds of games against the AI and see if you can tell a difference.

Or you could ask the wonderfully talented and all knowing experts here on these forums.

Or you could ask someone that has played the AI hundreds of times.

Or you could just let the gamers that play the scenarios tell you.

Let's just take one small item into consideration. Why would BFC add a Computer Experience Bonus to the game, take all that development time and the resource space on the disk if it has no effect?

Good Hunting.

MR

[Mad Russian]

Originally posted by rocketman:

OK, so the AI will still be as "stupíd" as before the adjustment. What do you consider the major flaw in the AI behaviour? The thing that annoys me the most, since it is so obvious, is how troops are sent into the open despite overwhelming presence of enemy units.

The major flaw in with the AI is two fold. First most game designers aren't experienced enough to make it work. Second you are asking a machine to be as smart as a human player.

I'll post some comments for some of the scenarios I know about. My own. You can decide from there.

Anyone with other concrete evidence and not just a cookie cutter opinion can then start the debate about stupid the AI is.

Good Hunting.

MR

[Mad Russian]

At The Scenario Depot II my HSG KC Untstfhr Muehleck is second on the all time downloads list for CMBB scenarios. It is first in review ratings. Not bad for a vs the AI scenario...here are some of the comments that it has earned:

If you're looking for a fun "thinking mans" shoot'em up, high action, use the map or die, never a dull moment solo play scenario then this is the one for you.

Played as Axis vs AI.

I can see why this is top rated, it's great. The first time through, I got stomped into the dirt.

Exellent scenario, Gotta have good skills using Panther Tanks or it could be a nightmare. Loved it.

Have played through this scenario several times and its keeps getting better.

Its one hell of a fight and all in all one hell of a scenario.

I played this one back when it was posted at the depot and smashed the Russian AI with the loss of only a few men and a couple of halftracks. Had a blast doing it as well. Just played it again and suffered a total defeat turning all 3 of my big cats into flaming hulks. Oh well, it is an intense and fun scenario vs. the AI.

I thought this scenario was a complete blast! Played Axis vs. AI with a +1 bonus. I was amazed at how this scenario did not let up turn after turn, especially given my situation. If you like high action scenarios, short sharp engagments, I highly recommend this nailbiting slugfest.

I played at +2 vs the AI and achieved a major victory, but it was never 'plain sailing' with losses on both sides and blazing hulks littering the battlefield.

This scenario is all action from start to finish and without a doubt, any German victory here is going to win someone a Knights Cross!

Excellent Work by the Mad Russian! A pleasure to play and highly recommended.

Smart scenario with plenty of action, the map was nicely made with several options for a good defense. The AI put up a good fight with a last minute infantry rush over the hills at the VF. I managed a Major Victory and lost one Panther in the process, although the hero of the piece lost his head to MG fire when he very unwisely poked his head out for a look see.

Take particular note that all of my scenarios are recommended to be played as:

Add computer experience bonus relative to the experience you have as a player. Good player +1, above average player +2 and very experienced player +3.

Also, please note the comments about how the AI played. Seems that there are comments other than how the AI is dumb as a box of rocks.

An Indy race car is just a car unless you put it on the right track with the right driver. Let an experienced designer with all the tools that CM gives you and let the gamer FOLLOW DIRECTIONS and the AI will fight just fine.

That includes using the CEB to increase the game play in vs AI scenarios.

Good Hunting.

MR

[Mad Russian]

But you know what, maybe that first scenario was just a fluke. I mean even a broken clock is right twice a day right?

So then, let's move on the the number 3 downloaded scenario for CMBB on The Scenario Depot II. It is also 24th in the ratings for CMBB scenarios.

That would be HSG EW Panthers Roll.

Here some of the comments it has earned:

Excellent fight against the AI. Gave them a +1 experience bonus and managed to score a tactical victory.

As indicated before you really feel overwhelmed at first, but hang in there and you will experience a very fun and intense armor heavy fight.

Per the designer's advice, I played as Germans vs. AI, and it rapidly became a desperate defense. You will have your hands full holding back the Red hordes, praying for your Panthers to arrive by the next turn. Not that they guarantee your salvation!

This is how it should be done.... when I play such a battle in only 2 episodes (migraine interval doh!) it means it's cool!

I played this a long time ago and I remembered I got crushed.

Now I've got some experience I played a Major victory.

Again MR succeeds in creating a 2 phased battle.

Retreat and counterattack.

Plus there's a good PantherQuality/RussianQuantityRate.

You should/must play this one!

Very good scenario. Went down to the wire with Axis achieving a 55%/45% DRAW and a FLAG changing hands on the last turn. After the first three turns I actually wanted to quit, due to all the demoralizing things that had occurred. I'm glad I played it out. I plan on playing it again, because I think I can achieve a modest level of victory with the Axis, by setting up differently.

Ripping good time...

Well now, that doesn't seem to be the kind of reviews that one would expect from a "box of rocks" is it?

Again, I expect the gamer to use the CEB. Yes, makes a difference in how the game plays.

Good Hunting.

MR

[Mad Russian]

Alright then...but still that's only two scenarios...so let's take a look at some others....

How about HSG B Bussard

I played all games with +2 AI experience and managed a score of zero first attempt and 100% 4th attempt. The addiction is trying to better your previous score and trying out different tactics.

Or HSG B Each in Turn...

Well I reckon the 'B' in the scenario title stands for 'Bitch'... I started this baby thinking 'a tank battle vs the AI...? how hard can it be...?

Trust me, with a +3 CEB, it's hard enough. Do yourself a favour, download and play it.

Thanks for a tense couple of hours gaming vs the AI - and that's not something I say every day...

Maybe you want some CMAK examples like HSG KC Untscfhr Barkmann..this one is not only against the AI but an exit scenario for the player as well...

I found this scenario to be a mix of intense close action mixed with suspense as you poke your way forward trying to get out of the pocket with your force in tact.

The map is good and allows for multiple escape routes and the AI responds well which results in a good vs the AI scenario.

If one looks in the discussion area they will find my old AAR, a story of recon, suspense and constant ambushes!

I have never done really well with exit scenarios and to be honest I did not fare that well against the AI as the Germans in this fight either. Having said that it is still a well made and enjoyable battle and as the attackers you can never be quite sure what was around the next bend in the road.

We can go on with more examples if anyone would like to discuss this further.

The bottom line is that the Computer Experience Bonus has the AI playing better. I can show that. I've seen it enough time to include the disclaimer on how to use the CEB in my scenarios.

Is it tough to get the AI to fight? Somewhat. It's not the easiest thing in CM to make happen. Can it be dumb as a box of rocks. Yes. Can some of the experts in these forums talk about things they know nothing about...yes. Can the AI fight?

I think the download totals for The Scenario Depot II shows it can. I seriously doubt that a scenario that wasn't really good would be the second most downloaded scenario for CMBB and at the same time be the highest rated scenario. That scenario is only recommended for vs AI play.

Those that do QB's vs the AI get what you ask for. A quick throw together. Will the AI fight then? Most likely not. Will the AI require some adjustment and a bit of help here and there? Probably.

So what? All scenarios require playbalancing if they are going to be worth playing. Scenarios vs the AI are no different.

Good Hunting.

MR

[Mad Russian]

Originally posted by JasonC:

Designers usually suggest + AI settings because they balanced the scenario for human vs. human play, and think they are supposed to provide some change to even chances for an AI game. Few of them have any real idea what level of AI setting would be appropriate. A few do - those who frequently fight the AI themselves and like using + AI settings. I suspect many of the others are just trying to make their fight more challenging, on the principle "the harder it is, the more fun it will be". Which needless to say is completely false.

But do whatever you like. Personally, I don't use + AI settings.

Try not to tell me how I balance my scenarios.

I playtest and balance my vs the AI scenarios with a +2 CEB. They are set that way for a reason.

I thought the entire reason for playbalancing is, "to make the fights more challenging". That's the reason I playtest and balance my scenarios. Can't speak for your ultimate goals as a designer.

Personally, if you don't use +AI settings how can you even comment on their usefulness or effectiveness?

Good Hunting.

MR

[McIvan]

What a load of completely uncessary sarcasm.

The question is: Does the CEB bonus simply increase the experience level of the Ai troops, or does it alter the tactics of the AI?

All you've said is that a good designer can work with the AI's predilictions to create a good fight. Well, hallelujah, and I am (sarcasm aside) deeply indebted to the designers such as yourself who take up their spare time to do so......but I don't think that was really the issue.

[Mad Russian]

The issue is can the AI do better with the CEB at a higher level.

The answer I think is yes. I thought I was showing that gamers feel that way too. It was not sarcasm on my part. (Well, the part where Jason says he doesn't use it but has alot to say about how it doesn't work, was sarcasm.)

Let's look at this another way.

If all that needs be done is to raise the experience levels the designer can do that simply. I could make all the troops in the scenario, at say crack,, or elite, and then there wouldn't be a need for a CEB setting in the first place. Much easier for BFC to do that as well. Just put a statement in the rules booklet. Instead of all the programming and development time to actually add the CEB setting in the editor.

Does that make more sense to you?

I've played hundreds of games vs the AI. I playtest everyone of my scenarios at least once, vs the AI for various aspects of playbalance. All I know is that if I move the CEB setting higher it plays better. It will do such things as flank you, rush the flags, stalk your tanks, do combined arms attacks. etc...if that's just from getting better experience levels okay. I only see those kinds of responses from the AI when the CEB goes higher. Not when I just add to the experience levels of the troops themselves.

The question was does the CEB make a difference. The answer, for me at least, is YES.

There are alot of comments that I didn't put on here. They get pretty redundant after awhile. All of them in awe that the CEB addition has the AI playing better than they have ever seen it before.

Is it that I've learned to make the AI work by itself? I don't think so since I always use a +2 CEB. That doesn't mean that I get crack and expert troops everytime. I sometimes have conscript troops in my scenarios that get moved up to "only" veteran. If the experience levels are all that make the difference then why wouldn't they be giving you a good fight with 0 CEB instead of adding to it?

Good Hunting.

MR

[ December 12, 2007, 05:24 PM: Message edited by: Mad Russian ]

[Mad Russian]

Double post. Been awhile since I posted here.

Good Hunting.

MR

[Kingfish]

Originally posted by Mad Russian:

If all that needs be done is to raise the experience levels the designer can do that simply. I could make all the troops in the scenario at say crack or elite and then there wouldn't be a need for a CEB setting in the first place. Much easier for BFC to do that as well. Just put a statement in the rules booklet. Instead of all the programming and development time to actually add the CEB setting in the editor.

I don't think the CEB was intended for designer scenarios as much as it was for quick battles, where the players are restricted to a point limit.

Your points are valid and worth considering, but lets look at it another way. If BFC was able to code a function in the game that made the AI perform better and smarter with a simple click of the mouse, then why not have that function be the overlying code for the entire game?

IOW, if it is universally understood that a smarter AI is a huge selling point, why would they code it so you can only get it as a selectable option, one that few people really use?

[Mad Russian]

Originally posted by Kingfish:

</font><blockquote>quote:</font><hr />Originally posted by Mad Russian:

If all that needs be done is to raise the experience levels the designer can do that simply. I could make all the troops in the scenario at say crack or elite and then there wouldn't be a need for a CEB setting in the first place. Much easier for BFC to do that as well. Just put a statement in the rules booklet. Instead of all the programming and development time to actually add the CEB setting in the editor.

I don't think the CEB was intended for designer scenarios as much as it was for quick battles, where the players are restricted to a point limit.

Your points are valid and worth considering, but lets look at it another way. If BFC was able to code a function in the game that made the AI perform better and smarter with a simple click of the mouse, then why not have that function be the overlying code for the entire game?

IOW, if it is universally understood that a smarter AI is a huge selling point, why would they code it so you can only get it as a selectable option, one that few people really use? </font>

[Kingfish]

Originally posted by Mad Russian:

I've tried the same scenario with both just higher experience levels and then adding the higher level CEB. The best game play I get out of the AI is with the higher level CEB.

I have no doubt that you get a better game out of the AI when the CEB is increased, but that is a function of the AI using better functioning troops rather than making smarter decisions.

IMO, the AI will still act as predictably with regards to battlefield tactics. When on the attack it will still choose the one axis of advance. When on the defense it will leave 300 point flags unguarded while it launches piece-meal counterattacks in an attempt to recapture a 100 pt flag.

Now, having said all that, I do acknowledge a marked improvement over the AI's performance when comparing CMBO to CMAK. In the latter's case the AI will actually deploy armor in overwatch while the infantry screens forward. The infantry employs cover fire in support of advancing elements. Heavy weapons deploy to cover the main attack. Artillery spotters actually call in strikes, both prep and in response to battle situations. All this is observed in games with and without CEB, only you get a much better version of it obviously with it.

BTW, thanks for the pub. I appreciate it.

[Mad Russian]

Originally posted by Kingfish:

I have no doubt that you get a better game out of the AI when the CEB is increased, but that is a function of the AI using better functioning troops rather than making smarter decisions.

IMO, the AI will still act as predictably with regards to battlefield tactics. When on the attack it will still choose the one axis of advance. When on the defense it will leave 300 point flags unguarded while it launches piece-meal counterattacks in an attempt to recapture a 100 pt flag.

Now, having said all that, I do acknowledge a marked improvement over the AI's performance when comparing CMBO to CMAK. In the latter's case the AI will actually deploy armor in overwatch while the infantry screens forward. The infantry employs cover fire in support of advancing elements. Heavy weapons deploy to cover the main attack. Artillery spotters actually call in strikes, both prep and in response to battle situations. All this is observed in games with and without CEB, only you get a much better version of it obviously with it.

BTW, thanks for the pub. I appreciate it.

Not when you make the scenario with elite troops and play it out at 0 CEB. Then play the same scenario with an added CEB and it plays better.

You guys can use the CEB or not in your own game play. I recommend it because it does a better job for me. But then I don't have a dumb as a box of rocks AI either.

Funny you should mention the differences between games. I see a marked reduction in AI capability between CMBB and CMAK. In CMAK, it seems to me, the AI wants all games to be Meeting Engagements and will advance no matter what. I have had to use reinforcement groups to get the AI to defend instead of attack.

Yes, the designer has to do some work with the system as well. It's back to the Indy car. Just any driver can't win the Indy 500 with just any car. Both have to do their job. The scneario has to be playtested and balanced for the AI to do it's job. But that's no less true for H2H play. If a gamer plays bad he'll lose too.

Well you're welcome. IMO, you're one of the designers that doesn't get mentioned enough for the work you do. And since we are talking about scenarios and what works and what doesn't, gamers should know that your scenarios work.

Good Hunting.

MR

[JasonC]

I kept waiting for Mad Russian to make a claim.

I'm still waiting, because he hasn't.

[Mad Russian]

Originally posted by JasonC:

I kept waiting for Mad Russian to make a claim.

I'm still waiting, because he hasn't.

Sorry you missed my claim.

I made one claim Jason, over and over. The game plays better with higher CEB's applied.

I didn't say I never use it and then tell others why it doesn't work though.

Good Hunting.

MR

[GSX]

Wow, I came over here for a bit of peace from the SF Forums and guess what. Its real guys talking about the mechanics of a great game.

Keep it up! Its refreshing.

[Mad Russian]

Originally posted by Kingfish:

I have no doubt that you get a better game out of the AI when the CEB is increased, but that is a function of the AI using better functioning troops rather than making smarter decisions.

[/QB]

Let's just take that comment right there at face value. If you have no doubt the AI gives you a better game with the CEB increased isn't that what the original question was?

The strangest thing to me in all of the CM world is what difference does it make what BFC did to allow for better game play? If using a higher CEB gives you better results in the game what difference does it make? If that happens because of an increase in CEB, the designer doing a better job, a combination of the two, or the little miniature men my disks came with to install inside my case so when I play they direct the action?

The little guys eat quite a bit but my CM games play better because of them.... :eek:

I play the computer mostly because I check all of my scenarios vs the AI to see how it does. I play them Hotseat as well. Between the two of them I get the information I need to put that scenario close enough for actual playtesting in most instances. I also play alot of vs the AI scenarios for designers at TPG. So, I play the AI alot.

For me personally I wouldn't play a vs the AI without at least a +2 CEB. But again, that's just me. If it works for ou without using it so much the better.

Good Hunting.

MR

[McIvan]

Well, the question is more like this:

Does CEB mean the AI implements smart plans instead of stupid plans, or does it just mean that the troops are more capable of implementing exactly the same stupid plans as before.