QBs and AI

[Tank Hunter]

I usually like playing small QBs. It's well known that AI routines for picking forces are not that suitable for smaller battles due to how TOE works which often results in strange force selections.

One of the other issues is that AI does not know what forces he has and how to use them with the AI plans. I just played a battle yesterday where the AI (German) had a support wpns company together with an Inf Company. He used different plans for each of these forces. The battle was a meeting engagement with two different VPs.

As the battle started I decided to split my forces and advance with half of the force to the nearest objective while the rest of the forces were to protect the flank of my advancing unit. Once the first objective was taken the flank protection force was to sweep around and take the second objective.

After few minutes I realized that the AI was actually trying to outflank me and my flank protection force engaged. The problem was that the AI was trying to flank me with it's heavy wpns support company all mounted in their kubelwagens and trucks! It kind of looked more like an ambush of a force that was on their way to the front line rather than someone who was trying to flank me.

The AI can't really handle support formations in a good way. The AI plans should have more options to allow the designer to define which type of units should be able to use them. It should also be more selective with the employement of heavy weapons and use them across the battlefield in a support role rather than direct assault mode.

[MikeyD]

Nothing beats a genuine hand-crafted scenario. Everything else is second best. I hope people don't content themselves with just posting H2H and QB maps to Repository. Take your map and also do a proper scenario with it! There no rule against the same map doing triple-duty as standalone scenario, QB map and embedded scenario in a campaign.

[eltorrente]

I like playing QB against the AI, but when I do I play with a big handicap. It's also fun sometimes to see if I can win a battle with nothing but one elite platoon, an a ton of artillery, in a large battle for instance. Maybe make a badass little SEAL team or something and see if I can take out a whole company - just fun stuff like that. I learn to use them most efficiently, and it's good practice and fun.

Overall though, the AI simply isn't a good opponent. It's too easy to out smart it, and it doesn't resond to what you are doing like a human would. It's best when it's defending in a scenario or campaign.

Once you play a PBEM, this game goes to a different level in both fun and challenge.

[Tank Hunter]

Nothing beats a genuine hand-crafted scenario. Everything else is second best. I hope people don't content themselves with just posting H2H and QB maps to Repository. Take your map and also do a proper scenario with it! There no rule against the same map doing triple-duty as standalone scenario, QB map and embedded scenario in a campaign.

I agree that scenarios are best when it comes to challenge but QBs are still by many players considered an important part and some actually only play QBs. Having the possibility to select your own forces is quite entertaining. I don't consider lack of generated maps a problem but having a bad AI is kind of a game breaker if you prefer quick battles.

Improving the way AI works with battle plans in QBs is the way I see it a required improvement if you wan't the game to survive out there, something that BF is aiming at due to their announced upgrade policy.

[womble]

I agree that scenarios are best when it comes to challenge but QBs are still by many players considered an important part and some actually only play QBs. Having the possibility to select your own forces is quite entertaining. I don't consider lack of generated maps a problem but having a bad AI is kind of a game breaker if you prefer quick battles.

Being able to pick my own forces has always, since the early days of my wargaming "career" been a big, massive, nay huge attraction. If scenarios could permit some greater or lesser or designer-bounded selection of the human force, I'd be a very happy bunny indeed. If QBs allowed a force selection (or a selection of selections) to be associated with the AI plan(s), it would make the plans more of a challenge to beat.

In more general terms, having OOB files that you could load into any given environment, whether that be scenario generation, or game setup, for AI or player, would be fantastic.