Strategy 101 resources wanted

[ralphtricky]

I do want to spend as much time as I can in doing other things.

I mean in the AI other than the strategic layer.

[Martin_Goliath]

Ralph,
I think I remember you talking about using A-star search to improve AI movement a long time ago. By chance, I came across a report at work where they use the same method to let an AI select the best operational plan http://www.foi.se/rapport?rNo=FOI-R--3635--SE. Of course, those guys simulate tens of thousands of cases, and that would turn TOAW into a CPU-eating monster. However, there might be something in it that can move us beyond objective tracks without increasing computing needs excessively.
Regards,
Martin

[governato]

ORIGINAL: MarGol

Ralph,
I think I remember you talking about using A-star search to improve AI movement a long time ago. By chance, I came across a report at work where they use the same method to let an AI select the best operational plan http://www.foi.se/rapport?rNo=FOI-R--3635--SE. Of course, those guys simulate tens of thousands of cases, and that would turn TOAW into a CPU-eating monster. However, there might be something in it that can move us beyond objective tracks without increasing computing needs excessively.
Regards,
Martin

Again, with an eye firmly on time and resources constraints, I'd be happy to play with different levels of PO 'smartness', or having to wait a few minutes for the PO move in larger scenarios. And computers are a lot faster than when TOAW (or even v3.4) came up.

[ralphtricky]

ORIGINAL: governato
Again, with an eye firmly on time and resources constraints, I'd be happy to play with different levels of PO 'smartness', or having to wait a few minutes for the PO move in larger scenarios. And computers are a lot faster than when TOAW (or even v3.4) came up.

Much as I wish that were true, any one core of a computer really isn't all that much faster. To really speed up parts of TOAW, they need to be rewritten for multi-core CPUs. Those same pieces can also have their core routines rewritten to new algorithms, which is much simpler.

Chess is one area where throwing more hardware at is has made the AI unbeatable. Go and many other games really need smarter algorithms, just adding more cpu power doesn't help a lot.

[ralphtricky]

Here's a good talk about the difference between 'good' and 'fun' AI. Eventually, I also want to add 'personality' to the AIs, some might be more aggressive, some might be more passive. Usage of armor for breakthroughs, different levels of reserves might be kept, etc.

http://www.youtube.com/watch?v=IJcuQQ1eWWI&feature=sub

[sPzAbt653]

Thanks for the link. It seems he might allude to 'strategic' level stuff at the 54 minute mark.

TOAW is obviously far more complex than chess or checkers. I don't know anything about Civilization so I don't know how TOAW compares to that, but it seems TOAW is much simplier (no Temple building or Tech Trading).

[Foggy]

I always PBEM - PO playing is just wasting time. I don't think the majority of TAOW play the PO except for scen tweaking or deciding on a scen [8|]

[tcarusil]

Speaking for those of us who are not 5 star generals, PO playing provides me with my only chance of success.[;)]

ORIGINAL: Foggy

I always PBEM - PO playing is just wasting time. I don't think the majority of TAOW play the PO except for scen tweaking or deciding on a scen [8|]

TomC

[ralphtricky]

ORIGINAL: sPzAbt653
Thanks for the link. It seems he might allude to 'strategic' level stuff at the 54 minute mark.

TOAW is obviously far more complex than chess or checkers. I don't know anything about Civilization so I don't know how TOAW compares to that, but it seems TOAW is much simplier (no Temple building or Tech Trading).

It's both from an AI standpoint. It's more complex in that it has several different systems and a random map. You also want to stage units behind the line for an assault which I haven't seen a lot of in TOAW. In TOAW, people replay scenarios and want different results. In addition, it should follow the historic path is some scenarios which isn't always the 'best' path.

[Lobster]

ORIGINAL: Foggy

I always PBEM - PO playing is just wasting time. I don't think the majority of TAOW play the PO except for scen tweaking or deciding on a scen [8|]

I would hope this is not true. If it were then this game is in a very sorry state in regards to how many play. Looking at the opponents wanted at several different sites it would then seem only about two dozen or so actually play the game according to your 'the majority' statement.

What you don't see on this site as well as the others are the number of people who might be playing and do not bother posting because they play the PO. I prefer to think there is a large silent majority and the posters are only the tip of an iceberg. Otherwise it would be rather depressing thinking so few think this great game is worth bothering with.

[golden delicious]

ORIGINAL: ralphtrick

Here's a good talk about the difference between 'good' and 'fun' AI. Eventually, I also want to add 'personality' to the AIs, some might be more aggressive, some might be more passive. Usage of armor for breakthroughs, different levels of reserves might be kept, etc.

http://www.youtube.com/watch?v=IJcuQQ1eWWI&feature=sub

IF year = 1900 AND playerwinning = "Yes" THEN DOWHuman()

Wow.

[sPzAbt653]

ORIGINAL: Foggy

I always PBEM - PO playing is just wasting time. I don't think the majority of TAOW play the PO except for scen tweaking or deciding on a scen [8|]

I don't share the same opinion. For example, if I consider there are 200 players enjoying their Saturday by playing TOAW, I can't imagine there are 80 pbem matches going on while only 40 play solitaire. I would think the numbers would be reversed.

In that video, the Civ developers stated that 95% of their players play solitaire. I would bet that TOAW has a number similar to that. I'm not lobbying for one experience being 'better' than the other, I'm saying that its more difficult to find an opponent than it is to play solitaire.

[Rodwonder]

I'm part of that silent majority... [;)]

[sPzAbt653]

... several different systems and a random map.

The Map ! I remember seeing a show about robots negotiating terrain on their own. It looked very impressive until they put the robots on an unfamiliar course. They quickly started running into stuff. And poor Elmer, he has no video eyes to see our maps, how can he play at all ??

... stage units behind the line for an assault which I haven't seen a lot of in TOAW.

Me either, and it annoys me. Along with playing with no reserves. I haven't come to a conclusion if this is a fault of scenario design or TOAW in general.

[sPzAbt653]

IF year = 1900 AND playerwinning = "Yes" THEN DOWHuman()

Wow.

Yes Ben, the computer CHEATS !

I KNEW IT !! [:@]

[sPzAbt653]

He should not be overstacking, I'll look at that and the flanking. Part of that is that he can be much less cautious depending on the formation stance.

I just realised that changes may have been made since 3.4 or early 3.5, so I may be using an outdated Elmer. Sorry about that.

[golden delicious]

ORIGINAL: sPzAbt653

I don't share the same opinion. For example, if I consider there are 200 players enjoying their Saturday by playing TOAW, I can't imagine there are 80 pbem matches going on while only 40 play solitaire. I would think the numbers would be reversed.

Well;
a) it's probably more like 20 and
b) otherwise you're factually correct. Even I've started more matches vs. PO than I have PBEM. However, a clear majority of the fun I've had has been from PBEM.

[golden delicious]

ORIGINAL: sPzAbt653

Me either, and it annoys me. Along with playing with no reserves. I haven't come to a conclusion if this is a fault of scenario design or TOAW in general.

Ok;
1. Staging units behind the line before an attack
a) I like to have units staged- but all over the place with inconspicuous deployments. If they just sit right where the offensive is going to be, the other guy is going to figure out what you're doing and move HIS reserves accordingly.
b) TOAW in general doesn't feature a lot of prepared attacks on static lines. Where these happen they happen at the beginning of the scenario and everything is laid out for you. Thereafter it's a lot more rewarding to exploit than it is to prepare a new attack.
c) I have the impression that, in the scope of a half-week turn, units won't be all nicely lined up together for more than a brief moment.

2. Reserves
a) Obviously in any small scenario units which aren't fighting for a prolonged period are a bit of a waste. If something goes wrong there are normally spare units on the line you can juggle around. A unit might come off the line, sure, but next turn it'll go back in.
b) In large scenarios I do routinely see reserves used and use reserves myself when acting on the defence. An excellent example of this is my recent AAR of Fall Grau v2.9. I used armoured reserves pretty effectively both to make counterattacks and to create a fear of counterattacks- although admittedly I lost the match.
c) If one is in the process of attacking, you shouldn't really be holding anything back. Units might be a couple of hexes behind because they're not needed for assaults this round- but they should be ready to go in as soon as they're needed. If you're making two prongs and you're not sure which one will need the reinforcement, you're probably better off swapping units between the prongs when matters become clear than weakening one or both from the outset.

[golden delicious]

ORIGINAL: sPzAbt653

Yes Ben, the computer CHEATS !

I KNEW IT !! [:@]

I knew the computer cheated- the production multipliers he talks about are pretty well known, and in Civ2 the AI cheats in some pretty brazenly obvious ways.

But that line of code he described is just outrageous. I guess in 1991 that's what game design was like.

[governato]

Ralph,

There is an hour long lecture from Dave O'Connor, (Arjuna), the developer of Command Ops here:

http://www.wargamer.com/hosted/dropzone ... index.html

You are probably familiar with the description of the game, but the audio file on the AI algorithms is quite interesting and the short articles are hopefully useful to improve the AI a bit.