Page 1 of 1

When the AI becomes predictable, why not play solitaire?

Posted: Sat Oct 11, 2003 8:33 pm
by gambler
Note: When I use the term solitaire, I mean it in the traditional board game sense: The player plays both sides.

Now, I don't know how many people on this forum are traditional board wargamers. For those who are, this first bit will probably seem unnecessary, but computer gamers might not realize that most board wargames are played solitaire (if at all... 3/4 of my games aren't even punched and clipped, let alone played). This comes from a variety of reasons. Some people might not have opponents (I now do, luckily) since pretty much any look at a professional sports stadium when a game is being played (except perhaps the Montreal Expos) has more people sitting there than have a copy of a given wargame. Some people just like solitaire better for exploring the battle. They treat it as an interactive history lesson (this is why I still play solitaire).

Some games have a bit of fog of war and can still be played solitaire. For instance, in some games you can't look at any units except the top attack capable unit. In others there's an orders' system that provides uncertainty as to when orders to various formations take effect so they can act on them (much like the Airborne Assault:HttR sounds like it will do). The absolute best fog of war probably comes from the 'double blind' games where each side gets a map board and rules indicate when you can see the enemy. Often these are naval games, but a defunct company called GDW made a few land games, the best of which was probably Market Garden (the Arnhem battles). This latter method however has the disadvantage of not being playable solitaire even if it's the best FoW out there.

So what does this have to do with Korsun Pocket and other computer games? Well, I've seen some people complaining the computer is predictable after awhile (not really a complaint for some, since it's expected). I've also seen numerous cases where otherwise good games are blasted for poor AI, and not played. Now I know there's PBEM and I'll likely get involved in it once I have the system known. However, for those who really aren't into PBEM, and who have played board wargames solitaire, why not give it a shot with KP? It's basically a double blind game that's eminantly solitairable. Sure you'll remember some of the plan of attack for each side as you switch, but no more than any board game and there will be much less visible on the map than any board game that can be solitaired.

Others might say, "I paid $40-50 for this game, why should I solitaire when there's an AI. They should have made it more challenging!" (caveat, some people say it's predictable, others say it's tough... I have actually yet to play it beyond the tutorial). Well, most board wargames cost at least as much, if not up to double the price for a similarly sized product. There's no AI provided with those (except your opponent, or except in a game specifically designed for solitaire like B-17, Carrier, Phase Line Smash).

Now.. I'm being somewhat hypocritical here. You want to know how many times I've solitaired a computer game I'm sure. Here it comes.. 0. However, it's been a long time since I've played a 2D wargame on the computer and I would have done so with my old V for Victory series had I thought of this then. I plan fully to at least give the concept a shot with this game if it becomes too easy against the CPU. Might also do the same with Korsun '44 by HPS. Airborne Assault: HttR sounds like it might be even more solitairable if the AI becomes easy to defeat. In some ways though it still 'feels' wrong to me, even if I've done the reverse (I used to let the CPU play both sides in ToaW while I did something else - I do the same for sports games...). And I'm trying to figure out why.

Maybe it's because at some level I like an autonomous opponent more than a solitaire opponent. With board games I rarely play one all to the end solitaire, even though I like exploring the history and playing with the mechanics. So maybe computer games only attract those that like an opponent and so therefore if the CPU AI isn't adequate or becomes predictable, the game is then not played anymore.

Posted: Sun Oct 12, 2003 1:33 am
by Fred98
There aremany former board gamers here. But the best way to play this game is by PBEM.


If you play against yourself, then you know that the major attack is on the left and the attack on the right is only a feint. There is now way that FOW could hide that. FOW can only hide the exact detail.

-

Posted: Sun Oct 12, 2003 3:45 am
by Rob Gjessing
PBEM is where this game really shines, Im sure you will love it.

I have played the game solitaire though.. infact its the best way to really test a scenario that you have designed. The only problem is that when I play against myself I keep losing!

Posted: Sun Oct 12, 2003 1:20 pm
by BrubakerII
I cannot agree enough on the PBEM comments. This (and most) game(s) are ALWAYS more fun against another human.

I have to admit it took me a while to first take that first leap of faith (ie. interact with others on the internet) but once done it is hard to go back.

Just do it. We look forward to playing you.

PS

Posted: Sun Oct 12, 2003 1:21 pm
by BrubakerII
Joe aren't you in Korea? I am surprised you have time to post. Have fun.

Scott

Posted: Sun Oct 12, 2003 2:24 pm
by Fred98
Nearly there.......I can see the lights out the window :)

Posted: Sun Oct 12, 2003 5:17 pm
by Canuck_jp
I can understand why someone would "solitaire" with a boardgame, given the time and coordination that it would take to find someone with similar interests and schedules, but solitairing with a computer game borders on the ridiculous!! It's like solitairing on a boardgame when the room nextdoor is packed to the ceiling with eager wargamers.

Posted: Mon Oct 13, 2003 3:56 am
by Quint
"I used to let the CPU play both sides in ToaW while I did something else"


I don't think you've quite got the hang of the concept of "games" ! :)

Posted: Tue Oct 14, 2003 12:01 am
by "Joss" Osborne
Quint wrote:"I used to let the CPU play both sides in ToaW while I did something else"


I don't think you've quite got the hang of the concept of "games" ! :)

Actually, this is the best method for testing new scenario designs (though not much fun, I agree, as a "gaming experience"!).

This is something I learned from the venerable Bill Wilder. In testing a new scenario you would want to have the AI play itself over and over several times. A truly "balanced" scenario, when tested this way, will yield mostly draws as a result.


Joss

Posted: Tue Oct 14, 2003 10:45 am
by Rob Gjessing
Right on Joss, the scenarios I am creating and testing involve the following:

Playing each side against myself.
Playing each side against the AI.
Playing each side against another human.
And finally allowing the AI to play each side against itself.

This last one is fundamental in my book. With the Bastogne scenario I am creating, I have found that the Germans are able to capture Bastogne (and win!) every 3rd game against the Allies. Which I think is about right, although Im still fine tuning and testing it.

Posted: Wed Oct 15, 2003 9:33 pm
by gambler
To explain a bit, I often do the CPU vs CPU as a way to decide which side I should play. If the CPU wins decisively as a side consistently, then I'll play the other side. I definitely think that CPU versus CPU play is a good playtesting method, it should give a general idea how the CPU handles the 'basic' situation.

Something that I've been thinking about (and this is off topic for the thread) is that it might be easier to create AI for games that focus more on orders than on setting up that 5-1 attack in a crucial area which will let you surround another unit and then exploit through. That's why I'm somewhat disappointed by various AARs for the new HttR by Matrix Games. It seems like it would be easier to develop AI to create a plan, than one to create a plan and then have to implement all the details like which attack first, what odds, which units to not move to take advantage of the attack, etc. I might be wrong because the playtest AARs have mostly been decisive victories for the player in HttR.

Anyhow, I'll probably give KP a shot PBEM fairly shortly after playing with it some more this weekend.