c# Timed Out error is back :( can't test anything else this evening :(

Timeout happens only for C# bots?

yes, afaik

Yup, it had something to do with them running two queues.

C# bots take a split second longer to load up due to the hooks into the GAC, but when some other bots that are running at the same time are requiring a lot more processing power, it caused that load-up to take more than 2 seconds. If you could see the output for the C# bots when they crashed, the preferred starting armies gets a null value set, then the first command sent to the game host for place armies is the actual string that was supposed to be the preferred starting regions move.

Frustrating when you're trying to figure out what new bugs you've introduced into your code and even the test match crashes :(

-cringe-

Yeah it's painful to watch. And it doesn't help very much when trying to determine if a loss is because the other bot played better or because he got luckier.

At least the problem with C# bots should be solved now, the time limit for picking starts is up to 10 seconds.

i also sometimes get "Timed out" error messages ruining the game mid game (without occuring on the first round).

getting harder to measure the results of recent changes. hope this gets fixed soon.

Trogatog, is that an actual game that you had? This is clearly going wrong somehow. Can you send me the link?

But on a very different note:

I'd really like your guys' help. You're probably fully aware that the finals are really close now and we are already working on new competitions. But we want to start working on a new Warlight AI Challenge as well. This time with different game settings though.

That's why I'm asking you all for suggestions now, because you know the game better than most. A lot better than me at least. The idea was to have the next competitions with more than 2 bots on the same map, probably 4-6 and then do a Free For All game. Fizzer suggested adding wastelands, which seems like a really good idea to me. But we also need a new (more balanced!) map.

So do you guys have any suggestions for new game settings and maps? All help would be most welcome.

i would like to see the same bot getting tested on different random maps, not just with wastelands.

so that the bots would be forced to calculate the strategy based on the initial map layout information instead of being hardcoded strategic guidelines. there will be a higher learning curve for new folks ofcourse, but i would also hope that the folks who worked on bots for this tournament would open source them, or atleast provide improved base bots for others to work with in the future. how many people are re-implementing A* path finding for example? these kind of well known algorithms and functions shouldn't be such of a timewaster for folks who want to focus on the strategic side of things. if folks like 125ch (who didn't have much prior coding experience) can pull off decent bots as it is, imagine what they could do if they didn't have to spend so much time coding the grunt work.

i dont like FFA myself, and real warlight FFA isn't actually much about expansion strategy itself but more about being able to form alliances and stabbing people in the back without getting yourself remembered as such (for future games), which with bots is much harder to accomplish unless you also implement a messaging system standard (public chat or private message) where they can message each other some instructions and the receiving bot is left to decide to use that information or ignore it. and/or persistent memory, to know who usually stabs you in the back and who usually keeps it honorable. although there is not much fun in being honorable when there can only be one, some FFA's in warlight usually end with vote to end, so the guys who got wiped out get a loss, and the ones who voted to end don't. while a win is always more precious.

Edited 5/9/2014 09:22:01

FFAs have a bad reptutation with the community. They were quickly removed from the real time ladder by popular demand. They are much less about playing strategy. Paradoxically most FFAs end up being decided by the worst player rather than the best. Bad players either get kicked from the game or play terribly giving those near them a game winning advantage.

If you want to put lots of bots on a map then maybe you should try team games with 2 teams. They could be decided by 'who gets the least bad teammate?' but that will always be a problem when you put more than 2 bots on a map together. You could try team games but each team is made up of bots all written by the same person. 3v3 Europe is often played between 2 players in that manner and is very dynamic.

Try the strategic 1v1 template! That way I wont have to work so hard writing a bot.

You say you want a more balanced map. I don't think balance is what you mean because any template you use will be balanced because both players start the game with the same things. Imbalanced implies one side has an advantage.

i agree with piggy that co-op would be more interesting than FFA, still, would want a messaging system to abstract the coordination factor.

and i also agree with piggy that i'd also be more personally interested in a more advanced strategic 1vs1 instead of any FFA/co-op. With random maps, added wastelands, etc.

Edited 5/11/2014 23:04:33

even tho i would have to improve my coding skills a lot, i do think it is a good idea to make a random template challenge, or at least a 1v1 challenge with a bigger map and more balanced map. But to me the n°1 thing i would like is to lower the luck factor, 100% is just terrible for a strategic game. 2v2 games seems like a big leap, i think we should stick to 1v1 for now at least, there is still plenty interesting things to develop in 1v1

@piggy: the map being balanced or not doesn't depend on the players or the number of players or the fact that players start with the same picking choice; it is just a characteristic of the map. if we follow your reasonning, lotteries are balanced map...

Hello jim

The easiest way for your site getting the most possible attention from warlight players is by implementing the warlight 1v1 ladder settings.
--> Medium Earth Map
--> 1 starting spot per bonus
--> 7 random wastelands
--> Luck is currently 16% including some rounding luck. Luck can also be turned completely off and people will also be happy.
--> Those cards can be turned off and people will also be happy. (Especially the +5 armies card makes calculations to complex and unprecise imo.)

With those settings there are great variations in games and there is a huge competitive warlight playing base playing these settings. If the 1v1 ladder was just about following the same moves each game it would have died out by now. I think the current warlight AI challenge will quickly die out after the finals (people stop updating their bots) but with ladder settings I think there will always be players keep updating their bot and new players writing new bots as long as you keep it online.

I strongly advise against ffa. There is no competitive warlight ffa playing base. The reason is that in ffa if you get to strong other players attack you so you trying to win the game makes you actually lose. I played some small earth ffa and the way I usually try to win them is by first taking Australia and then sit for the next 30 turns in Australia while the other players kill themselves. If they are smart enough not to go on the offensive then there aren't any fights until someone gets bored and goes suicide into an opponent. Also dead piggy is right about the worst player deciding the game. If you start next to a d...bag who keeps attacking you although you are both losing then you have no chance winning the game.

Lotteries are balanced. What do you think balanced means?

I don't think that we should take for granted that settings which are enjoyable or dull for humans will cause the same effect when coding bots to play. While I don't particularly favour FFA it seems to me that they could be potentially much more entertaining for bots then for humans. If you take quitters and alliance junkies away from FFA it's actually fun :)

If the key is long term interest and creating a player(coder) base, I'd say the following things are crucial:

*Different maps:
Small Earth, Medium Earth, Europe... etc. None too big, none with overlapping bonuses or other complications, but some variety is definitely needed to keep people interested.

*No limitations on starting picks, or Warlord distribution mode:
Right now all the games start more or less in the same way, with both players deadlocked from turn 1. Make every province available to pick (or one per bonus: Warlord style). That way people can code up some elaborate starting routines.

*Standardized, reasonable settings.
Personally I think that all the games should have the same basic setting like kill rates, starting armies, fog settings, etc. Custom settings are great fun for humans, but coding bots to take all of these things into account could be a nightmare. Luck as well: 16% is kind of ideal / standard, 100% is madness. The only exception I could possible see working is the number of starting regions.

Looking forward to whatever you come up with!

Edited 5/9/2014 15:20:03

Hey Jim! Welcome to the WL community :)

Trogatog, is that an actual game that you had? This is clearly going wrong somehow. Can you send me the link?

Yup, this actually happened. I don't have the link anymore, it was a while ago, I just came across the pic in my pic folder yesterday and remembered I wanted to post it. With the game settings the way they are, this can very much happen (100% luck). I think I linked a game a couple pages ago where an opponent stack of 6 attack into a stack of 7 of mine and the resulting armies left were in favor of the opponent.

I want to say this happens all the time, but in reality, they happen in the games we lose, which makes them more memorable. Looking at my history, when I stack armies just so they survive an attack, 75% of the time it works as expected.

balanced (ˈbælənst)
adj : having weight evenly distributed

one territory worth 1000 armies, the rest worth 0....doesn't seem very balanced to me. What do YOU think balanced means?

Maybe piggy means _lotteries_ are balanced because everyone has the same probability on starting on the 1000 army tile.

And 123ch means lotteries _maps_ are imbalanced.

Lotteries are balanced because neither player has a advantage before the game starts. Same as evey other game of warlight. An example of an imbalanced game would be chess where one player starts without a queen.

Ok, i understand the misunderstanding now. You are talking about balanced games while Jim was talking about balanced maps