Suggestion: Chess Clock

[thespaceinvader]

The only issue that strikes me is how to determine who's won in the event of the timer elapsing. In chess, you do this by adding up the value of taken pieces AFAIK. In wesnoth, this doesn't work so well.

Usually, it would be fairly obvious that someone was in the lead (though there's no in-game means to assess this AFAIK). But that doesn't take into account stalemate situations, or situations where a player has a clear advantage, but places their leader in a vulnerable situation late on.

[Dixie]

Maybe I'm missing something, but that statement seems false. If you want to control total time without affecting the maximum number of turns, could you not set a large "Initial Timer" and then take away turn and action bonus?

Edit: OK, I just checked in 1.8 and the maximum "init timer" you can set is 25 minutes. I am not sure when that changed, but the limit was much higher in some previous versions.

I think for the main idea could be implemented without adding any new features, (ie. relatively easily, enough if I lack the technical knowledge to say for sure) it would just tweaking the actual timer. Instead of capping at 25 mins (1500 secs), it could cap at an hour or two (3600 or 7200 secs), which should be enough to accomodate any game that's gonna use a chess clock, I think. Another such required tweak would be the minimal turn bonus: atm it's a minimum of 10 seconds, which, while not extravagant, is contrary to the idea of a chess-clock (0 turn bonus).

The team-shared timer is a nice idea too, but that would probably require additionnal coding which I am not qualified to evaluate.

[pauxlo]

The only issue that strikes me is how to determine who's won in the event of the timer elapsing. In chess, you do this by adding up the value of taken pieces AFAIK. In wesnoth, this doesn't work so well.

The one whose timer elapsed is the one who loses, I would think.
As timers only run alternately, they can't run out at the same time.

[thespaceinvader]

That would appear to penalise the person doing better, if anything. If you're winning, you'll have more terrain to cover and units to manage. Your timer will run down more quickly. Surely the way then to win would be to more or less ignore the time, recruit a bunch of the solidest units you were able to, and sit in your keep whilst the other guy tried to track you down. If anything it could promote turtling, rather than quick play.

[Dixie]

But luckily, the present timer includes an ation bonus option! I'm not 100% clear on what actions trigger that bonus (any action, just attacking, etc?), and you may use it or not, but it's there to help in such situations.

Edit: it's for attack, recruit and capture; not simple moving around.

[eyerouge]

The only issue that strikes me is how to determine who's won in the event of the timer elapsing. In chess, you do this by adding up the value of taken pieces AFAIK. In wesnoth, this doesn't work so well.

Usually, it would be fairly obvious that someone was in the lead (though there's no in-game means to assess this AFAIK). But that doesn't take into account stalemate situations, or situations where a player has a clear advantage, but places their leader in a vulnerable situation late on.

You're right, it's no easy or obvious nut to crack.

My own take on it would be to give the users a few customizable options about what type of penalty would be dealt to the person who's total time ran out. Two of them are insta-loss and random army punishment (x% of units die, or all have their life reduced with 80%, etc).

I just came up with something that's perhaps more interesting: The punishment could be that you only get x% of the time you would normally get per turn (if that is enabled that is). That's an interesting way of using a chess timer.

About the evauation of the end-game
This is a cool topic. There is actually already enough of variables in BfW to make simple evaluations possible, yet pretty meaningful and interesting. Gold cost of every unit is already around. In many ways it has a parallel to how the creatures in chess are valued when trying to determine (from that alone) who is leading or not.

An example would be a formula that could somehow look at, compare and/or combine one or several of the following by, perhaps, giving each "a score" and then adding it all up together, and then comparing with the opponents:

• Each players total army worth in gold in play right at that time
• Ratio between your total damage dealt and received during the whole game
• Same, but only the last x recent turns.
• Number of leveled up units
• A reduction from the total "score" for the player that used up his/her time first.

Now, I am not saying these are the ones to go with. They're just simple examples. One can think what (s)he wants about them. They're not very innovative, but in the end, the priciple is the same for all such solutions.

I also am not suggesting that any formula in the world can predict the future and be correct always. I believe many mistake what this is with the function of having a digital oracle that sees into the future. A formula can never be made superbly accurate, but, it can fill it's function and be what it's supposed to be: A way of declaring a winner in the case where time runs out for one of the players. Notice there's a huge difference in declaring one, and claiming that person would have won if the game continued. In best case scenarios, and with a good formula, that declaration would however be correct more than 50% of the times. If so, then the formula is on the right track. (One would know if testing this empirically or by watching replays, or making simulations maybe maybe)


Personally, I don't have an issue with the above type of solution to declare a winner as long as it's reasonable and a type of factors that are often and generally relevant for the outcome of a game of BfW, even if we can never foresee the future or factor in all variables we'd want to in the best of worlds.

I would be happy if there was some type of decent formula and if that could be an option the players could choose to activate when creating the game with a chess timer. I would almost be equally happy with any other, less sophisticated, punishments for running out of time.


exploiting
You suggest that one could exploit the timer by doing nothing or next to nothing for all the game, just waiting for the other persons timer to reach zero, so he/she'd lose/get punished.

I'm not sure a person that really does nothing or close to nothing would actually ever stand a chance against a rational player that took his/her time and actually attacked or played aggro in some way at all (i.e. capturing the other other players villages). I guess your fear could be seen between two newcomers to the game, where one exploits and the other just sits there and is clueless and, for some reason, doesn't even scout or attack in 20 turns straight But is that whom one should design a game around? And how is this a problem really? Wouldn't that person learn from the experience, as with any other experience in BfW? I myself have no issues with people that pretend to play but just stare at the game as if it was a screensaver - if they enjoy it, then good for them.

But, let's go back here and suggest that your worry is valid - how can it mended?

Well, one way would be to build a function that checks to see if the other player is actually alive and kicking it. In FPS there are often anti-camp mods that can be loaded: When a player camps on the same spot or even larger area, he/she would start losing HP. Maybe something similar in that direction could be devised and activated when playing with the chessclock.

I have a hard time being constructive here as I'm mentally stuck in the assumption that this would very seldom be a problem with people that actually chose to play with the chess clock. That coupled with the notion that a player that does nothing but run away and avoid confrontation will surely lose the game makes it seem like more of a theoretical worry than one that captures how it would almost always play out.

I might be wrong, I don't know.

Sidenote: While I played competitive chess we used to declare a person the loser if his/her clock ran out. A person could not exploit that in chess, since he/she'd lose against anyone that could think, and I see little difference from that and Wesnoth (granted one scouts if FoW is enabled)

[tekelili]

I havent read all the the posts on this issue, so sorry if I repeat something already said. Just 2 quick coments:

1- I could be wrong, but I had intended that most professional chess games are played now with an electronic clock that adds some time each player turn, because is considered unfair that a player that has virtually won a game just lose because he needs 10 moves to check mate his oponent and currently has just 3 seconds of time reminding. This leads to the fact that chess games could last for a lot of time if a game takes a lot of turns. Games havent a fixed time limit.

2- In chess the minimal amount of time you need to move all your units is the same for both players (they just have to move 1 unit). In wesnoth this kind of clock would be unfair for cheap factions like orcs that have to move more units each turn than expensive ones like drakes.

[Yogibear]

For me, the question is this: Is it more important to have an interesting and challenging game being ended in time or do we want to keep it that way?

If the answer is yes, then a chess clock like it was originally suggested is perfectly valid. It is just an additional element of strategy to have your opponent reach the time limit first (and at the same time invest enough of your own time to not lose before).

If the answer is no, but you still want some kind of planning, i suggest to have a fixed time limit for a fixed number of turns: For example 90 minutes for the first 20 turns. After that 20 turns, you get 90 minutes again for the next 20 turns (either with or without preserving what is left from the previous turns). IIRC, it works the same way in chess, just that there it is the first 40 turns and the rest of the game. Which makes sense, because in contrast to wesnoth, the number of units in chess constantly decreases and so normally an endgame is less challenging than a midgame.

[ancestral]

You might consider adding an option to this feature which adds incremental time, somewhere between 1 and 5 seconds each turn. For many games, the longer the game, the longer the turns become.

[sanaris]

Lets talk about common time manner in turn games as ~60-120 minutes per game with some low ~60 sec bonus per turn.

Existing system is almost identical to chess, just need to set turn bonus to ~60, and init time to big value - and you will get it.

Trouble is, wesnoth is not a chess/

So, there is no time measure being valuable as much as in chess.

Action bonus is good try to achieve the goal. For moving unit, attack, and anything else could be rewarded with time. But time prices are matter of map, race, etc.
Really, as already was proposed, it could be implemented more internally in some markup, so you can script the game events.

So the suggestion is:
INCREASE maximum possible init time in menus, please. Because its too small ^^ 1500 sec = 25 min.

It will allow chess game clock implementing.

OR
Remove time feature at all, because its not usable in time common manner. Instead, implement game events more. And implement some bindings. Players are free to use any other source of balancing - like web timers, timer programs etc.