Closed fsmosca closed 5 years ago
What about an option like:
supported_matches:
casual: "bot"
rated: "human"
We can extend but for now will work
That is fine. But of course if you can do it, I welcome a human playing casual or rated.
In this case if the human can play both modes then just leave: casual: "bot"
, if the option supported_matches
is not present then the engine accept bots and humans
That is Excellent. Thank you.
Download and test from here: https://github.com/tttony/lichess-bot/tree/bot_matches
I have tested a rated match human to bot with: rated: "bot"
and it was declined, so it seems to work, test and let me know
Bot is busy currently, I will try it later and let you know the test.
I got a single hit so far.
Decline Bullet rated challenge from BOT XxGODxX(2223)
I am using,
supported_matches:
casual: "bot"
I am waiting challenge from human and see its behaviour.
With the following setting,
supported_matches:
casual: "bot"
All human challenges on both modes were declined.
With,
supported_matches:
casual: "bot"
rated: "human"
Human rated is accepted but human casual is declined.
@fsmosca it's implemented in https://github.com/careless25/lichess-bot/commit/81566b4dc1010a0bc316f57e0496e12e2bbd95a5 You just have to apply https://github.com/careless25/lichess-bot/pull/81 also.
@fsmosca Oops! Right! There are non title players, fixed that: https://github.com/tttony/lichess-bot/tree/bot_matches
@fsmosca Can I close this as this was already implemented in 81566b4
Feature request: My bot is not designed to play against a bot, it is designed to play against humans. It is weak around 1800 CCRL 40/4, and I am not improving its strength instead am trying to make it weaker.
There are bots that challenges it on rated mode which I don't really like as I can test engine vs engine matches on my pc at anytime, so I have enough engine vs engine games. Bot that challenges it in casual mode is just fine of course.
Another idea could be to develop a lichess bot protocol altogether where bot and api will communicate. Example if there is a challenge to my engine named my_engine the api will send to my_engine,
challenge type bot name eeee rating n mode rated
where: type is the type of player can be [bot, human] eeee is the name of the engine challenging my_engine n is the rating of the eeee engine can be [none, integer] mode can be [rated, casual]
note that,
challenge type bot name eeee rating n mode rated
can be expanded to include tc, example
challenge type bot name eeee rating n mode rated tc b32
where: tc value b32 can be a code, b32 = Blitz 3+2 b52 = Blitza 5+2
Or other coding system design, above is just an example.
my_engine understand this as eeee is challenging me and it is of type bot and has rating of n on a rated mode. With this info I can decide whether I will accept the challenge or decline it.
or if a human challenges my_engine, the api will send,
challenge type human name hhhh rating n mode rated
The bot my_engine can reply with,
reply accept challenge type human name hhhh rating n mode rated
where: reply [accept, decline] would determine if the game is a go or not.
If reply value is accept, the api can proceed the match.
If value is decline, api will inform the challenger,
reply decline [some message]
This is just a general idea details is totaly different thing. Perhaps adapt existing rules in fics and probably improve it further.