Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
Add description of playing strength and difficulty levels to AI tooltips #419
The AIs installed by default in this game are listed by alphabetical order of
AI names and tooltips are translated properly.
The playing strength of AIs relative to each other has been measured by
Hover AI, Nexus, SemperFi and Turtle AI were not tested.
I have considered trying to improve upon Prot's results with an AI
Wins, losses and draws can be stored in an AI's JSON file, as below:
If such data are available, AI tooltips list them as percentages in
More detailed descriptions of difficulty levels can also help players to
Just like the playing strength data, they are optional. This ensures
All AIs share the cheats defined in data/mp/multiplay/skirmish/rules.js:
Nexus cheats outside of multiplayer games for all difficulties except
TurtleAI behaves like NullBot. So does HoverAI, but since it has only a
If an AI cannot be changed by a player, its difficulty tooltip is shown.
Challenges no longer show the "Difficulty locked" tooltip:
Players who join multiplayer games will also see AI difficulty tooltips:
Note that this PR contains new messages that will need to be translated.
The attached ZIP file contains
Looks pretty neat. Will test some later time.
Unfortunately, there are no accurate tournament ratios for current SemperFi-JS. The ratios you used were from how it was in 3.2.3 and if you try current SemperFi-JS against NullBot, Cobra, or BoneCrusher! a few times you'll easily see it doesn't lose 86% of matches, 1v1 or teamed, on T1 no bases. The most accurate would be the 2018 variant, however, that has a lot less games played with. :/
Bots can build units as soon as they can research the necessary components. It's up to the bot maker to choose what rules to follow, or invent, for their unique bot.
Garbage character(s) are printed before the difficulty tips, both when players join a game or the player is in a challenge when using autotools and optimizations with GCC 9.1. Tested on O1, O2, and O3. No optimizations don't exhibit this.
Building with Cmake and GCC 8.3 causes something else. The cmake build keeps corrupting the tips to something different when switching between a skirmish and a challenge, and repeating the real tip by whatever player number the AI is, which was mildly interesting.
Other than that everything else seems to work as expected.
Thank you for testing! Could you check whether the errors still occur?