I threw together a histogram of the total games played for players on the win rate leaderboard. It looks like the majority of the players on the leaderboard have less than 200 games. I think a 100 or 150 game minimum would probably be best, but the minimum could probably be lower if something like @Flaxative 's idea was implemented.
I think a minimum of 50-100 might work just fine, as long as you don't count matches vs the AI. Actually, you might just as well have a minimum rank which has to be met in order to be listed on the win% leaderboards. I think both measures combined together will both prevent new players from topping the leaderboards b/c of crazy win% fluctuations and also measure win % in a more reliable way (winning vs AI, even cheatotron, is relatively trivial)
I hacked an optional parameter into the live board so you can test out different min game counts, like so: http://cardhuntermeta.farbs.org/winrateleaderboard.php?min=100 Have a play around with it if you like and let me know what a good default would be.
That's an interesting idea, but I think it's much cooler to see your placing on The win rate leaderboard than on A win rate leaderboard.
Setting the number of games to 1 brings up a list of perfect players whose names I don't recognize. Cool stuff!
Following Vakaz's lead (though my charts aren't quite as nice)... Minimum 100: Minimum 150: Minimum 200: 250: 300: Personally I don't think the data starts to get interesting until at least 200.
Actually... It's sort of there. Each player request via the API now returns that player's rating, but I don't want to hit the API from the [META] server every time someone visits a player profile page, and I certainly don't have a path to storing the player rating on [META] to show on leaderboards etc. So, the most viable integration right now would involve a clientside JS call to the API to pick up a player's rating, and only on the profile page. Since I haven't worked with JS much lately this'd take me quite a while, and would be of limited use, so I'm not really in a hurry to do it.
Is there anyone who would NOT prefer it to be 100 rather than 50? I vote for 100 also because it's a nice round number
Yeah, 100 sounds like a much more reasonable default value. It is pretty trivial for a good player to get a 80%+ win rate for her 50 first matches (which has proven to be an unsustainable win rate in the longer run) so a more realistic picture of player ability is currently sacrificed for temporary new player thrills. Which might be reasonable depending on one's goals with the statistics. Many might argue for even higher limits. 200 matches might not be unreasonable at all.
Another "maybe someday" suggestion: Since in a 4 league grouping (the norm atm) a league is repeated every 40 hours, if the scenarios page had a 48 hour period instead of 24 to clear data, it would allow the accumulation of all the data for the 4 leagues in a given cycle over the 2-3 weeks that league cycle is active rather than holding only the data from that day's leagues.
Oh smart! Done. Old data has been discarded already but it should start piling up now. I also raised the required games for win rate leaderboard to 100.
Yeah, I've been monitoring number of league games played for a while and it always was annoying to have to re-check twice a day to compare them. Great idea and glad to see it implemented.
Something I've been thinking about: would it be reasonable to have a scenario that confers the ability to recruit players to a guild, without making them the guild principal?
I've asked Farbs for this... he said he didn't want to do it because it would make the Sorcererers grow even faster!
Just saw this in the forum. How this Guild thing works? And also how does the standing system work? It seems I got negative values since this standing system launched. It kept almost unchanged even i climbed up the elo from 800 to 1700.
