Interesting. I haven’t read the article too closely, but the way it converges is interesting.
Clearly I was still stuck in the ye-olde days mindset of ELO where losses essentially had to be directly offset by wins. This seems much more robust, especially since losses from past restarts are (presumably?) no longer in the rating period, and it can just converge from where people’s scores are. (even if I’m still traumatized by accumulating upwards of 20 losses to oracle + zero alone this week )
Wait what is even happening…
Why am I 5. all of the sudden and @n-sanders is going down. I was not expecting that, season 3 really is weird.
I’m quite sure the cause is that the top ten are now getting matched up with people from the top 10-20
and whoever wins against those is going up and if you lose your going down.
Up to this point, I’ve managed to avoid following the match history of any of my algos (just to avoid the stress of it!). But having fallen so far, I had to take a look.
I’m 63 - 36 over the last 99 matches right now, with a string of losses to some algo “families” in the 1900s.
My turtle cave algo has great matchups with the bulk of those in the top 10, but it’s achilles heel is becoming it’s’ downfall.
Ugh, at this point I may just need to stop following it’s performance and just see where it ends up on Monday
Edit to add: despite the shuffling, 9 of the top 10 are still the same. @IWannaWin has joined the group with LuckyR slipping to 12. Looking back at the first leaderboard freeze, 9 of the 10 from then are still in the top 10 now (LuckyR was in with Michigan_Difference out)
A dedicated front-line emp strat completely unravels me. The Madrox boss is an example. I build my defenses forward enough that they can pretty much all be reached behind a front-line wall, and I procrastinated adding any building adjustments in favor of other tweaks during the last few days of the competition.
The EMP nerf has reduced the number of algos that commit to such strats and since I have decent match-ups against most of the remaining EMP mazes (as they generally build row 11 before 13), I neglected to fix the front-line vulnerability.
Quite a lot of shuffling going on in the lower ranks of the top 10! Looks like it will come down to the wire for some players with less than 24 hours to go.
Looks like matches have stopped.
So … @RegularRyan - any particular rules for ties? If it’s alphabetical then I might have to rethink my emoji naming schemes
I’d guess the leaderboard defaults to sorting by algo id (you’re 80065 to my 80522). So I could see them going with that as a recognition for the “older” algo.
Or perhaps they’ll break the tie with a head-to-head match (pretty sure all your algos have been named Mazer, so not sure how my turtle does against this one).
Interestingly enough, given the odd/even group format that comes next, I think one of the groups has an extra two wins in it for my turtle algo (if I’m remembering my match history right). I won’t bother saying if it’s the odd or even group as I’m not looking to influence the decision by it.
We mentioned cases like this would be handled at C1s discretion
I’ll check the match history and see who’s algo wins the 1v1, and manutally increase their glicko2 score by one point so its reflected on the leaderboard
Wait … is the Leaderboard for S3 final ?
I was floating at #17 place … and havent updated my algo for at least a month … just checking the forum and waitihg for the new season.
Now it is on 9 place … what happend