Albert Aranye

Meta || data, games, AI

[Albert Aranye @AlbertEzem]

(Finally) launching xGRAPM 😅

The only live-updating public player impact metric in football.

Plus match forecasts, player comparisons, and Bayesian-stabilized per-90s across Europe's top 5 leagues.

Live now. Link below.

[Albert Aranye @AlbertEzem]

@FPL_Armo @fpl__core I remember you were doing research into price rises, did you ever finish the series?

[Albert Aranye @AlbertEzem]

@Dr0P1loT @HalfSpaceMC caicedo better as a deep playmaker than enzo??

[Albert Aranye @AlbertEzem]

@bineymos @Eigengrau674 @Loyt_t it wouldn’t hurt sure but he already did it when they first broke up and the internet will always peddle the false narrative around, so when is he meant to stop? plus they are still exes, it would be different if they had a different relationship

[Albert Aranye @AlbertEzem]

@V2_Romario if the traditional big 6 is still allowed to compete i dont think theres any permutation of players outside the top 6 that could win it. defo wouldnt give this team higher odds than arsenal/city

[Albert Aranye @AlbertEzem]

@FPLProp a no haaland challenge 🫩

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @RansomArsom @Finchinator2 so, are we disagreeing? smogon took a step in the right direction. finch said bots are an existential threat, i said that’s a bit strong and there’s a playbook here, one that smogon is following. i’m not sure what i said that you have an issue with

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @RansomArsom @Finchinator2 yes, and bots should be prohibited competitively. to enforce that, Pokemon probably needs some bot detection + controlled bot environments so people don’t need to cheat. or even better, provide bot tools yourself, and maybe use them to educate new players.

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @RansomArsom @Finchinator2 true, pokemon does have more board states than chess, and a host of other differences too. do any of these differences imply pokemon will die a sad death when chess is doing fine?

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @Finchinator2 oh i missed this. i’m aware of jax - that was my reference. jax beat michael 5-3, i think? deep blue beat kasparov 3.5 - 2.5 20 years ago. this is *barely* superhuman. in chess, bots are superhuman, like, they don’t lose to humans ever. poke bots really aren’t good by comparison

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @RansomArsom @Finchinator2 pokémon doesn’t have more possible move choices per turn than chess. it does have a teambuilding phase with more choices, but no bots exist that can teambuild from scratch yet.

a bit rude of you at the end, but i’m free to type this out for others who may see this thread

[Albert Aranye @AlbertEzem]

@RansomArsom @TTV_LuckyCharms @Finchinator2 and btw, 2 is the important bit. pokemon is difference from chess ofc: one has pokemon, the other has pieces. but ive seen a lot of ppl talk about the differences and no one talk about whether they matter. and its funny because meanwhile, the Pokemon bot creators all copy chess!

[Albert Aranye @AlbertEzem]

@RansomArsom @TTV_LuckyCharms @Finchinator2 1. define ‘best move’. the nash equilibrium strategy? the move associated with the highest reward (RL)? not sure there’s a significant difference regardless

2. if there is a difference, is Pokemon more susceptible to bots because of it?

[Albert Aranye @AlbertEzem]

@TTV_LuckyCharms @Finchinator2 … no, pretty sure that’s not true. AlphaZero is a superpowered AI bot capable of learning, and chess is fine.
in fact, Pokemon bots aren’t even superpowered by comparison - humans don’t beat chess bots, period. the chess equivalent to current Pokemon was 20 yrs ago

[Albert Aranye @AlbertEzem]

@Moritzva2 im anime only, but wasn’t it only in frieren’s era that humans/elves became strong enough to fight back? so about 100 years ago. and historically humans have hunted prey that could fight back, so not that unusual

[Albert Aranye @AlbertEzem]

@mmmjkphm @BenisKen lol i was thinking of that, and there are definitely *some* that would argue AI is a moral actor.

it’s like the duck test: if it walks like a duck and talks like a duck, it’s probably a duck. so there’s a dissonance between behavior and narrative portrayal

[Albert Aranye @AlbertEzem]

@zlats4ever @Paeblo32 @dumkenee 34/41/35 > 36/26/36, and won UCL + never fumbled league. try again

[Albert Aranye @AlbertEzem]

@jonathanstoop in a sense, everything reduces to data tbf - the best process is in itself data. but more to your point, there are definitely edges to be found even with ‘traditional’ data, football clubs aren’t that efficient to claim that they don’t exist

[Albert Aranye @AlbertEzem]

@zlats4ever bro has already joined silky to start cooking excuses 😭