really cool work, honestly I would suggest casting a wider net and including Diamond and maybe a tier below as well, think for low playrate champions your results may not be statistically significant or may be skewed by a few mains who spam a certain skin

The main reason I didn't include diamond is because it's a bit more complicated to crawl diamond players versus challenger/master. But feedback received, it's definitely not statistically significant for some of the uncommon skins (only 35k matches total, or 350k sightings of being played since there are 10 selections per match). I'll try to add that wider net this week.

no worries at all, like I said this is really cool/impressive haha. used to be a casual player and now working with statistics every day so figured I'd comment :)

HFT is still a massive thing in volume weighted (we call them pro-rata) markets, and it's even more toxic for retail as lots of people submit large orders that are unlikely to get filled immediately (which retail doesn't have the $ to do) to secure a bigger share of the pie

yes you are re-inventing the wheel, the CME has pro-rata markets: e.g Soybeans and Wheat where the matching engine is not FIFO

there is, $GOOGL stock

we have so much oil and natural gas, we're going to transition off of petroleum before we run out

And coal. Massive amounts of coal which could be gassified in a pinch if there were nothing else.

hah, disagree strongly tbh, there may be an adversarial engine that specifically finds positions that stockfish doesn't evaluate well. they did the same to AlphaGo:

https://arstechnica.com/information-technology/2023/02/man-b...

True that's a reasonable possibility. But in go the top engines are far from perfection but it's not clear in chess

He actually didn't lie lol, they knew they just didn't care as he was a very profitable customer

"Website blocked due to malware Your Malwarebytes blocked this website because it may contain malware."

It's back up for me now in NYC

Down for me in NYC as well