Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!
You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!
Use Smashboards links to get your gaming stuff and support the site, Super Smash Bros. Ultimate, GameCube Controller Super Smash Bros. Ultimate Edition, Nintendo Super Smash Bros. Ultimate Edition Pro Controller - Switch, a Nintendo Switch Super Smash Bros. Ultimate Edition - Switch or a Poké Ball Plus. As an Amazon Associate Smashboards earns from qualifying purchases.
The algorithm on this is something like
Didn't realise garpr used trueskill. I did read a pretty interesting write up on some of the maths behind trueskill, it's pretty cool. Forgot most of it though lol. Wonder how good trueskill is at dealing with the disparate nature of our scenes though. Would want to be able to have SD stay above Gerd say even if they only play once a year.I think the Microsoft trueskill rating sitting behind garpr doesn't work in the same way that Chess type ELO does. I'm not 100% sure but I think if a high skill player (with correspondingly low variance) plays and beats a newcomer with default skill and high variance, nothing much happens to the old players rating except that perhaps their variance shrinks a bit.
Here is a write up http://research.microsoft.com/en-us/projects/trueskill/details.aspx
Also you can play around with this calculator: http://boson.research.microsoft.com/trueskill/rankcalculator.aspx to experiment with what happens when you introduce a fresh player to a stable ranking. Run the default tournament 3-4 times you'll see skill estimates asymptotically approach some values. Introducing a new player and having them get bodied (set last place to 25 +- 8.33 i.e. fresh) doesn't change the asymptotic approach of the higher tier players.