Boards›Maps & Modes

Hey Data Analysis guys at Rito, I have a very humble request

Z Statistic·8/30/2014, 8:17:34 PM·7 votes·874 views

I think it would be really, really interesting to have a 3D cumulative distribution plane with the following axes: x gold difference at, y minute, and z proportions ending in victory on SR. It would show what the probability of victory would be given only that the enemy team is up by 5k at 13 minutes. You guys can do that for 50k games, right? It could really shed some light on just how snowbally the game is for the community or maybe not?

My hypothesis: the probability of victory given "high" gold at time before a certain breakpoint K, is significantly higher than 50%. However, the same gold or even a lot higher gold after a certain breakpoint L is not significantly different from 50%. In plain English: snowballing only works for a certain period of time after which it actually comes back to an even game; it's basically a temporary advantage is always checked (barring ending the game "early") in the long run.

4 Comments

DocRedgrave8/30/2014, 8:32:56 PM1 votes

My head hurts.....

Kyleman1438/31/2014, 2:44:06 AM1 votes

I like this idea the problem is i do not believe many people would understand how the whole system would be working... But getting to know at what point a certain snowball plateaus would be absolutely wonderful

RocksCanFly9/1/2014, 11:26:16 AM1 votes

While the math on that looks interesting (and I would like to see it done anyway) I dont think it is as simple as gold wins games. The level cap is 18, any advantage throughout the earlier stages of the game stops at the point where you can't get any higher but they can. Minions give gold and xp, more gold means more xp, higher levels and gold earlier when the advantage is still effective can have a profound impact on the game. But as anyone who has urgot knows, your momentum can drop off rapidly even before 18.

You can only buy six items, have all the gold you want but the most expensive build I have takes 19600 gold to complete, the moment I get that, my power curve drops off like a stone as they catch up because my effective advantage cannot go higher.

Different champs peak at different times, pre six LB is laughable but level 4 pantheon is a nightmare, some champions with a more utility R like olaf are pretty weak 6 and 7 but climb back up to scary at level 8.

I think that your hypothesis is really only relevant to a few champs (hypercarries like vayne and fiora mostly) that peak on items rather than levels as some champs will power on through, good cs or not, and some will fade by the end of midgame even with a perfect cs.

Absolutely love the idea of seeing some data though. <3