Originally Posted by Melvear
Again, I don't get this:
c4tuna's stats (just cause he posted and has stats close to mine): Dominion: 2010 [203W/187L] (52%)
Here are mine: Dominion: 1704 [174W/145L] (54.5%)
And the difference in ELO is quite significant, while Waterd103 and Nekro, despite a quite large gap in win rate, have barely any difference in their ELO.
I have suspicions that normal elo (and by extension the dominion elo which most likely uses the same system) has certain tweaks that ranked games do not have, but they're all under the radar since you can't actively check your normal elo and it isn't common knowledge that you can check your dominion one.
Possible "tweaks" include how it matches you up/rewards Elo for team/duo queues, putting you in higher elo pools (and therefore affecting your elo gain) for significant win streaks, even normalizing for typical elo shifts at different times of day. I doubt Riot wants to make most of it public, so we're going to have to speculate for the most part.
That said, probably half or more of these dominion wins on this account were pre-30 so there could also be different ways to calculate those numbers, or different pools for high-elo low level players. And if more "serious" players play draft than blind pick, having a higher draft elo would be more meaningful since the average skill level would be much higher (this is all speculation, but keep in mind that elo is based around a theoretical "average" at 1200 or thereabouts depending on inflation/deflation).