Talk:Elo rating system/@comment-178.128.101.153-20130108213113/@comment-80.221.9.199-20130110035943

The first 50 games have higher elo impacts. It's possible to have the same win ratio but a considerable difference in rating, and vice versa. Average ingame stats don't matter in this regard.

for example: 2 guys with a skill level of 1400

- the first one wins and gets luckily carried in most of 50 games, popping his rating somewhere near 1800, then in the next 100 games he loses most of them, due to being matched agaist better opponents. He ends up in 1400 elo with a 33% win rate.

- the second one has some serious unluck in his high impact games, dropping somewhere around 800 with nearly no wins. However, since he is continually placed against opponents under his skill level he manages to win most of the 100 next games bringing him to 1400 rating. His winrate would be 67%.

This is of course an exaggeration and nearly impossible to actually happen.

By playing a large number of games (at least a couple hundred), the system will bring you where you actually belong, no exceptions.

And btw, if you never troll/afk yourself the changes of the troll being in your team are 4/9 or 44% and in the enemy team 55%. This means that for every 4 games lost by your team because of a troll, you will win 5 games because of a troll in enemy team

-> eventual elo gain.