Rating freefall.

that's not freefall that's just a few losses, I went from 1600 bullet to 1300 in a few weeks and only just now started recovering my ratings..

*****Confessions of  Tilter - A Cautionary tale*******

I reached my goal of geting to a 1000.  5 games later I lost then I went on a tilt bing and lost 14 IN A ROW! Got no sleep.  Was pissed, snarled at my wife and kids teh next day. They  got pissed at me. Went back the next night and lost more, stayed up too late. again. Snarled at family again. Wife begged me to take stop playing chess. Son is laughing at me.

My rating is now mid 800s. Took me a year to get to high 900s and "lost it all" in 2 nights tilting.

Interseiting to note:  I was so desperate to win I wasn't really playing chess. My accuracy dropped 4 percentage points which was the difference even in sub 1000 BlunderLand.

In my first response on this thread I suggested that the game does this to you. I realize, obviously, that I play worse sometimes and better other times, but I used to let losing streaks really get to me until I realized one day that losing streaks and winning streaks might be partly a design by this site. One way I determined this was, I started to investigate the history of the players I played and realized that many of them had super high ratings in the past. Some as high as 2000! Of course I’m going to lose to those people. And then when I had winning streaks I discovered those players had pretty low ratings in their history. You know what else I discovered? Whenever I was doing a lot of the lessons on the site, I would be put on what I call the winning streak channel—determined again by investigating the history of my appointments. Again this is all hypothetical on my part, but one thing it did for me is, I have never been stressed by a losing streak since this notion occurred to me. I just think to myself, “HMM…I guess I’m on a ‘losing streak channel’ now, lol.” Whatever you think, this is probably how I would run a Chess site if I had one. Because it encourages people to do the lessons and it’s an efficient and clever way to help people improve.

The only real way you can sort’a track if you’re getting better or worse is if you play opponents that you’ve played before. Even then, that’s all dependent upon how often you play compared to them, do they have coaches, etc.

So what I've noticed is that if my rating is in freefall, like I've just lost 10 games in a row, the site will start matching me up with unusually lower rated players, like 200 points below what I'm usually matched against. My initial guess why this was the case is because if people keep losing for 20 games or something they'll just get fed up with the site. But that is as much of a guess as anything. I've noticed the opposite too. If I'm on a winning streak eventually I'll start getting really high matchups like plus 100-150 points. Again I just assume that they programmed the algorythm to reflect theories on how to make a compelling game experience by making things just hard enough to keep you playing. I do agree that it sure seems the matching algorythm is not just a straight elo system. In my case I just assumed it was to encourage user engagement.

Another interesting thing is that when you look at those higher rated players a lot of them have that crazy high score because they opted to start at a super high rating and then lost a bunch of games. BUT sometimes they did use to have a legitimately much higher ELO like maybe 300-500 points. I am totally guessing but maybe those players are weighted in some way when the program determines matches.