Seeing the graph randomly fluctuate by up to 200 points makes me surprised the top players have much more stable Elo ratings in comparison. I suppose this is mainly due to them playing opponents of comparable rating, which reduces luck-driven fluctuation range.
Another takeaway I had from your graph is that the normal tendency to attribute big drops or gains to catchy narratives like "he lost confidence in his play which led to a downward spiral" -- which I'd be tempted to conclude from the first half of the chart, if I didn't know the data-generating process behind it -- is really unnecessary, from an Occam's razor perspective. It really can be *all* luck-driven. I think this supports your recommendation to focus on process over results.
Two differences between top players and the simulation: 1. Top players have a lower K-factor than the player in the simulation, and 2. Top players have a lot of draws (the simulation has no draws).
The aspect of luck that I often find the most frustrating is not being able to control how well your opponent plays, those grueling weeks when it seems like no one will gift you an easy win and you've got to grind for everything while ending up on the losing side of a brilliancy or two.
"After gaining 100 rating points they will probably conclude that all their training is paying off and write blog posts explaining how they did it." This is exactly right, the humdrum of chess Twitter.
If you play too much, you will start rocketing until you burn out, then you will plummet. I've seen that very often to addicts of local tournaments. So 1 tourney is enough to not lose your OTB expected performance without the need to play every event.
Poker has a deck where no player knows what order the cards are in. Chess has a position where, most of the time, neither player completely understands all the imbalances. The two situations are not that different from each other. The better player usually wins, but sometimes the worse player gets lucky.
Seeing the graph randomly fluctuate by up to 200 points makes me surprised the top players have much more stable Elo ratings in comparison. I suppose this is mainly due to them playing opponents of comparable rating, which reduces luck-driven fluctuation range.
Another takeaway I had from your graph is that the normal tendency to attribute big drops or gains to catchy narratives like "he lost confidence in his play which led to a downward spiral" -- which I'd be tempted to conclude from the first half of the chart, if I didn't know the data-generating process behind it -- is really unnecessary, from an Occam's razor perspective. It really can be *all* luck-driven. I think this supports your recommendation to focus on process over results.
Two differences between top players and the simulation: 1. Top players have a lower K-factor than the player in the simulation, and 2. Top players have a lot of draws (the simulation has no draws).
The aspect of luck that I often find the most frustrating is not being able to control how well your opponent plays, those grueling weeks when it seems like no one will gift you an easy win and you've got to grind for everything while ending up on the losing side of a brilliancy or two.
"After gaining 100 rating points they will probably conclude that all their training is paying off and write blog posts explaining how they did it." This is exactly right, the humdrum of chess Twitter.
Great post that matches my real life experience so closely.
If you play too much, you will start rocketing until you burn out, then you will plummet. I've seen that very often to addicts of local tournaments. So 1 tourney is enough to not lose your OTB expected performance without the need to play every event.
Thanks for your article FM ;)
I gotta say, I laughed out loud when I saw the title of this article, but after I read it, it really makes sense. Thanks for another great article!
Poker has a deck where no player knows what order the cards are in. Chess has a position where, most of the time, neither player completely understands all the imbalances. The two situations are not that different from each other. The better player usually wins, but sometimes the worse player gets lucky.
Well said!