Sometimes, I get to the upper 1400s, occasionally 1500s, and once I got to 1602. However, my peak is often followed by a fairly rapid descent, to the lower 1400s or even worse. Please note that this is nothing to do with sandbagging. I never resign with, say, just a piece down, and I never try to manipulate ratings. Is my lack of form likely to be weariness, or sickness, or.....?
When your rating rises, it is either because you actually got stronger, or you just got lucky. If you didn't get stronger, the natural tendency of the rating system is to self-correct, bringing you back down where you belong.
There's a blog post on this specific phenomenon.
lichess.org/@/CheckRaiseMate/blog/is-chess-a-game-of-luck/kDSzCwLu
You will always fluctuate based on luck with opponents, how you're feeling, whether you're actually improving or getting worse at chess etc. My purely speculative theory is that as you reach higher Elos you feel happier and spend less effort on each game so you sink back down, so you start to focus more on each move so you reach higher Elos, and so on.
lichess.org/@/CheckRaiseMate/blog/is-chess-a-game-of-luck/kDSzCwLu
You will always fluctuate based on luck with opponents, how you're feeling, whether you're actually improving or getting worse at chess etc. My purely speculative theory is that as you reach higher Elos you feel happier and spend less effort on each game so you sink back down, so you start to focus more on each move so you reach higher Elos, and so on.
That's why rating deviations are never beneath 45 - there is always a certainty of 95% that your current "true strength" is somewhere plus or minus 90 points from your displayed rating.
Don't know...but I like the title!
@coledavis said in #1:
> Sometimes, I get to the upper 1400s, occasionally 1500s, and once I got to 1602. However, my peak is often followed by a fairly rapid descent, to the lower 1400s or even worse. Please note that this is nothing to do with sandbagging. I never resign with, say, just a piece down, and I never try to manipulate ratings. Is my lack of form likely to be weariness, or sickness, or.....?
I think this effect may be explained by the pseudo-random opponent selection algorithms implemented in the pairing scripts.
> Sometimes, I get to the upper 1400s, occasionally 1500s, and once I got to 1602. However, my peak is often followed by a fairly rapid descent, to the lower 1400s or even worse. Please note that this is nothing to do with sandbagging. I never resign with, say, just a piece down, and I never try to manipulate ratings. Is my lack of form likely to be weariness, or sickness, or.....?
I think this effect may be explained by the pseudo-random opponent selection algorithms implemented in the pairing scripts.
@coledavis said in #1:
Repeat after me
"Peak is not the average" "Peak is not the average"
Got it? You cannot play at your peak all the time, you will bounce back to your average.
Same effect happens if you hit the floor. You will bounce back.
Repeat after me
"Peak is not the average" "Peak is not the average"
Got it? You cannot play at your peak all the time, you will bounce back to your average.
Same effect happens if you hit the floor. You will bounce back.
@Alientcp
Repeat after me: "don't be condescending".
I understand the difference between peak and average. However, if you read a little more closely, you will see that I seem to travel between peaks and troughs in a fairly dramatic way. My performance rarely seems to include a period hovering around within confidence limits (statistical term).
Repeat after me: "don't be condescending".
I understand the difference between peak and average. However, if you read a little more closely, you will see that I seem to travel between peaks and troughs in a fairly dramatic way. My performance rarely seems to include a period hovering around within confidence limits (statistical term).
@heallan The opponent selection algorithm that you suggest is an interesting theory.
@coledavis said in #8:
> @Alientcp
> Repeat after me: "don't be condescending".
I am not being condescending, but it apparent that you are clearly confused on the actual meaning of the terms, or like in this case, the consequences.
> I understand the difference between peak and average. However, if you read a little more closely, you will see that I seem to travel between peaks and troughs in a fairly dramatic way. My performance rarely seems to include a period hovering around within confidence limits (statistical term).
See this graph?
ibb.co/37m5F29
That is your level from about jan 2022 to date. You can clearly see that your average is something around 1420-1430.
100 points above is quite a lot. You will always violently bounce back because you are way out of your league. There is no soft landing from there. Way above your league.
Though, when in free fall, most of the times you go way below your average because a very specific reason, if you climb too high, and you dont belong there, you are raising the average, you need to go below the average for about the same period of time then climb up, because if you dont go below, the average will present a rise, which you cannot maintain.
It is always that way when you climb too high and you are in fact not getting better. You will rise way above, and crash way below.
ibb.co/MGJyvKT
But the point is this. You have an average , whatever that might be. If you go way to high and you dont belong there, there is only 1 way to make equilibrium to get your average correctly. To be below, if you dont go below, your average will rise, meaning you got stronger. And there are mostly 2 possibilities, When you are on the top, either you crash and burn as quickly as you went up, and then it normalizes or, after the rise, you stabilize a bit in a certain point, then you go slightly below for a prolonged time. In any case, whatever time you spent on top, you need to spend below to make equilibrium.
That is the difference of peak and average.
> @Alientcp
> Repeat after me: "don't be condescending".
I am not being condescending, but it apparent that you are clearly confused on the actual meaning of the terms, or like in this case, the consequences.
> I understand the difference between peak and average. However, if you read a little more closely, you will see that I seem to travel between peaks and troughs in a fairly dramatic way. My performance rarely seems to include a period hovering around within confidence limits (statistical term).
See this graph?
ibb.co/37m5F29
That is your level from about jan 2022 to date. You can clearly see that your average is something around 1420-1430.
100 points above is quite a lot. You will always violently bounce back because you are way out of your league. There is no soft landing from there. Way above your league.
Though, when in free fall, most of the times you go way below your average because a very specific reason, if you climb too high, and you dont belong there, you are raising the average, you need to go below the average for about the same period of time then climb up, because if you dont go below, the average will present a rise, which you cannot maintain.
It is always that way when you climb too high and you are in fact not getting better. You will rise way above, and crash way below.
ibb.co/MGJyvKT
But the point is this. You have an average , whatever that might be. If you go way to high and you dont belong there, there is only 1 way to make equilibrium to get your average correctly. To be below, if you dont go below, your average will rise, meaning you got stronger. And there are mostly 2 possibilities, When you are on the top, either you crash and burn as quickly as you went up, and then it normalizes or, after the rise, you stabilize a bit in a certain point, then you go slightly below for a prolonged time. In any case, whatever time you spent on top, you need to spend below to make equilibrium.
That is the difference of peak and average.
This topic has been archived and can no longer be replied to.