Frank, have a break. Your posts on page 10 and 11 of this thread have been pretty wild to be honest. You delve soo much into side topics and sometimes misrepresent points. It has become very tiring and annoying to discuss with you. I give you 3 examples what I mean.
(1) The "circular argument"
FrankJones wrote: ↑06 April 2026, 23:03
ChiefPointThief wrote: ↑06 April 2026, 20:02
I think you are still missing his point. The gap is cut in half after only 15 games in aes and 30 games regular not 100. Which is literally nothing for luck based games. To give an example I recently had a bad game stretch in captain flip losing 7 of 10. Going from 408 to 333 which is 17th to 144th all time. The difference in chess is that you aren't going to lose x amount of games based off of luck. I've seen games in which my opponent made suboptimal moves all game and still won
this isn't possible in chess.
And I will continue to point out the obvious circular argument being made:
You are saying Elo in these games is severely flawed. Yet, you are utterly convinced that a player with a 408 Elo should actually have that Elo, an assumption which you then use to prove your argument, when in fact, one of your primary premises is contradicted by your conclusion.
This seems to be your main gripe in the discussion about changing the K factor. But you misrepresented or misunderstood the point here. The core of the complaint is not that the original high Elo must be correct and kept for all times. "
you are utterly convinced that a player with a 408 Elo should actually have that Elo" That is
not what is written above in ChiefPointThiefs post, it's something you made up yourself. Instead, the main point is that people notice if Elo moves so rapidly then it can never be correct at all in high-luck games. People being people, they notice the speed of Elo changes more strongly when their Elo goes down rather than up. If you have a closer look at the complaint posts, it's mostly not that people assume their highest ever Elo is their correct Elo, it's rather people say, "WTF is that?" To give you an intuitive analogy: Imagine you have a weighing scale (a kitchen scale) and some mass laying on it, and the scale displays some weight value. Whenever you add a thing that weighs roughly 100g (or 80g to 120g), the display changes by 300g, whenever you remove a thing of roughly 100g from the mass, the display changes by 300g, too. Clearly we observe that the weighing scale is broken. We don't even need to know what the true value of the original mass is or even what the exact true masses of the smaller things are. Just the fact that the weighing scale display changes so quickly is enough to conclude the weighing scale is broken. Now exchange "weight" with "Elo" and "weighing scale" with "Elo rating system" and you have the argument being made. When a measurement value changes much quicker than the underlying construct to be measured is changing, then the measurement value must be off. And that would be mitigated by changing the K factor.
FrankJones, by concentrating so much on the idea that people want their high Elos preserved, you are ignoring the strong version of the argument. The weak version is, "High Elos should always stay high", which is stupid, yes, I agree. But we should ignore this rather than making it the center point of the discusssion.
(2) The "why 2 players's Elos should be preserved" argument
FrankJones wrote: ↑05 April 2026, 16:20
You are simultaneously saying "Elo here at BGA is inaccurate" and "It's very important to preserve the Elo of any 2 players at one snapshot in time."
That "2 players preserve their Elo" thing was only an absolute side aspect of the simulation I made. In the simulation, I fixed the true Elos of both players at 300 and I wanted to measure accuracy. Clearly yes, as a trivial consequence of those fixed true Elos, it is better (= more accurate) when Elos are preserved. I didn't say the (possibly inaccurate by BGA) starting Elos need to be preserved. I thought it should be obvious that I care about the true Elos, not the starting Elos, but you make it sound that starting values fixed at true Elos is a problem in my reasoning because in my reasoning, Elos on BGA are inaccurate. The point is I could have easily set inaccurate starting Elos in my simulation and it wouldn't really change the outcome of my simulation. The fact that I need to explain this is annoying. The fact that the starting values of both players was the same was not very important. It was only the same because I wanted to create equal conditions, the 2 players are equal in win probability, true Elo, and starting Elo, yet their calculated Elo values deviate strongly many times. That is literally the definition of mismeasurement, two exact same conditions lead to widely different measurement values, and the degree of mismeasurement is stronger with higher K factors. You have not engaged with that point, you haven't really engaged with my simulation.
Instead you make statements like,
"We need to keep in mind that "fluctuate too much" is in itself a vague, subjective phrase. What constitutes "too much"?" Man, "fluctuate too much" is not vague in the context of our discussion. I relatively precisely showed in the simulation what level of fluctuation amounts to what level of inaccuracy. It was clearly shown that when true Elo does not change in a small series of games (which is approximately a very realistic and usual scenario), then calculated Elo is very inaccurate a lot of times. Now if you wanted to say that people have different tastes in preferred accuracy levels, that's a very bold statement. Why would people want less accuracy, I don't get it? The job of Elo is to measure skill, that is it's definition. If you personally want Elo for other purposes, that is fine but then you should spell out what these are and you shouldn't assume that other poeple do not care if Elo doesn't represent its label. People hate wrong labels.
(3) The "100 equal game results should lead to equal Elo" argument
dschingis27 wrote: ↑05 April 2026, 07:21
100 games are way to few to make a definitive statement about skill or Elo for that matter. Every competitive Backgammon or Can't Stop player can tell you that stretches of 100 games are usually not representative, both mathematics and simulations show that there is a high variance in win rate in only 100 games. If player A had 200 wins out of 300 games before and player B had 100 wins out of 300 games before, and then they have equal performance in 100 games, every sane person should assume that player A is much better than player B.
You bring this up again and again saying that 100 equal results should lead to equal Elo. I repeatedly said that this is not the main point anyway and if you look at the above quote, it was just an illustration of the actual real point I made in the sentences before. My point was that larger historical series of games are more accurate than smaller recent series. The fact you can refute my illustration doesn't refute my point, you haven't refuted math and simulations. (That's what I mean when I say you are so much into side topics.) You haven't engaged with my point. Instead you create hypotheticals and details in the illustrative example so that it doesn't work anymore. Yes, of course, if player A had weaker opponents at first or player B learned a lot, that explains the scenario. That is obviously not how I wanted the example to be understood. I could also add "what ifs" to make it obviously right, like what if player B had much more luck than player A in the 100 recent games? Shouldn't we consider this possibilty?, to use a wording you used yourself. We can go endlessly through different hypotheticals. But I need good-faith readers to not fill in details in my examples that don't have a relation to the point I was making. You found one situation where my example doesn't work but that doesn't add much to the general discussion. I know you wanted to refute my indeed dramatic statement that "every sane person should think A is better than B" and you did that. But that deflects so much from the point about Elo and K factors that we were discussing.
To give you another example: I could say, "There is a lot of animal life in the pond." and you could respond by saying, "But not when the whole pond is completely frozen." That means you made a correct statement - but no one really cares because the main statement wasn't about that, it was about another greater thing, the observable lifeliness of the pond, it wasn't directed at the frozen pond. You don't refute biology saying there is a lot of life in the pond by saying "but not always", in the same way you don't refute a general statistical point about Elo by saying, "but not always".
Extra:
You write so many stories from low-luck games that also don't add anything. The complaints about Elo are not targeted at these games like Earth.
