Posts from Player rating for tournaments across all installments of Heroes.

The difficulty here is that a player can participate in one Tournament but not another. This leads to a stratification of the player base. And the true strength of the game can only be determined in relation to the leaders of a particular version, right? But the main factor here will not be the place taken in the Tournament, but the result, how close it is to the leader in terms of the number of days spent completing the game.

Maybe this is only a problem in the initial stage, and then the coefficient will be made up of the Tournament rating (which is determined, for example, from the 5 best players in this Tournament)

My point is that if I'm not playing Heroes of Might and Magic II, and you're not playing Heroes of Might and Magic III, then I don't care how our strength ratings compare. We don't have a way to verify this in equal conditions (on the same map). Would a chess player be interested in the rating of checkers players if they don't play checkers themselves? No, they wouldn't...

I checked it in my head... Nonsense! You can't reduce the second parameter proportionally to the first... it's wrong... there's no direct relationship because the first depends only on the leader's result, and the second depends on the number of participants...

I agree, but I'm just afraid that you're leading to a comparison of players from different versions...

Added 1 minute ago
There's no problem, though - these players already have a rating, and at the moment it's the same for everyone, so this doesn't interfere with calculating the final formula... I'm just pointing out that this parameter will be dimensionless, i.e., fractions of a whole - for example, 0.75 or 1.3... And it won't be related to the nominal rating of the leaders (which could be 150 or 1500)...

Let me explain: the parameter that determines the strength (qualification) of the tournament will be calculated, for example, like this:

the ratio of the average sum of the ratings of the five strongest participants in the tournament (regardless of their placement) to the average sum of the ratings of all participants in this tournament. The higher the coefficient, the higher the tournament rating.

Either sum the ratings of all participants and divide by the number of participants. I don't know beyond that, but the figure is fair, though I don't know what to do with it—I'm not great at math :(

I'll also mention the pros of the system. After all, tournaments can be based on a main objective for resource gathering or experience, where the duration is equal. Well then, just plug the amount of experience or resources into my formula instead of days and calculate the rating further.

2 IronAxe.

I suggest simplifying the situation for the start. Let's forget about the oscillator for a while and use a purely nonlinear coefficient for calculating the rating (Example No. 1). We will also consider it from 0 to 100 – no concessions for underdogs. We only round down to the nearest whole number, with a minimum rating of 1, i.e., we don't set zero...

Now, let's calculate the tournament rating based on the average rating of the top five players. For the first tournament, the rating will be 1 (because the ratings of all participants are currently equal). Then, we will recalculate it based on the results of the first tournament.

Let the first tournament be the recent BB. We need statistics. It doesn't matter that this tournament is not included in the VK – no tournament is included there yet... :)

Why shouldn't the rating be zero? Zeros represent ghost players, one-time users, who have permanently left this resource, etc. They will automatically be removed from the leaderboard when their rating reaches 0. Or are you suggesting that we should remove everyone with a rating of 1 from the leaderboard?

That's exactly what it is – the average rating of all players. And what percentage of them are the strongest? That's why I'm saying that dividing the average rating of the strongest by the average rating of all players... We'll get a number that's just a little over 1. This will be the parameter, and the higher it is, the stronger the players in the tournament... It's simple for mathematicians :)

We don't assign 0 for participation in the tournament, even if the calculation results in 0.05. The minimum is 1.

That's for mathematicians, but this is mine. Remember, the VDV condition was simplicity; it suited me – I quickly derived the formula – one divided by the other, but I understood, and still understand, that it's not complete. I agree with your calculations. I think you'll bring it into a proper form, but if they ask me to calculate, I'll need a table in Excel from you, in case I can't manage it. I somehow crammed my simple formula in there.

No problem. We already have a ranking of 26 players... Let's see what we can do with a real example, not a virtual one... Question: which scale should we use? 1-100 or 1-50? In essence, there is no difference, only 1-100 is more convenient and clear (psychologically more familiar and directly related to percentages), in 1-50 you would have to enter decimal digits after the comma, which I would not like. I suggest 1-100. I think no one will object...

I suggest moving all three results from the normal difficulty without coefficients to the impossible difficulty. Their save wasn't easier. We've already figured that out.

I wouldn't mind if you took on the task of calculating the ratings; I can help or fill in if needed :) Because I'm already starting to get lost in the complex formulas, like in a triathlon :)

No way, if you've decided to start calculating, do it yourselves :D No need to shift the burden onto others. I'll help with the calculation part and make a template for you. You'll only need to copy and paste the numbers there...

I'll calculate for all 26 participants. We'll see from there. As for the two saves in the tournament—I think we can overcome that! Or we simply won't count those who play on a low difficulty level—we already decided that...

In the same place, someone will eventually have to age it, not counting the current rating if they missed it—it's quite a hassle, so you'll be consulting according to your template then. I might send mine so you know what explanations to expect :)

And if they already have some rating, will it be impossible to age it? Damn, there are still going to be misunderstandings :)
Anyway, we'll think about aging as we go; for now, there's nothing and no one to age :)

- What is the aging rate?
- What is the adjustment ratio? 1/3 is too much, in my opinion... let's make it 1/10... it will reduce the rating of an inactive player by almost the entire amount over a year, but it won't be as painful for a missed tournament.