HeroesWorld
  • Log in
  • Create account
  • Mode
    Interface scale

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

Rating systems.
5.0 (1 rating)
Open
Which rating system should I choose? (see the text for details)
33 voters
Single choice
Sign in to participate in this poll.
Poll results will be visible after you vote or when the poll closes.
In reply to AlexeyPank
User Avatar
Offline AmberSoler
Registered:
Posts: 1,394
#363
Auto-translated
AlexeyPank;64759
In other words, as I understand it, you suggest that my bonus should only apply to the linear component?
A little clarification.

Let's agree to consider the oscillator as it is - reducing the rating of players at the bottom of the table is not a bug, but a deliberately implemented balancing mechanism... a "carrot and stick" method, hidden within the oscillator itself. This is not a calculation error, but one of the ways to transfer a player's result to their tournament rating.

It doesn't halve the rating of the last player; it creates a rating for them in the form it should be according to the formula I proposed. It is equal to 1/2 of the non-linear component only because the second part starts from zero. If the second part starts from 10, 20, or 100 points, the effect will be completely different. The formula is just a shell, and the content is up to the rating creator. The oscillator is a general term for a function of several independent variables...

Iron, without delving into the essence of the mathematics, used the term "bug." And no matter how many times I told him that this is just my balancing proposal, he didn't understand. But at least he admitted that he doesn't understand numbers; but surely everyone else doesn't understand either? :) This "bug" can be fixed in any direction, if there is a desire... But in any case, one will get less, and others will get more.

Now, to the point:

1. As I already said, the linear component can start with any number. Therefore, it is possible to set that the minimum will be, for example, 10 points. That's it - there are no problems with the calculation... :) It is also possible to calculate grals in the same way - if a minimum level for awarding points above zero is introduced (as previously suggested).

2. Regarding the bonus. It can be calculated in any way. The only condition is to think about the meaning of this coefficient. After all, every tournament will definitely have newcomers with a zero rating - which means that bonuses will be given to them by definition for any place except the last... If you played poorly in relation to yourself, then you will get a minus (because others came with zero and pulled the points over to themselves, which is understandable).

Set a condition, formulate the task - and it will not be difficult to draw up a formula.
Сначала было слово...
FizMiG v.2.0 *** Очередное обновление справочника! *** http://vk.com/fizmig

0
⚖️
0
In reply to AmberSoler
User Avatar
Offline IronAxe
Registered:
Posts: 1,495
#364
Auto-translated
AmberSoler;64780
He doesn't halve the rating of the last player; he assigns a rating to him in the form it should be according to the formula I proposed.
Iron, without delving into the essence of the mathematics, used the term "flaw." And no matter how many times I told him that this is just my suggestion for balancing, he still didn't understand.
:) Without delving into the essence of mathematics :) I remember someone here saying that logic comes first, and mathematics is secondary. So, explain logically why, if players complete the game in, say, the same number of days, have the same resource accumulation and oil production, but one takes second to last place and the other takes last place, why should the player in last place receive half the rating of the player in second to last place?
I understand that this is a balancing proposal. However, when comparing the results of a linear system and an oscillator, I didn't like this balancing. At least for the simple reason that, in the non-linear component, everyone honestly receives what they earned in days, and there is no need to artificially inflate anyone's rating in one direction or another. Moreover, I provided examples in numbers, both hypothetical and based on real calculations. Therefore, I used the term "flaw" only to shorten my explanations. Yes, for me, this balancing is a flaw. And it's not just about the last place. Look at the real calculations done by Kyle and AlexPank-o. I won't comment on them anymore, as the others continue to remain silent about it :) And I have already given my comments.
AmberSoler;64780

But at least he admitted that he doesn't understand numbers, but does everyone else not understand either? :)
:) I can at least admit something, unlike some. I don't understand complex formulas, because sometimes I don't even want to delve into them, but if necessary, I can understand something. But admit, Amber Soler, who found the error in the formula in your documentation for the third game (it's good and helped me in the tournament because I haven't played in a long time)? I remember you sending me to it. You admitted in a private message that the formula wasn't yours. But how did such a nonsensical formula end up in your documentation if you understand mathematics so well? :)
And why didn't anyone else notice this except me?
AmberSoler;64780

This "flaw" can be corrected in any direction, if there is a desire... But in any case, one will get less, and the other more.
I agree, but why don't you admit that the linear part gives each player as much as they earned in days relative to the leader. No more, no less. What they earned, they get. Is that unfair to you?

And according to AlexPanka's coefficient, there also seems to be an unsolvable situation, when a player with a zero rating receives a bonus after the game, like a starting bonus. Personally, I don't see the point of any bonuses when calculating the rating.
0
⚖️
0
User Avatar
Offline AlexeyPank
Registered:
Posts: 2,396
#365
Auto-translated
Yes, a player with a zero rating will receive a bonus. However, as could be seen, this bonus is not very significant and will not be able to substantially change the overall situation. And over time, when the number of regular players noticeably exceeds the number of newcomers, it will be completely unnoticeable.
In general, this bonus can address the following tasks:
1 Account for the number of players.
2 Account for a player's previous results in relation to new ones.
3 Account for the average rating of the tournament (i.e., if there is one strong player and others are much weaker, the average rating will be lower and, consequently, the bonuses will be small).
I am not arguing that the formula I proposed may be incorrect. But has anyone proposed another one that takes into account all the points I listed?
P.S. And still, it is better to calculate the linear component not from zero, otherwise what is the point for a weak player to participate among the strongest?
Мы спасемся? Я не знаю. Верьте...
Геройский Интернационал
Клан Homo homini - основатель клана
Член РКСМ с 07.07.07.
ФЛУДОВСКИЕ ССЫЛКИ!- сообщить о бесполезной теме/сообщении.
0
⚖️
0
In reply to AlexeyPank
User Avatar
Offline IronAxe
Registered:
Posts: 1,495
#366
Auto-translated
AlexeyPank;64790
Yes, a player with a zero rating will receive a bonus. But, as you could see, this bonus is not very significant and will not be able to significantly change the overall picture. And over time, when the number of regular players noticeably exceeds the number of newcomers, it will become completely unnoticeable.
In general, this bonus can solve the following problems:
1. Take into account the number of players.
2. Take into account the player's previous results in relation to new ones.
3. Take into account the average rating of the tournament (i.e., if there is one strong player, and the others are much weaker, then the average rating will be lower and, therefore, the bonuses will be small).
I'm not arguing that the formula I proposed may not be correct. But has anyone suggested another one that takes into account all the points I listed?
You are absolutely right. But AmberSoler is also right. Everyone comes up with a rating based on the logic of what they want to see and take into account. For example, I believe that the number of players and the place are not so important for playing Heroes for a rating. Especially if you think about the overall rating among online resources. Therefore, everyone is right in their own way.
AlexeyPank;64790

P.S. And yet, the linear component should be calculated not from zero, because then what is the point for a weak player to participate among the strongest?
If it is not calculated from zero, then the oscillator can be completely discarded, in the sense that the non-linear component that takes into account the result in days in relation to the leader will probably be reduced to nothing, so why use it, just leave the linear one and calculate it not from zero.
0
⚖️
0
In reply to IronAxe
User Avatar
Offline AmberSoler
Registered:
Posts: 1,394
#367
Auto-translated
IronAxe;64787
...explain logically why, if players complete the game in the same number of days, have the same resources and oil, but one takes second to last place and the other takes last place...
You are twisting things. In the case you mentioned, two players will take the same place, and their tournament rating for this tournament will be the same.
IronAxe;64787
...It's a different matter when comparing the results of a linear system and an oscillator; I didn't like that balancing.
That's a different matter. It would have been worse if you liked everything. :)
IronAxe;64787
...And admit it, Amber Soler, who found the error in the formula in your documentation for the third game (it's good and helped me in the tournament because I haven't played in a long time)? I remember you sent me to it. You admitted in a private message that the formula wasn't yours. But how did such a nonsensical formula end up in your documentation if you are so good at math? :) And why didn't anyone else notice this except for me?
Yes, that's absolutely correct. The reference book doesn't only contain my calculations. It also contains a lot of information that was taken from other sources - that's why there are links everywhere. It's impossible to keep track of everything and check every number manually. Sometimes inconsistencies pop up. And then I myself, rolling up my sleeves, try to analyze everything in detail. But now, I think, there are practically no such moments left. Why didn't others notice? I suppose they did. They just didn't want to participate in this issue... So they kept silent. And you - well done. By the way, I reworked that formula and made it right.
IronAxe;64787
...And according to AlexPanka's coefficient, there also seems to be an unsolvable situation, when a player with a zero rating receives a bonus after the game, like a starting bonus. I personally don't see the point of any bonuses when calculating the rating.
Not just a starting bonus. Any player whose current rating has dropped due to low activity will receive this bonus regardless of their skill level. But, I agree, the bonus is still not very big. And this is a reasonable price to pay for a more balanced system. You insist on not taking anything into account except the player's result. I believe that it is more reasonable to take into account all factors. And that's enough about that. There is no argument here.
Сначала было слово...
FizMiG v.2.0 *** Очередное обновление справочника! *** http://vk.com/fizmig

0
⚖️
0
In reply to AmberSoler
User Avatar
Offline IronAxe
Registered:
Posts: 1,495
#368
Auto-translated
AmberSoler;64800

Not just for beginners. Any player whose current rating has dropped due to low activity will receive this bonus, regardless of their skill level. But, I agree, the bonus is still not very large. And this is a reasonable price to pay for a more balanced system. You insist on only considering the player's results. I believe it is wiser to consider all factors. And let's leave it at that. There is no argument here.
That's what I said – everyone is right in their own way. You believe that all factors should be taken into account and bonuses given for them, and I believe that some of these factors lead to an incorrect balance, distorting the real results of players. And again, I can calculate a rating for our player based on a game in a tournament on another resource. You are unlikely to be able to do this, since no rating is kept there. What do we conclude? That anyone who wants can come up with their own rating system and, if they wish, keep it and publish it in their own separate thread. The forum rules do not prohibit this. And who will look at which table is a personal matter for everyone, and which system each person will use for their own needs is also a personal matter. Here's an example. Kyle took a non-linear system, calculated the ratings for all clan players, and calculated the clan rating based on this. If he wants, who will stop him from calculating the same thing using the oscillator and also taking into account the AlexPanka coefficient? Therefore, I absolutely agree with you – there are no arguments and there cannot be. Everyone has simply promoted their own rating system in their own way
0
⚖️
0
User Avatar
Offline AlexeyPank
Registered:
Posts: 2,396
#369
Auto-translated
So at the moment we have three options for calculating the rating:
1 - Non-linear (IronAxe's version), here there are two variants: 10% decay and 33% decay, plus taking into account "allowable limits" when playing for speed.
2 - Oscillator (AmberSoler's version), also several variants: 10% decay, 33% decay, no decay, the linear component is calculated from 0 or from any value.
3 - Oscillator+Bonus (my version) - non-linear component without allowable limits, linear component is calculated from 5 points, a bonus is added to the total result, no decay.
All these systems have a right to exist, and time will tell which of them is better.
For now, I will try to calculate the general rating for everyone who played on the portal across all versions. I will calculate it using my system. I'll see what comes of it.
Мы спасемся? Я не знаю. Верьте...
Геройский Интернационал
Клан Homo homini - основатель клана
Член РКСМ с 07.07.07.
ФЛУДОВСКИЕ ССЫЛКИ!- сообщить о бесполезной теме/сообщении.
0
⚖️
0
In reply to AlexeyPank
User Avatar
Offline AmberSoler
Registered:
Posts: 1,394
#370
Auto-translated
AlexeyPank;64862
I will try to calculate the overall rating for everyone who has played on the portal, across all versions. I will use my own system for the calculation. Let's see what comes out.
All the experiments are reasonable. However, it makes sense to conduct them only on a statistically significant sample. For now, we don't have enough material, and we won't be able to obtain reliable results. A couple of tournaments won't help you realistically assess the pros and cons of all rating calculation systems.
Сначала было слово...
FizMiG v.2.0 *** Очередное обновление справочника! *** http://vk.com/fizmig

0
⚖️
0
User Avatar
Offline AlexeyPank
Registered:
Posts: 2,396
#371
Auto-translated
Here are the results of all Tournaments calculated: /topic/at/64866/. Of course, there isn't much information yet, but we should start now. I suggest everyone take a look at the results.
Мы спасемся? Я не знаю. Верьте...
Геройский Интернационал
Клан Homo homini - основатель клана
Член РКСМ с 07.07.07.
ФЛУДОВСКИЕ ССЫЛКИ!- сообщить о бесполезной теме/сообщении.
0
⚖️
0
User Avatar
Offline Kyle
Registered:
Posts: 2,153
#372
Auto-translated
to IronAxe: I've been meaning to say this for a while, but I keep forgetting – there's another drawback to your system... it can't be used to calculate a rating for a tournament like "Arena." Because in this type of tournament, the player's placement is what matters most.
This relates to your plans for:
IronAxe;64793
Especially if you think about an overall rating among online resources.

Best regards.
0
⚖️
0
User Avatar
Offline Qwerty
Registered:
Posts: 2,090
#373
Auto-translated
Guys, no offense.
1. The idea of a "decaying" rating is good. But we need to face reality. How much worse can a player perform after a quarter of inactivity? Maybe 10 percent, probably no more. After a year, it's probably around 40 percent. Results older than 15 months are not taken into account in the rating.
So, the decay coefficient is 10% per quarter. And it should be based on the time factor, not the tournament factor.
2. You can't see the forest for the trees. There are 13 pages in the topic, but it's not clear how the rating is calculated.

Орден Феникса – харизматический :) лидер
0
⚖️
0
In reply to Qwerty
User Avatar
Offline AmberSoler
Registered:
Posts: 1,394
#374
Auto-translated
Qwerty;64892
Guys, no offense.
1. The idea of a "decaying" rating is good. But we need to face reality. How much worse can a player perform after a quarter of inactivity? Maybe about 10%, probably no more. After a year - about 40%, it seems. Results older than 15 months are not taken into account in the rating.
So, the decay coefficient is 10% per quarter. And it should be based on the time factor, not the tournament factor.
2. You can't see the forest for the trees. There are 13 pages in the thread, but it's not clear how the rating is calculated.
The rating is not calculated yet. Or rather, there are several methods. The final one will be determined over time...

And regarding the rating decay - the player's skill is not a factor here. :) It's simply tied to the player's activity. There is no doubt that a strong player will remain strong. And the "rating" in this case is a synthetic parameter that takes into account both the player's abilities and his activity at the same time.

P.s. After a couple more tournaments, you'll be like "dear Leonid Ilyich" :)
Сначала было слово...
FizMiG v.2.0 *** Очередное обновление справочника! *** http://vk.com/fizmig

0
⚖️
0
In reply to Kyle
User Avatar
Offline IronAxe
Registered:
Posts: 1,495
#375
Auto-translated
Kyle;64877
to IronAxe: I've been meaning to say this for a while, but I keep forgetting. There's another drawback to your system... it can't be used to calculate a rating for a tournament like "Arena." Because in this type of tournament, the player's placement is what matters!
The Prophet told me this at the very beginning. So, I don't count Arena towards the rating. Arena is a different type of offline tournament, as far as I know. It involves gathering troops and fighting. Has there been a lot of Arenas on this resource? If a player is participating in an Arena on another resource, I simply won't update their rating. Look for drawbacks in other systems; I've even shown them to you. The oscillator won't work in the Arena either because it uses one of the
0
⚖️
0
In reply to IronAxe
User Avatar
Offline AmberSoler
Registered:
Posts: 1,394
#376
Auto-translated
IronAxe;65005
The oscillator won't work in the arena either because it uses one of the formulas from my system.
I'm about to snap...:smile19:
Сначала было слово...
FizMiG v.2.0 *** Очередное обновление справочника! *** http://vk.com/fizmig

0
⚖️
0
In reply to AmberSoler
User Avatar
Offline IronAxe
Registered:
Posts: 1,495
#377
Auto-translated
AmberSoler;65007
I'm gonna crush someone...:smile19:
:smile17:
Or, is it not like that? :)
0
⚖️
0