Connection ErrorBad RequestThis page went staleSession ExpiredSign in requiredForbiddenNot FoundToo Many RequestsServer ErrorBad GatewayService UnavailableGateway TimeoutHTTP Version Not Supported
Session Expired
Your session has expired. Please log in to continue.
Sign in required
You need an account to do that. Sign in or create one — it only takes a minute.
This page went stale
Your security token was rejected, usually because the page sat open too long or you signed in elsewhere. Nothing was saved. Reload the page and try again.
Request Failed
We encountered an issue while processing your request.
Please check your internet connection and try again.
Error Details:
Share Link
Posts from Player rating for tournaments across all installments of Heroes.Auto-translated
If a player didn't participate, what should be taken into account? I don't understand.
A player who skipped should have zero points taken into account, while those who played should have a bonus added based on the second component of the oscillator. After all, the number of participants in the ВК tournament is n, and the number of participants in each stage is also n!
But why don't you come up with a logically and mathematically sound mechanism for dividing... In my opinion, this is an unsolvable problem, so either leave it as it is or abandon it altogether...
Hooray:smile03: you finally admitted it yourself! I dare to suggest that you abandon this idea! Or create two rankings... but with no aging!
I dare to give you some advice: take a piece of graph paper, a pen, and experiment with a dozen abstract characters in a dozen abstract tournaments... And then tell me what observations you made about accounting for their results... It's a very interesting exercise, I tell you... :)
You might not believe it, but I did just that... it was after this that I realized that this task has no solution!
What has been proposed is currently the optimal option, taking into account the interests of all players...
An unsubstantiated statement!
Justify it. And what does "not needed yet" mean? Does that mean it will become "needed" later? And what needs to change for that to happen?
Later, it may turn out that it will be needed. Time must pass... By the way, maybe instead of introducing an increasing coefficient, it would be better to dynamically change the "Scale" (i.e., the maximum number of points for a tournament).
HOORAY:smile03: you finally admitted it yourself! I dare to suggest that you abandon this idea! Or create two ratings... but no aging!
Let me tell you a secret - just between us - I was never a supporter of aging, as such. Initially, I considered the goal to be building a balanced calculation for the tournament rating. I proposed and continue to propose that it be taken into account in the form of a cumulative annual rating... Then - reset it. This is much closer to me because I don't like "nitpicking" :) The current rating reflects exactly such a not very flattering comparison... Cosmic races, in a word...
But my colleague is a staunch supporter of the current rating calculation method. I admit that the idea itself is not bad and can also be considered - for those who like it. Therefore, at one time, I also took on helping him develop and balance this method. That is, my opinion is neutral. I am simply, in a sense, a co-author of this calculation. No more, no less. Whether to apply it in practice or not is up to specially trained people to decide. :) So you don't need to persuade me... Kyle;63436
You might not believe it, but I did... it was after that that I realized that this task has no solution!
Exactly. If the goal of the rating is to reflect the "Player's Strength," then the current rating does not do this under any circumstances. Simply, at the moment, of all the options that come close to assessing strength, the resulting current rating is relatively close. All other options are much further away.Kyle;63436
By the way, maybe instead of introducing an increasing coefficient, it would be better to dynamically change the "Scale" (i.e., the maximum number of points for a tournament).
2 IronAxe Here's a small comment regarding your "lyrical" approach. I was considering the situation where out of one hundred players, one player completes the map very quickly (they pulled off a clever trick – the original source is here) – let's say in 10 days, the second takes 1000 days, and all the other 98 are even worse. The example is exaggerated, just to illustrate the point. In this case, all 99 people, compared to the leader (who will have 100 points), will receive a rating from 0 to 1, which is disproportionate to the effort they put into competing for the prize.
And I told you how I would avoid this situation. I will calculate each player's rating up to 7 decimal places, then each tournament rating will be added to each player's old rating, taking aging into account. There should no longer be zero due to the presence of an old rating. However, if this does happen, there will be numbers after the comma. In the final general table, multiply each result by 10 or 100. As a result, proportions will not be violated, and the current strength of each player will be correctly expressed in numbers, without any distortions associated with linearity.AmberSoler;63330
You – I believe – simply didn't grasp the idea behind my example. Naturally, I don't think you intentionally distorted it... In your comments, you omitted the first "fast" player from the analysis and considered all the others, so in that case, the very essence of my example is lost. Let's read on: This is incorrect on your part because it is the oscillator that smooths out this imbalance ...
The oscillator actually introduces an imbalance. Regarding the last players, I already talked about it using an example – the last one has their tournament rating halved. You all like to say that we should look at real examples, so go ahead and look – many people have already downloaded your Excel spreadsheet. See how Izualu's rating was reduced by half. Now let's look at the top places: -------- Days -- Nonlinear -- Oscillator Dmitry 36 -- 100 -- 100 Radik 39 -- 92.3-- 93.9 VDV 56 -- 64.3--77.6 .. Axe 71 -- 50.7--64.0
Don't you notice anything? The oscillator straightened out my nonlinear rating for Radik by as much as 1.6 :) But then, for VDV who followed, the oscillator only smoothed out my nonlinear rating by a total of 14.3 :) And what's the logic there? Why is the player who took second place with a small gap from the leader being penalized in this way? And why did player VDV perform better, taking third place with a large gap from the leader in terms of days? Well, and further down, everyone has been nicely "smoothed" out; it even almost corrected mine by about 14. I have already shown two specific flaws using a real example. Show me a flaw in my nonlinear column on this same example ???
Regarding the second component, it's good to calculate. Without considering the nonlinear component. There, yes, it doesn't matter how many days I completed the map – but if I happened to take second place after 1000 days, give me a bag of grails, but what does the tournament rating having half the value of the leader have to do with it if the result in terms of days is far below average?!
Added after 13 minutes AmberSoler;63398
Well, that oscillator corrects this flaw... :) What else should we develop? In the oscillator, there is no such flaw - I showed this flaw in IronAxe's methodology solely as an example of inefficiency... So, if you are adopting the oscillator, then you shouldn't think about random discrepancies. This sometimes happens, but it will not affect the final calculation... Take the calculator that I suggested and conduct an experiment. Second place, regardless of the gap from the leader, will be rewarded with at least 50 points, the rest – evenly from 0 to 50.
The oscillator doesn't correct the flaw; it pulls up a mediocre player by the ears based on their rating to someone who is actually good, if it turns out that second place was taken by a player with 1000 days. But in a single tournament, this is still somewhat acceptable, but what does it have to do with people who didn't play in the tournament and suddenly become neighbors in the ranking? Even aging by 33% won't create such a mess if calculated purely based on the nonlinear component :)
And I answered you how I will avoid this situation. I will calculate the rating of each to 7 decimal places...
Seven decimal places or ten — the result will still be close to zero regardless. I wanted to ask if it was even worth fighting for SUCH a rating? :) I think not. In my calculation, I give a BONUS for high placements. The consequence of this is that a few people at the tail end lose... Yes. No one is without sin... :)
As for all other questions — we are just repeating ourselves... I see no point in starting the polemic anew. I ask you not to distort the facts and my thoughts. For some reason, you missed this point in my previous post and are once again comparing our respective calculation methods.
I wrote to you about your lyrical opus, that you are referencing me there and distorting my words... This is incorrect. What does the rating have to do with it? Forget it. I had no intention of reviving this old discussion. We have finished discussing calculation methods.
Let me tell you a secret – just between us – I have never been a proponent of aging as such. Initially, I considered the goal to be building a balanced calculation for the tournament rating. I suggested, and still suggest, that it be taken into account in the form of a cumulative annual rating... Then – reset it. This is much closer to my liking, because I don't like "nitpicking" :) The current rating reflects exactly this, not very flattering, comparison... Cosmic races, in a word...
No, you say – should a player's rating always increase, and if it decreases, then only due to aging?
That is, I'm not an expert, but is that how it is in real life? If a player plays well and suddenly shows poor results, doesn't their rating decrease?
Added 4 minutes ago AmberSoler;63443
Seven digits after the decimal point or ten – the result will still be close to zero. I wanted to ask if it was worth fighting for SUCH a rating? :) I don't think so. In my calculation, I give a BONUS for high places. As a result, several people at the bottom lose... Yes. No one is without sin... :)
But should they (the underdogs) have fought for your linear oscillator? :)
AmberSoler;63443
All other questions – we are just repeating ourselves... I don't see the point in starting a debate again.
This is how tactfully we refuse to comment on the situation with Radik and VDV. Note that we are talking about 2nd and 3rd place.
No, you say – a player's rating should always increase, and if it decreases, it should be exclusively due to aging?
I'm not an expert, but is that how it works in real life? If a player performs well and suddenly shows poor results, doesn't their rating drop?
His rating "drops" relative to other players because of a smaller increase in absolute value. That is, let's say I received +10, and you received +20 points. Both of our ratings are increasing, but mine is decreasing relative to yours. The relative component of the rating is important, not the absolute one...
What difference does it make if my rating dropped or increased compared to the previous value of my rating? What matters is how my rating has changed compared to other players. That's it. Everything else is irrelevant. How can you not understand this? We've repeated this so many times.
Added after 14 minutes IronAxe;63445
This is how tactfully we refuse to comment on the situation with Radik and VDV. Note that we are talking about 2nd and 3rd place.
I have long ago stated my conclusion that you do not understand the meaning of the oscillator... Its purpose is to partially compensate the player for falling behind the leader if the player took a high place, and vice versa.
The second participant did not fall behind the leader, and there is nothing to compensate for – he is almost at the maximum. The player who took third place has a significant gap. Your non-linear calculation hits him very hard. But he took third place (!). And the player receives compensation for taking a high place. And if a player with the same result took last place, he would receive a significant minus for taking last place. The meaning of the oscillator was created precisely for this!
Because the place matters regardless of the result. Can you even understand the meaning of several words in a row? I don't even know what to tell you anymore...
His rating "drops" relative to other players due to a smaller increase in absolute terms. That is, let's say I received +10, and you received +20 points. Both of our ratings are increasing, but mine is decreasing relative to yours. The relative component of the rating is important, not the absolute one...
Correct. Let's say I got +5, and you got -5. It's all the same. Therefore, the rating does not necessarily have to constantly increase. And even less so that someone else linearly increases or decreases it.AmberSoler;63447
What difference does it make if my rating has fallen or risen compared to the previous value of my rating? It is important how my rating has changed compared to other players. That's it. Everything else is irrelevant. How can you not understand this? We've repeated this so many times.
All players' ratings will change proportionally to their results for completing the map and relative to the leader's results. That's why I don't understand why you don't understand that there are no games with linearity - it's just a sport :)
Added 4 minutes ago AmberSoler;63447
Its goal is to compensate the player for lagging behind the leader if the player took a high place, and vice versa.
Why is there any compensation for lagging behind? You play, and you get what you deserve.
Added 11 minutes ago AlexeyPank;63354
We are talking about how to calculate the rating for ALL parts of Heroes. The VDV Tournament has a convenient property that it includes all versions, so it is often discussed. I am also against aging, because a player who has played one tournament may be pushed down from his position in the second, because his rating has not increased, but the ratings of those who played have increased.
I explained to you how the rating will be calculated in the VDV tournaments. I provided you with additional information about the introduction of permissible limits, which I came to after a discussion with AmberSoler, although he says that I did not understand him correctly, but that is not important, the main thing is that the flaw was identified and, if possible, smoothed out. Regarding the aging of the rating, we will have a flexible system. VDV also does not want to age by 33%. However, I think that we will come to the conclusion that we will age current regular tournaments by 10%, and super tournaments by 33%. Players with a high rating should be afraid of aging by 33%, as it affects the average and weak players less, and the leaders must prove that they have not lost their strength and can still compensate for the aging and get an increase. And where else to do this if not in super tournaments?
Added 34 minutes ago AmberSoler;63447
Added 14 minutes ago [/U]I have long ago expressed my conclusion that you do not understand the meaning of the oscillator... Its goal is to partially compensate the player for lagging behind the leader if the player took a high place, and vice versa.
Well, I guess I don't. Imagine, I don't understand. The leader worked hard to show a great result. And instead of slightly pushing away his rivals, or at least leaving them with what they have earned themselves, they are also given compensation (and think about it, for what - partially compensate the player for lagging behind the leader) - cool :) The player with 1000 days of completion gets half the rating of the leader for second place (50 versus 100), even if the leader completed it in one day. I wish I lived like that :)
Your approach is dogmatic. Mine is based on balance... We simply have different approaches. You stubbornly focus on the difference in results, while I want to encourage those who took second, third, or fourth place in the tournament... A silver medal in a championship doesn't represent a specific result; it signifies second place. And that's what matters. No one remembers the seconds, centimeters, or kilograms. Everyone knows who the second-place winner of the championship is. And that's important.
I suggest we end our conversation. It's pointless. And as I see it, you have no concept of compromise. Because not everything can be determined by mathematics. There's also psychology, ethics, and human relationships. And even in the most rigorous sport, all of this exists. You deny it. But people are not robots.
My approach doesn't rely solely on dry statistics. The goal is to create motivation. Because ranking is ultimately secondary. If it lacks variability, if it's not flexible, it's not interesting. Moreover, all the aspects that make it up are purely game-related. I don't bring anything in from the outside. I simply consider ALL the factors that should be taken into account. Namely: the actual outcome, the place taken, the number and strength of opponents. All of this equally shapes the player's assessment of their achievement. But with you, it's just a series of duels with the leader, and everything else doesn't matter.
These are simply completely different approaches. There's no point in comparing them. And let's leave it at that...
Could you please share your thoughts on comparing results not to the leader, but to the average result? The leader, having a result significantly higher than the average, has a chance to pull ahead. The underdog doesn't lose motivation because they will be "helped" by stronger players.
Your version is dogmatism. Mine is a bet on balance... It's simply that you and I have different approaches. You dogmatically insist on the difference in results, whereas I want to reward those who took second, third, and fourth place in the tournament... A silver medal in a championship isn't based on a specific result. It signifies second place. And that is the main thing. No one remembers the seconds, centimeters, or kilograms. Everyone knows that person is the championship's second-place winner. And that is what matters.
The rating remembers the kilograms. But you are right—the medal will remain with the second-place winner. No one is stopping you from rewarding them with certificates and medals. Shower them with grails using a linear formula: a bag for those at the top, nothing for those at the bottom. That is correct, and your linear formula, even without the first component, would handle this job perfectly. But what does this have to do with a game rating that reflects the players' skill in Heroes??? The non-linear system rating was devised precisely for this reason—to remember the kilograms and seconds. As for the medals, the rewards in the form of grails already serve that purpose. It turns out you want grails by a linear formula and the rating to follow the same path. In that case, the VDV rating is almost the same thing. Everything there is linear and great, depending on the place taken, and nothing needs to be "straightened out." Just account for the number of people and that's it. I have already stated my position. Tournament organizers are free to choose how and what they calculate. As long as VDV agrees to calculate at least its own tournaments using the experimental rating, I am at peace :) That is why I say that ALL rating tables should be published. Why—are you worried about space? Some people want to see the non-linear rating, some the oscillator. And some are fine with the VDV rating.AmberSoler;63462
I suggest we end our conversation. It is pointless. And as I see it, you have no concept of compromise.
This is not a case where one needs to compromise. Because I am certain there is a category of people who don't just want to admire their numbers, but actually want to compare their results to see how far or close they are from the leaders in terms of off-line Heroes skill. Regardless, some of your ideas have been accepted. This includes aging the rating in regular tournaments by 10% without any privileges. Relative permissible limits for leaders have been introduced to exclude cases where there are no strong players in a tournament. Yes, those in prize-winning positions will receive the same certificates and the same amounts of grails. However, it will be impossible to artificially inflate the game rating. So, thank you for your help in the discussions.AmberSoler;63462
After all, not everything is determined by mathematics. There is also psychology, ethics, and human relationships, for that matter. :) Even in the harshest sports, all of this exists. You deny this. But people are not robots.
I am not denying it; in fact, I constantly say that there are already rewards for prize-winning places, and they are in no way diminished or belittled based on the leaders' Walkthrough days. But you, for some reason, deny the additional experimental system, which accounts only for the result in days and the strength of the leader. You want to have both the "grail pot" and the rating, and for them to be the same balls, just from a different side :)AmberSoler;63462
My version does not rely on dry statistics alone. The goal is to create motivation. Because the rating is ultimately secondary. If it lacks variability, if it is not flexible—it is not interesting. Moreover, all its components are purely game-related. I don't take anything from the outside. I simply account for ALL the points that deserve attention. Namely: the actual result, the place taken, and the number and strength of opponents. All of these equally form the player's assessment of the achieved result.
That is wonderful, and all of that will be reflected in the grails rewards.AmberSoler;63462
And your version is just a set of duels with the leader, and nothing else matters.
Exactly—my system was designed specifically for that, as an alternative to the grail pots and the VDV rating.
Added after 25 minutes AlexeyPank;63470
Please, say something about comparing the result not with the leader, but with the average result? The leader, having a result significantly above average, has a chance to pull away. An underdog doesn't lose motivation, as they will be "pulled up" by the stronger players.
results of tournament 1 for Heroes-3: player 1 - 1 day player 2 - 100 days player 3 - 1000 days
1101/3=367 player 1 - 367*100/1=36700 player 2 - 367*100/100=367 player 3 - 367*100/1000=3.67
results of tournament 2 for Heroes-3: those three players skipped it, and we won't even age anything. assume the map Walkthrough difficulty is the same.
player 4 - 1 day player 5 - 10 days player 6 - 20 days 31/3=10.3
player 4 - 10.3*100/1=10300 player 5 - 10.3*100/10=1030 player 6 - 10.3*100/20=515
Now merge everyone into a general rating table based on the results of the two tournaments. You'll get nonsense. There's nothing left to talk about. If it doesn't work on a simple example, then some kind of bug will appear in reality too. Now calculate this same example using my non-linear system, taking into account the permissible time limits. You will get a normal general table that specifically shows who is capable of what. Try it for fun. Or you can even consider the oscillator.
That's not the right approach. Not only do you not have the right, but you are OBLIGED to use numbers and examples to show the shortcomings of the system I proposed if you plan to actively participate in tournaments for the game rating on this resource!
And what do we get as a result: four people (AmberSoler, AlexeyPank, Sage, and myself) expressed their opinion that the tournament rating should be calculated using an oscillator; they all provided arguments for their opinions, but you simply don't want to hear them or take them into account... because your system is supposedly better from the start! It's not good to act like that...
Further regarding aging and the current rating: I already spoke about this, and AmberSoler did too. But I will repeat: combining a player's "result" and his "activity" into **one number** is not the correct solution! Or rather, it's completely wrong. Why? I answered this question here. Therefore, we need to abandon the current rating in favor of, for example, an "accumulative annual rating" or some other one, but **it must definitely be limited in time**!
P.S. Sorry if I came across as a bit harsh.
Sincerely.
to IronAxe: You mentioned something here earlier. And what do we get in the end: four people (AmberSoler, AlexeyPank, Sage and I) expressed the opinion that the tournament rating should be calculated using an oscillator, they all argued their point of view, but you simply don't want to hear them and take it into account... because your system is supposedly better a priori! It's not right to act like this...
The thing is - I gave examples with numbers where the oscillator specifically distorts the player's strength rating, artificially increasing it for some players (leaders) and decreasing it for others (outsiders). With a non-linear system, everyone gets what they earned in terms of days and how close they got to the leader. Then I provided results of calculating two abstract tournaments, where this linear distortion shows strange results, when a player from the second tournament takes much longer in terms of days, but by taking second place, he surpasses a player from the first tournament who plays better, as can be seen from his results. Conclusion: I showed you a situation where the linear system works incorrectly, and the non-linear one works correctly. No one has shown me another situation where it would be the other way around, either in numbers or otherwise. In the discussion, one obvious flaw of the non-linear system was revealed, but it also exists in the linear system. I introduced acceptable limits and slightly smoothed out this situation, but this drawback remained in the linear system and even in the oscillator. I am not forcing you to calculate the ranking in Kanighem's Tournaments using my system – calculate it however you want, and you can use the non-linear part of the formula that I proposed as you wish. In VDV's 3rd Heroes Tournaments, I will keep my own rating. This is agreed with VDV and Prophet. I believe that there are no such flaws in this calculation as with the oscillator, and this rating more realistically reflects the playing strength of players in Heroes. Since I did not see any corresponding examples in the figures comparing the non-linear system with the oscillator in favor of the oscillator.Kyle;63893
Further about aging and current rating: I have already expressed myself on this issue, and AmberSoler too. But I repeat: combining a "player result" and his "activity" into one number is not the right solution! Or rather, it’s completely wrong. Why? I answered this question here. Therefore, we need to abandon the current rating in favor of, for example, an "accumulative annual rating" or something else, but it must be limited in time!
Aging reflects not only activity. If a player played worse compared to his previous game, and his tournament rating cannot cover the aging, then his rating may also decrease slightly - which is also correct. The rating has the property that it does not always increase, but can also decrease depending on the shown game and results. Missing tournaments should also not be encouraged. But I went halfway and agreed to age the rating by 10% instead of 33% in regular tournaments. In super tournaments, I think it should be aged by 1/3. You can take on keeping an accumulative rating or any other you like. I undertook to calculate the current rating because, in my opinion, it most realistically reflects the current playing strength of a player on this resource. I repeat once again that the cumulative system will work on Kukle and Grails; why count the rating there as well? But I'm not forbidding anyone who wants to do so. Personally, I don't need the same balls, just from different sides. I hope I have clearly expressed my position. Understand my idea, and I will understand yours too.
Added after 31 minutes Kyle;63893
to IronAxe: You mentioned something here earlier. And what do we get in the end: four people (AmberSoler, AlexeyPank, Sage and I) expressed the opinion that the tournament rating should be calculated using an oscillator, they all argued their point of view, but you simply don't want to hear them and take it into account... because your system is supposedly better a priori! It's not right to act like this...
Regarding the spontaneously formed vote here, I can only answer one thing. Even if all four of you say that the oscillator is better a priori, but cannot refute the flawed examples I provided, then this means nothing. Moreover, Sage has not appeared for a long time and did not express his final opinion, and AlexPank seems to doubt it - since he plans to calculate Kanighem's Tournament using both systems. Only AmberSoler and you clearly spoke out in favor of the oscillator, without refuting the incorrect operation of the oscillator in the examples I provided.
I have come to a final conclusion for myself, which sounds approximately like this: The specifics of playing Heroes at speed involve non-linear results, expressed in days passed, which more realistically shows the playing strength of each player. Places are always linear and introduce strong distortions. A player in second place, having taken 1000 days, will receive a rating of at least 50 :eek: No, second place shouldn't be worth so much for a player who plays like that!
to IronAxe: Regarding the first part of your post, I won't argue, as this discussion has been going on for 10 pages... and it's just funny to me:smile40:
Regarding the second part: You write:
I decided to calculate the current rating because, in my opinion, it most accurately reflects the current skill level of a player on this resource.
meaning any player, right? But I already wrote above that the current rating will show us the leader, but there will be complete confusion regarding the rest. because it will be unclear whether the player is not playing, or playing but poorly, or etc.. i.e., it will show us the skill level of only one player on this resource! What about the rest?
Regarding the third part of the post: sophistry...
to IronAxe: Regarding the first part of your post, I won't argue, as this discussion has been going on for 10 pages... and it's just funny to me:smile40:
---POST 312 BY Kyle--- Let's go into more detail here... explain what tournament, current, cyclical, and annual ratings are. This is needed so that we can speak the "same" language. ------------------------- Yeah, it's funny, considering you only started to understand what we're talking about from post 312 onwards. And you won't argue because you couldn't say anything then, and you can't say anything now. Kyle;63969
regarding the second part: You write: which implies any player, right? But I already wrote above that the current rating will show us the leader, but there will be complete confusion regarding the rest, because it will be unclear whether a player is not playing, or playing poorly, or something else. That is, it will show us the strength of the game of only one player on this resource! What about the others?
:) And there's nothing to understand here; the current rating reflects the current state of affairs; if they haven't played or played poorly, the rating will drop. What's so unclear about that? And you can see how good someone is on their profile.Kyle;63969
regarding the third part of the post: sophistry...
I've long understood that it's pointless. Therefore, I'm stopping the argument here and calculating the VDV tournament rating according to the non-linear scheme approved by the organizer. You personally don't have to look at it - no one is forcing you to - just accumulate grails and the cumulative rating. In general, I suggest you calculate the grails for a tournament and then enter the sum into the rating table - THAT'S YOUR RATING :)
to IronAxe: Regarding the first part of your post: You are exaggerating... and this only confirms what I wrote in the first part of this post.
Further:
There is nothing to understand here; the current rating reflects the current state of affairs. If you haven't played or played poorly, your rating will drop. What's not clear about that? And look at the avatar for coolness.
You didn't answer the question. What is the purpose of the current rating? To display the strength of play of all players or just one? Answer this question clearly.
Kyle
