Connection ErrorBad RequestSession 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.
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.
If the losing player had a higher rating, they would lose a larger number of units of X (which would be added to the player with the lowest rating, but the winning player). This number of units was determined by how long the game lasted; the shorter the game, the larger the value of X. Also, if the losing player had a lower rating, they would lose a smaller number of units of Y (which would be added to the player with the highest rating, but the winning player). This number of units was determined by how long the game lasted; the shorter the game, the larger the value of Y. However, the value of X was always greater than the value of Y, meaning it is more advantageous for a player with the lowest rating to win (they would receive a larger bonus), and it would more severely reduce the rating of the player with the higher rating if they lost.
That's an interesting idea, but if it were to be implemented in an offline mode, our mathematicians would be upset because all the discussions and even arguments they had would be for nothing, and the calculations would have to be redone (new formulas, new discussions...). Do you think anyone would take on that task?
"Единственное, что можно сказать с уверенностью об удаче - она изменит..."
KyleStatistics
Registered:
Posts:2,071
Auto-translated
Friends, Please note that quoting a large message in its entirety is considered bad etiquette... not everyone has unlimited data... It's easier to simply address the person directly. For example: to K0Lb@zzeR
Thank you for your attention.
P.S. to Kamikaze and Ley, please correct your messages.
Кто состоянием своим доволен, тому жить весело. (с)Алекса́ндр Васи́льевич Суво́ров
LeyStatistics
Registered:
Posts:1,299
Auto-translated
I also think it makes sense to consider the location – because the result directly depends on it. The only question is, how do we correctly determine this relationship?
to Kyle: Is this okay? :)
Kyle: Yes. Thank you.
"Единственное, что можно сказать с уверенностью об удаче - она изменит..."
An interesting option, but if it's implemented for OFFLINE mode, I would be upset on behalf of our mathematicians – because all that they discussed and even argued about for so long will become irrelevant, and the calculations will have to be done again (new formulas, discussions...). Do you think anyone will take this on?
An old friend is better than two new ones!!! I think the author of that system isn't a stupid person :) And why did Iron keep silent... he was supposedly involved in discussing that system!!!:smile10:
Iron, be so kind! come up with a more realistic example... for example, at least 8 participants!!! 3 participants is nonsense IMHO!!!
:) Good people. I already said the example was exaggerated, to make it easier for anyone to calculate using all systems and see what the flaw would be and understand it. If you don't want to understand it with such an example, who on earth is going to deal with an example where I whip up a couple of tables with a hundred people in each? :) I already provided a real example too (the 4th stage of Kanigem).
Added after 13 minutes Fireball;105847
This is impossible, since the participants' ratings are unknown. And it makes no sense, since the example is not real.
The example examines how a newly devised system will behave when, conventionally speaking, no one has ratings yet :)Fireball;105847
2Kyle I see that you are also no stranger to the competitive spirit. I understand you, and that's why I cannot conceive of a rating without a place. Results are important in training, but in Competitive events, almost everything is decided by your place in the final protocol.
All your patriotic examples about Competitive events and the importance of placing are applicable when talking about competitions where players go through a bunch of qualifying stages to get in. In other words, there are practically no weak players in such competitions; they've been filtered out. In our case, everyone is lumped together, the tournament lineups are inconsistent and random in nature. Maximix was in the "double" — he dragged everyone's rating down a bit and then he's gone. I suffered too, even though I was the one who invited him back then :) And for the "four," no one came to push others down by place. All the factors I mentioned are evident: the randomness of the lineup from tournament to tournament and version to version, the inconsistency of strength in the lineup, i.e., the rating flaw heavily depends on the random factor due to the place occupied. Moreover, I see no competitive element here. This is a clear substitution of the real result with the occupied place, and consequently, a distortion in the rating.
Added after 11 minutes Kyle;105849
Exactly, and at that point it's no longer important by how much you lagged behind the leader! Only the place matters. Because many ratings for competitions (with a large number of stages) are calculated by points which directly depend on the occupied place, and do not depend on the result as such.
Alright, stop exaggerating. It's not ratings being calculated at the stages, but POINTS, to tally up the money for the entire tournament based on points. And ratings are calculated later :) And one rating is not like another. Some ratings can, for example, take into account even broken knees of your runners :) and the results shown by a player at tournaments. The place may be of a supporting nature. If the rating simply depended on the place — then what would bookmakers live on? The unpredictability of results in matches ultimately affects the places. Therefore, the rating should to a greater extent take into account specifically the results obtained by the player, and not the occupied place. And if we take your example with runners, one finished in 10 minutes, and the whole crowd caught up a couple of hours later, then it's unlikely that the second place would get a good rating from bookmakers, despite their silver medal :) PS. Basically, I'm tired of this again, look, Amber is smarter, she doesn't interfere now. You are developing some thing to calculate a rating with which you can then determine the finalists for the Kanigem tournament. But that is not a rating, it's a system for counting some points to determine winners in a multi-stage tournament. My idea was different from the start. It's already clear who took which places when we arrange the players according to their result. And now, taking the first place as 100 rating points, we calculate the rating for all players based on the result in DAYS. I don't get why the place is needed there? :) My rating was not, roughly speaking, tasked with identifying winners of multi-stage tournaments :)
Added after 8 minutes Kamikaze;105852
You mean it should be simplified to the impossible, and important components should be ignored. :rolleyes: Moreover, I showed you that my rating turns out the same :D PS. There won't be three participants in official tournaments
That means you've picked the coefficients well, which remove the flaws for the occupied place :)
Added after 7 minutes Alex_R;105855
to IronAxe May I ask the reason for the suspension of the rating calculation according to your system since April... and this also concerns the nomination of the 13 best warriors!!!:confused:
There are several reasons: 1. I'm waiting for the most optimal formula to be developed 2. Lack of time due to a tight schedule of tournaments and discussions of the most optimal formula. 3. It's not hard to recalculate all tournaments using my formula, but updating the page on this forum is very difficult for me. 4. Checking the need for an alternative rating shows that besides you and me, nobody needs it, so is it worth continuing?
Added after 2 minutes Alex_R;105893
Why did Iron keep quiet... he seemed to participate in the discussion of that system!!!:smile10:
I didn't keep anything quiet, I expressed in that same thread that it's unlikely to be suitable for the official side. At that time, I suggested the NON-LINEAR part of the formula for the rating, taking into account only the result in DAYS, to which they've now added god-knows-what :)
FireballStatistics
Registered:
Posts:863
Auto-translated
Good people. I said the example was exaggerated to make it easier for anyone to calculate it using all the systems and see what the flaw is and understand it. If you don't want to understand it with such an example, then who will deal with an example where I add a couple of tables with a hundred people in each?
Okay, let's analyze your example.
The example considers how the newly invented system will behave when, let's say, no one has any ratings yet.
Therefore, I assume that all players are beginners and assign them an average rating of 25. Fireball system: Player A - 5 B - 1.25 C - 7.5 D - 4.79 E - 1.35 Indeed, it's not very clear why E received a whole tenth of a rating more. But all this is minor compared to your system, where players receive 100 rating points for some unknown walkthrough and victory in a tournament with 2-3 participants. Even with this flawed example, my system works, but yours...
I already gave a real example (the 4th stage of Kanighem).
I also gave one (the 6th stage of VK).
Maximix was on the second tier - he lowered everyone's rating a little, and he's gone. I also suffered, although it was I who invited him at that time.
But this is a good thing you did. I would be very happy to see Maximix in the tournaments of our portal in the future.
And no one came to the fourth tier to compete for the spot.
Therefore, I assume all players are beginners and assign them an average rating of 25. Fireball system: Player A - 5 B - 1.25 C - 7.5 D - 4.79 E - 1.35 Indeed, it's not very clear why E received a whole tenth of a rating more. But all of this is insignificant compared to your system, where players receive 100 rating points for some unknown Walkthrough and victory in a tournament with 2-3 participants. Even with this flawed example, my system works, but yours...
Yours doesn't even account for the differences among leaders in equally matched tournaments of 2.5. You yourself admitted that you don't really understand why players receive different ratings there.
Mine: A - 5 days - 100 B - 20 days - 25 C - 5 days - 100 D - 6 days - 83.33 E - 20 days - 25 With mine, it's clear who received what and how much. 100 points are awarded for a pretty decent result in 5 days. All the others are based on this result. But the ratings of players from different matches for the same results are not different, which is important.
FireballStatistics
Registered:
Posts:863
Auto-translated
You haven't even mentioned the differences between leaders in equally-matched tournaments on 2.5.
The tournaments aren't equally matched. Player C beat twice as many players, so his rating is higher. But the ratings themselves are so insignificant that you can safely say that the players wasted their time.
A hundred points is a pretty decent result in 5 days.
And what makes you think that's a decent result? Maybe an average player can complete the map in a day. And in the tournament, only super-skilled players participated.
Another example for the sake of discussion :) Let's assume, for simplicity, we calculate the rating based on the "trio" [map]. 100 games played. We have the final rating table according to my system and according to the system that takes into account the places taken in the tournament. It's clear that players have played different numbers of games. But comparing the ratings of players who have played the same number of games according to my table, it can be unequivocally stated that the one who has a higher rating showed results in days closer to the leader in tournaments, or was even the leader themselves sometimes. And what can be said looking at the table where ratings soar taking into account the places taken in the tournament?? Unequivocally, nothing can be said, because a player might have jumped up with a hefty bonus in the tournaments due to the absence of the middle tier (like the example of the 4th stage), or they actually showed results in days close to the leader, or who knows what—in short, an "oscillator" in one word, a bonus with a coefficient, and nothing is clear :) Apparently, this is how they want to cloud people's minds :) However, I can thank my opponents for the discussion for teaching me how to argue scientifically. Now, when we're drinking in the garage and start arguing strongly about something, instead of the expressions known to everyone, I use words such as bonus, coefficient, oscillator, etc. Fortunately, it doesn't come to a fight, otherwise they'd probably kill me, because common expressions are more understandable to the people :)
Added after 12 minutes Fireball;105955
Tournaments are not equivalent. Player C beat twice as many players, therefore his rating is higher.
Tournaments are equivalent by condition, because it was agreed that the game is played from the same starting saves. Player D simply, out of stupidity, by chance, on a whim, signed up for the second tournament instead of the first—do I have to repeat it 100 times? What does the absence of fool D change for the leader of tournament N1? Nothing. Even if this player were there, nothing would change. But you stubbornly use the expression that he would have surpassed him :) Yes, he surpassed him anyway if you can't understand this from the example, even though the latter played in a different tournament :) This is visible from the results by days :) Do you actually not see this?Fireball;105955
But the ratings themselves are so insignificant that one can safely say that the players wasted their time.
Well, there it is. Now you've already admitted another flaw in your system. Like, players wasted their time for an insignificant rating according to your system. In mine, such nonsense is not foreseen. Even the outsiders got 25 each :)Fireball;105955
And why do you assume that this is a decent result. Maybe the map is cleared by an average player in a day. And super-noobs played in the tournament.
And why did you give everyone 25 at the start? That's exactly why the leader gets 100, and for 1st place. But nothing is given for free in mine :)Fireball;105955
Yeah, and also these ratings are full-weight and with a huge probability they are freebies. My flaw is in the tenths, yours is in the tens.
First he says he doesn't understand why he has different ratings for the same results, then that they are insignificant and the players wasted their time, then he pulled 25 for everyone out of thin air. And my system is flawed? :) I'm tired of laughing :( And he didn't notice a single one of these flaws in his own.
Added after 21 minutes And also regarding the "full-weight" ratings taking into account the strength of the tournament that you're trying to drag in. 4th stage. Suppose I showed the limit result on this map. But there were few players, and there were no strong players for this version of the game in offline. Accordingly, as the leader, taking into account all your bonuses and coefficients, I should receive less than if there were strong players and many players. Why is that? The result was the limit, and I couldn't care less; I am not responsible for the turnout of other players at the tournament. And then I go to the "trio" where there are plenty of strong players and many people, I try my hardest to stay in the top ten, but meanwhile, bonuses and coefficients will soar for the leading players for the strength of the tournament, and since I'll be displaced in rank, the oscillator will finish me off with kicks, EVEN if my result in days is not very bad. There was simply an influx of strong players for the trio. Offline differs from online in that no one personally interferes with the production of my personal result in days. Players can only interfere with me by their presence or absence if we start taking into account the places occupied. And the presence/absence of players is a random factor. And you want a random factor to affect the rating of each player??
PS. I already occupy a place in the trio in this tournament, which is distributed according to the number of days and secondary parameters. To take into account who wins in a multi-stage tournament, the VDV system would be quite suitable. But what does the rating have to do with this? What does my result in days, approximated to the leader as much as I could, have to do with this? Why the hell do I need a random factor of player presence/absence to affect my rating, and consequently a double counting of the place occupied, when there is a standalone result in days??
FireballStatistics
Registered:
Posts:863
Auto-translated
But when comparing the ratings of players who have played the same number of games according to my table, it can be unequivocally stated that the player with the higher rating showed results in the Tournaments that were closer to the leader, and sometimes even was the leader himself.
And you are wrong again. Let's say we have players A and B, and they both participated in 9 Tournaments. Player A consistently showed a result of 31 days in the first eight, and player B consistently showed 32 days. The winner always had 30 days, in short, maps like BB.
And in the ninth Tournament, B showed the same result as the winner (3 days), and took second place. The organizers decided to change the version of Heroes for variety, for example, to version four. Player A took third place with 4 days.
Who was closer to the winner more often? That's right, player A (8 out of 9 times).
And who has a higher rating? That's right, player B.
A - 849.19
B - 850.00
Well, now you have already admitted another flaw in your system. Like, players wasted their time for a negligible rating in your system.
This is not a flaw, but an advantage. If you play in a nonsensical mess instead of a Tournament, you won't see a rating.
And why did you give everyone 25 at the beginning?
This is the average level of a beginner (on a normal map).
First, he says that he doesn't understand why he has different ratings for the same results, then that they are insignificant and players wasted their time.
You know, I'm willing to forgive any system for a 0.1 point rating error.
Accordingly, as a leader, taking into account all your bonuses and coefficients, I should get less than if there were strong players and many players. But why is that? The result was maximum, and I don't care, I am not responsible for the attendance of other players in the Tournament. And then I go to the third Tournament, where there are many strong players and a lot of people, I try my best to stay in the top ten, but at the same time, the bonuses and coefficients will increase for the leading players due to the strength of the Tournament, and since I will be moved down in the ranking, the oscillator will kick me even more, EVEN if my result in days is not very bad. There was just an influx of strong players in the third Tournament.
Conclusion: beat strong players, and you will get a rating. Moreover, if you calculate separately for each version, then everything is fine. For the Hive, if you haven't forgotten, there are no coefficients.
KamikazeStatistics
Registered:
Posts:1,700
Auto-translated
To Fireball & IronAxe:
Guys, stop arguing – it's clear that both the location and the result are important aspects and cannot be ignored. Your argument itself shows how important both indicators are.
FireballStatistics
Registered:
Posts:863
Auto-translated
It is precisely from your argument that it becomes clear how important both metrics are.
The problem is that everyone sees something different. For example, you believe that the ratio of locations to days should be 3 to 7, while Alexey believes it should be approximately 1 to 2 (taking into account the bonus). Kyle and I believe it should be 1 to 1, and Topor considers locations to be evil and wants to ignore them altogether. If the overwhelming majority (or the administration) clearly expressed support for a specific ratio, then the issue could be resolved.
The problem is that everyone sees something different. For example, you think the ratio of places/days should be 3 to 7, while Alexey thinks it should be approximately 1 to 2 (taking into account the bonus). Kyle and I believe it should be 1 to 1, and Topor considers it an evil and wants to ignore it altogether. If the overwhelming majority (or the administration) clearly expressed support for one ratio, then the issue could be closed.
Exactly, we need to decide on a ratio and stick to it. I'm not insisting on 3 to 7 – it's just my point of view. We can use 1 to 1, or 2 to 3. Let's end this dragging on.
Added 3 minutes later I recalculated using the 1 to 1 formula for the first Heroes game and found that Prorock, who won the first tournament, is ahead of Alexey Pank, who participated in all three, by one and a half points. According to the old formula, Alexey would have won 22 points (6, 12, 6 places).
I recalculated using the 1 to 1 formula from the first Heroes game and found that Prorock, who won the first tournament, is 1.5 points ahead of Alexei Pank, who participated in all three. And according to the old formula, Lesha would have won 22 points (6th, 12th, 6th place).
And? Don't you like the first option? In my opinion, everything is correct. :smile01: Two sixths and one twelfth place are approximately equal to one first place. IMHO.
Кто состоянием своим доволен, тому жить весело. (с)Алекса́ндр Васи́льевич Суво́ров
Ley
