Let the highest Rank be the group level

then they would just have to change the system to make it non exploitable

I’m asking for the removal of SR and just show MMR

how could one exploit it? let’s say they took out PBSR for the sake of this discussion.

so I asked you these questions as I read your post, and now I’ll sum it up with my theory.

to make MMR public we would have to loose:
A.) getting a rating extra quick. this what PBSR does and would have to scraped

and that’s basically it, you really wouldn’t have an edge in gaming the system, because a bronze can get carried by 5 GMs whether they know the MMR formula or not.

and finally, this guy wants the MMR formula transparent.


(timestamped)

the fact that he never says anything about anti-exploitable trigger safes for a transparent ranking system, makes me think that it’s possible to publish the MMR formula and make it non exploitable.

it can be done, I have faith :slight_smile:

Possible exploitation of the system and other negative effects on the game aside, the MMR system is Blizzard’s IP and they won’t want to give that up easily. And why should they, from a legal perspective?

tell me one reasonable way to exploit it, it’s just a intransparent system and it’s not really doing what it should do. Look at true skill, it’S public and it works.

forgot to link the vid

the guy on there says they should make it public because those who love that kind of stuff will eat it up, and those that don’t really won’t care either way, so make it transparent. That’s what he says.

Some things about TrueSkill are public but the specifics of what statistics are used in each game are not. It is that sort of detail which would be exploitable. I.e. padding certain stats and not trying to achieve the match objective is exploiting the system.

Jeff Kaplan:

SR is not used in matchmaking (01:13:30).

i still don’t think it’s true, wouldn’T be the first time they lied. The Team SR is always too close for it to be really true.

I gave four sources. Are they all lying? If so, why/ What’s their motive?

the same for hiding it in the first place. And i showed you some proof too: The Team SR is always in a range of 200 max on both sides, and no this isn’t a proof of anything, as this is the case even in decayed matches.

It’s true, people like me would find it fascinating. That doesn’t change the fact that it’s their intellectual property which is very valuable to them. When you have a method for doing something which is, or has the potential to be, very profitable it’s not wise to just give it away for free.

contrary to what the MMR master in the video says

I don’t think it’s so awesome that they can’t publish it. Because 99% of matchmaking is common knowledge. Basically what I’m saying is that there are no awesome secrets worth them keeping it a secret. Sure there might be one or two things they want to keep secret, but I don’t think that’s the main cause.

I think the real reason the keep MMR secret is so they can tinker with it, without having to publish “updates” to the MMR system. And I don’t mean tinker like forced 50/50 or anything like conspiracy. What I mean is that they can tweak it to make it better, and know one is the wiser.

I just don’t think the competition would gain much ground on Blizzard if they gave out the MMR formula. Now maybe the matchmaker belongs to Activision and they don’t want to give it away.

Bottom Line: The participants of a sport must be privy to the rules. Anything less is sacrilege. That’s my opinion, IP or not.

I’m not going to repeat the reasons for not showing MMR and not publishing the method of calculating MMR. But let’s assume that it is of similar complexity to Microsoft’s TrueSkill2. Here’s a link to a paper (unpublished) which describes TrueSkill2.

h ttps://www.microsoft.com/en-us/research/uploads/prod/2018/03/trueskill2.pdf

I found it very interesting but to be honest that’s all it is. It doesn’t do anything that proponents of making the MMR calculation method public claim. It basically boils down to ‘play better and your TrueSkill rating goes up’. That’s hardly mind blowing.

Probably true. The paper from my previous post on TrueSkill2 shows that you can say a lot without bankrupting your company. However that paper also shows that even after a great deal of detail is made public the players still whinge about inaccurate rankings. People complain that the rankings in Halo 5 are trash just like people complain about Overwatch. This type of person will simply always complain.

I think you’re right. That in itself is a pretty good reason.

The rules of the game are not a secret. E.g. we all know that control (koth) maps are best of three rounds and that each round is first to 100% control. The rules of the game have nothing to do with MMR or matchmaking. I don’t see how knowing the MMR formula could help anyone win more matches. So in that respect it’s perfectly fine to have it remain a secret.

1 Like

Since SR is not MMR, and PBSR is only for SR, publishing the MMR gain / loss formula wouldn’t affect PBSR, it would show the min/max players how to gain the most MMR in a match. More MMR means you go higher and play against stronger opponents. If SR was gone, you’d just have MMR, which is a big hairy decimal in the range -3.0000 to +3.0000. Someone’s MMR might be 1.3349727293 currently.

Jeff says that MMR is an ugly number, and it moves in ways that players probably don’t like, so they have SR, which is a clean number (Integer), and which has restrictions on how it can move.

So, if you got rid of SR, you’d have MMR, which moves according to your performance, possibly in dramatic ways. MMR is probably more performance based than PBSR is, and therefore even more able to be gamed / exploited.

The matchmaker is not scary, most of the things people get all pissy about are myths. The matchmaker just tries to make a game with 12 people that have the exact same MMR. It usually can’t do that quickly, so it tries to balance the teams out with as close to a 50% win likelihood as possible (calculated on MMR).

There are no forced losses or wins or anything else. The matchmaker just uses MMR to create a match. MMR moves in ways that we don’t know about because if the formula were public, people would exploit it.

now wait, does Halo 5 officially endorse that paper? Do they publish the formula with excruciating detail and all? and the people still complain, is that what you’re saying.

now wait, with a public MMR they’d still be able to tinker with it, they would just have the added work of publishing their “tinkering” maybe the published MMR formula would be really mess and a hodgepodge so it just wouldn’t suit the Blizzard “polish” but nevertheless they should publish it because it’s the right thing to do imho

This is my stance. Maybe if I found out what the MMR formula was and why they keep it a secret, I might change my mind. But anything less than that is unlikely to change my opinion on the matter.

How is similar SR contradictory to not using SR for matchmaking? SR follows MMR closely and unless the account is brand new and the MMR has moved faster than SR can, MMR and SR are closely aligned.

Do you have any reason whatsoever to believe they’re lying about the matchmaker only using MMR for finding the right 12 players for a game?

Name one sport that has a Matchmaker at all.

1 Like

This is 100% incorrect. That’s like saying that the only thing that counts in basketball is how to win a game. And the season and the way the playoffs work should be ignored by the teams.

The way the season work may have little persuasion towards the basketball teams, but nevertheless, they are still made aware of how it works.

well, the last place team of the NFL gets first round draft picks. So if you’re not going to make the playoffs, it’s better to loose as many games as possible.

if this information was private, then there could be room for exploitation. but since the rule is public, everyone has a fair shot at exploitive behavior so that makes it fair.

Yes and no. The equations are all there but in terms of what stats are used they only give examples such as number of kills and number of deaths but that’s about it. Being vague in this respect is probably for a few reasons:

  1. The stats which are used depend on the game (TrueSkill2 is used for many games, not just Halo 5 and Gears of War 4). So in a paper about TrueSkill2 there’s not real reason to go in to all the details of one specific game.

  2. They may not want players to be able to manipulate their ranking via stat padding.

  3. It’s already a long paper and that sort of detail is superfluous. You can understand TrueSkill2 without it.

  4. It’s pretty clear that they try to model certain probability distributions for certain stats which allow them to predict match outcomes. Because they want to predict match outcomes the players really should just focus on winning to get the most out of the system.

Point 4 applies to Overwatch as well. People should focus more on learning what it takes to win a match than learning how MMR is actually calculated.

1 Like