Hmm, well I mean he'd be in a weaker division, that's true. But what's his motivation? That would just make it harder to be eligible for the top 200 right?

The only benefit is he would be guaranteed a #1 spot, which he would probably have in a more competitive division anyway. I guess he would have a huge lead over #2.

He would be playing the exact same players either way. I don't see the benefit tbh.

Things you need to keep in mind:

#1: Except for a very short period of time (while they're filling up), divisions have exactly 100 players in them.

#2: Players *never* get moved from one division to another within the same league. Once you've been put in a division, you stay there. When people get demoted out of your division, somebody else will be promoted into it. Otherwise divisions are static.

#3: You don't play people in your division. I suppose you could get matched against somebody in your division, but the game isn't primarily looking for in-division matches.

Points really have nothing whatsoever to do with divisions. You get points when you win. People who play more and are better will get more points.

You really are reading way too much into the division thing. They simply don't mean anything. It's a way to say "hey, I'm #1 in my division" instead of saying "ahshit, I'm rank #23728 on the ladder". That's all.

Of course he would.

People get demoted, you know, leaving an empty spot that has to be filled?

I thought this was the case for a long time too, but it seems that it is not. Go read that TL thread.

I never said points have anything to do with divisions. I said that players of similar skill will be placed in the same division. Points != skill nor was I trying to make that claim.

Matchmaking has nothing to do with divisions at all. I never said it did. It gives no preference to players in or out of your division. So we agree there I assume.

'He" was referring to #1 in my division. Just so that's clear.

If you really think that all divisions are equal that's fine. You can keep thinking that. I've given you the literature to suggest that's not true. I gain nothing by convincing you, so whatever.

After taking a glance at that thread, some of the conclusions drawn are absolutely laughable.

For example:

If blizzard's intent is that players who are active and skilled will reach the top of their division, then they should be separating top players out into *different* divisions, not trying to put players of similar skill in the same division.

If you have the top 100 players in a single division, the 100th player on that list is going to be "ranked" lower than some joe schmoe who just got promoted to diamond with a 50/49 win/loss record.

Seriously, division ranks are completely meaningless, based purely on the luck of the draw of the set of people who were promoted at the same day you were.

That entire thread is an attempt to answer the question "why doesn't blizzard's top 200 match sc2ranks top 200", and manages to come up with an extremely complicated answer that doesn't even make sense, when there are much simpler answers available.

Except they reverse engineered it perfectly through random chance right?

I'm not arguing about this anymore, it's stupid.

This can only possibly be true if people move laterally between divisions, rather than divisions filling up as they're created.

I happen to know from personal experience that divisions are created one at a time, and filled at that time. Have you never paid attention to the fact that when you get promoted, somehow magically all the people in your division were added at the same time? For the first week, the up/down arrow will be an ! instead, and hovering over it will show when the person was added to the division.

How can divisions possibly be intended to contain players of approximately equal skill level if they contain a set of exactly people who were promoted within the span of a day or two?

Not all divisions are equal. When have I ever suggested that they are? I have never said that, and the idea that they are is absolutely ridiculous. It's obvious that some divisions have stronger players in them than others.

What I *have* said is that divisions are completely meaningless, and they are. That doesn't make them equal.

The whole division concept is an attempt on Blizzard's part to give people reasonable targets to work towards that make logical sense, so that they'll hopefully play more. "Reach #1 in my division" is a much more sensible goal than "reach spot #17563 on the ladder".

You'd be surprised at how good the brain can be at finding patterns where none exist.

Seriously, what's complicated about this? A division gets created, the next 100 people who get promoted get added to it. The divisions that are full of top players have those players in them because those were the players who reached diamond the fastest.

That's how it works, as you would know if you had ever paid attention to the little exclamation marks when you got promoted.

I agree that the guy's conclusions regarding why blizzard designed the system this way make no sense, but the numbers all add up the way he has described them. This is what I meant when I said that you need to stop assuming the ranking system has any semblance of logic in its design at all.

You know what? I'm going to take a step back here, and say that it *is* possible that specifically for the top 200 posting, blizzard somehow tries to take division into account.

Since you don't get matched against people in your division, it would be absolutely *ridiculous* of them to do this, and I can't think of any logical reason that they would, and even if they are applying some sort of modifier based on division, that modifier would have to be a variable modifier, rather than a fixed one.

But even if they are doing something like that, it doesn't change the fact that divisions are in fact meaningless. It would just mean that they're including a factor in their top 200 ranking list that is completely meaningless, that they absolutely should *not* be including in any way shape or form, and that wouldn't change the fact that any claims that divisions are created to try to group people of similar skill together are clearly hogwash.

Perhaps there has been some miscommunication here. Allow me to clearly and concisely say my point.

Divisions are not filled in a "the next 100 that are ready to be promoted way" but in a "the next 100 that are ready and are sufficiently close to X skill level way".

There are ~134,000 players in diamond.

There are ~50 divisions represented in the top 200.

Think about it.

Determined how? At the point of promotion, especially early on, the number of games to base "X skill level" on would be very small.

Furthermore, there's no reason for blizzard to do this. It would be extra, unnecessary work that couldn't possibly be even remotely accurate anyway.

So why do it?

If divisions were created and filled one at a time, the first 50 diamond divisions created would contain the first 500 players to reach diamond.

Is it *really* that much of a stretch to say that those 500 players would be very likely to be the best and most active players from beta?

How can you say it wouldn't be accurate when you have no comprehension about how it would be done?



Yes, absolutely. The first 500 players (or even 5000...) to reach diamond representing the top 200 would be a ridiculous coincidence.

I am not too well versed in matchmaking/ranking. I really wish I was cause it's extremely interesting. I would have thought somehow in my math education that would have come up, but unfortunately it didn't. I even took an algorithms course that had some matchmaking implications but it was more focused on rankings without weights.

I think that a lot can be determined from a small-ish sample though. 30-50 games is probably enough to not necessarily pin point skill but group some players together, perhaps in some kind of divisional way.

Division rankings and points don't factor into Blizz Top 200.

Correct. But I think eligibility does? I'm not really understanding that entirely, I just took that from the thread I posted above.

I know you know more about this stuff than I do.

And I was under the impression that it was ranked by points?

I can say it wouldn't be accurate because it would have to be based on a small number of games.

No rating system that is based on a small number of games, in a game like starcraft, can *possibly* be accurate, not matter what the method is.

Reread what I said please.

It absolutely would *not* be a ridiculous coincidence for a significant number of the first 50 *divisions* created to represent a sizable percentage of the top 200.

Furthermore, it wouldn't even necessarily be the first 50 divisions created. In the first weeks, they were being created very furiously I'm sure. But the fact of the matter is that the top players from beta represent a large percentage of the top 200 right now, and those players would have been among the first to qualify for diamond, so the fact that so many of them appear in the same divisions can be perfectly explained by the first-come-first-served theory, which is *far* simpler and makes far more sense than any sort of convoluted "create divisions based on approximate skill" theory.

This is one place where Occam's razor doesn't apply imo.

WAT? I remember when they released the first one they stated that it is based on MMR. I assumed that activity was also a factor.

Activity can be considered indirectly by points. Bonus pool.

Something you should know about programmers, at least the ones that aren't horrible at their jobs.

They really like to do the minimum amount of work possible.

The simpler system is orders of magnitude more likely than the more complex one. Especially when there wouldn't be any good reasons to implement the more complex one.

Except that the one I'm suggesting is not only not that complicated but seems to fit the patterns perfectly.

I actually posted the reason as to why they would implement that system above.

I'm starting to think you're trolling me at this point. If so, well done.

Beat: Wants to gouge out eyes.

----

Let's play devil's advocate. Suppose I'm wrong. What is your evidence for your proposed system? Where has that ever been discussed? Link to blizz post or tl discussion?

And for the last time, this is lol false. Learn maths.

Ugh, my 2nd bronze match, PvZ, was awful. It lasted 100 minutes.

I expanded early and often and built up a force of gate units and colossi, which when I finally pushed 20-25 min in was enough to take out his main and his nat expansion. He took an expansion on the top left corner of the map, which is unreachable to ground units, and just built lots of brood lords, corruptors and hydras. Much of my original army that I rushed his base with was still alive, so I was supply blocked from building many air units to try to attack his base. I attempted to kill him with the few void rays I could build but his defenses were very strong and took them out quickly. I had expanded across the map and depleted most of the mineral fields. I have a ton of minerals in the bank and a supply maxed army.

He ends up attacking my main with brood lords and corruptors. I attempt to hold him off with stalkers but my army is decimated. With no way to stop him from destroying my main (which had like 8 gates, 6 stargates) so I start building lots of gates and tech structures by my two newest expos and start researching blink right away.

The game ended when the last of my stalkers died attempting to prevent the enemy from expanding. I was out of minerals and my expansions had all dried up. My opponent said that he never should have won. lol

Mistakes:
- Didn't expand across the map nearly quickly enough. Shouldn't have left him any opportunity to expand and acquire new minerals.
- When my main was taken out and my army was destroyed, I should have built lots of stargates, not warpgates, and massed/teched void rays, not stalkers.
- Didn't spread out pylons across the map enough which made warping in to defend expansions clumsy.
- Should have invested in more photon cannons to defend expansions.
- Need to get in the habit of at least sending a unit to sit on expansion sites so I at least know if my opponent is expanding.
- Didn't use warp prisms to attack his expansion. I thought they were a stargate unit and was confused that they weren't there and now realize they are a robo unit. /facepalm

The final resource collection count was 80k-50k in my favor.

Cliffs: It doesn't matter how many minerals you mine if you put them to poor use.

It happens, those types of games are soul crushing. I tend to get most angry at long games. I know some people get upset at cheese, but I find that's not really the case for me. Even when I win long games it can be tilting. Just put it behind you and play another. If you ever want to post a replay or anything, I know there are a lot of people that would be happy to help you out. I know it helped me a lot when I first started.