Outwitters WORLD CUP Winner - Alvendor (SWEDEN)

[Necrocat219]

What do you use points to determine? Do you do something like points divided by wins to guess our skill level? Just curious that's the only thing I can think of

[awpertunity]

(09-01-2012 05:30 AM)Necrocat219 Wrote:  What do you use points to determine? Do you do something like points divided by wins to guess our skill level? Just curious that's the only thing I can think of

I'm using mainly points to determine seed. But if the points are relatively close, I put the one with lower wins on top as you suggest. Here's the current list of seeds if you're curious (player number can be found on participants list):

11 - rank 1
15 - rank 6
4 - rank 8
7 - rank 25
17 - rank 27
2 - rank 54
20 - rank 59
19 - rank 74
21 - rank 98
1 - Master - 1110 points - 155 wins
10 - Master - 805 points - 76 wins
5 - Master - 585 points - 60 wins
13 - Master - 182 points - 15 wins
16 - Gifted - 739 points - 82 wins
9 - Gifted - 556 points - 57 wins
3 - Gifted - 570 points - 67 wins
6 - Gifted - 436 points - 63 wins
8 - Gifted - 369 points - 47 wins
23 - Gifted - 431 points - 86 wins
18 - Gifted - 381 points - 42 wins
14 - Clever - 948 points - 62 win
24 - Clever - 421 points - 25 wins
12 - Clever - 466 points - 40 wins

[GreatGonzales]

Interesting that there are no ST who are not in the top 100 so far. Actually - this gets back to my league percentiles thread - but if indeed there are only 200 ST players, maybe that's not surprising.

[CombatEX]

I understand

9 - Gifted - 556 points - 57 wins
3 - Gifted - 570 points - 67 wins

being switched but I don't understand this

8 - Gifted - 369 points - 47 wins
23 - Gifted - 431 points - 86 wins
18 - Gifted - 381 points - 42 wins

Surely 18 should be above both 8 and 23.

Anyway I calculated the Points/Win: img37.imageshack.us/img37/1663/owcadjustedseeding.png (for some reason when I make this a link it keeps changing to something else when I refresh the page and try to click it again so I just leave it in this ugly form =P) NOTE: I meant 6 in the adjusted order column when I wrote 17. That's because I looked at the 17 in the seed column by accident instead of the player # column.

And here is the adjusted order (League order still takes precedence over P/W).

Old Order | Adjusted Order

Top 100
11 | 11
15 | 15
4 | 4
7 | 7
17 | 17
2 | 2
20 | 20
19 | 19
21 | 21

Master
1 | 13 *
10 | 10
5 | 5
13 | 1

Gifted
16 | 9
9 | 18
3 | 16
6 | 3
8 | 8
23 | 6
18 | 23

Clever
14 | 24
24 | 14
12 | 12

I actually think this would be a better order.

This is based on Points/Win. Obviously the problem is that if you've played more 2v2s then your rating will suffer some, but I think this will ultimately still be a much more accurate indication of skill since many people haven't played 2s and out of those who have, they'll probably have played quite a lot more 1s than 2s. Raw points means absolutely nothing as far as skill goes so why not use a system that does factor skill in in at least some way than one that neglects skill almost entirely. (I say almost because obviously if you lose EVERY game you'll have 0 points, but anything over 0 points doesn't say much).

I'm happy to do the calculations for the full player list when it is ready. (I have it set up in spread sheet already so it will be easy)

*Note that 13 has very few games played however so I'm not sure about this one. Everyone else has 300+ points though so aside from the aforementioned 2v2 issue I think this should be a fairly good representation of W/L (and as a result, relative skill).

tldr; Regardless of how good or bad an indication of skill P/W is, it's better than raw points which is no indication of skill.

[awpertunity]

(09-01-2012 05:54 AM)CombatEX Wrote:  I don't understand this

8 - Gifted - 369 points - 47 wins
23 - Gifted - 431 points - 86 wins
18 - Gifted - 381 points - 42 wins

Surely 18 should be above both 8 and 23.

You're definitely right about that, thanks, I've fixed it on my copy.


As for the points/wins ratio, that was actually the way I intended to do it originally, and here is the reason I didn't in the end:

For a massive pool of players, I agree, points/wins would be the best way to quickly compare them. But for 32 players, I thought just doing it case by case as I've been doing it was better because of two things. One was the one you mentioned that 2v2 games are included in total wins which deflate the person's skill, but also to the people at 80+ wins. The points tend to get saturated (this is from my own experience only). When I started, I quickly rose from 0 to 1200 points in my first 60-80 games. In the last 60 games I've won I've only increased from 1200 to 1300 or so. Obviously this is because I am now losing games as well, but also I don't get as many points for wins anymore. But according to just points/win ratio, the more games I play at this point my "skill" is supposedly decreasing which is definitely not true.

The points/wins gives the best score to the people at around 40-50 games that are winning all their easy games and getting around 20-32 points for each win. No matter who I play at this point, I cannot get more than about 18 points for a win, and so points/win is no longer relative.



Doesn't matter too much to me either way, we can let people decide here if people care one way or the other. If people want a more structured way of calculating seed then points/win makes sense. But this definitely is not a true indication of skill.

[aaronINdayton]

I don't care how you seed beside the top 100 ranks being ordered as listed already. There's no prize money, right? I don't think anyone'll care too much how you seed

[CombatEX]

(09-01-2012 06:49 AM)awpertunity Wrote:  

(09-01-2012 05:54 AM)CombatEX Wrote:  I don't understand this

8 - Gifted - 369 points - 47 wins
23 - Gifted - 431 points - 86 wins
18 - Gifted - 381 points - 42 wins

Surely 18 should be above both 8 and 23.

You're definitely right about that, thanks, I've fixed it on my copy.


As for the points/wins ratio, that was actually the way I intended to do it originally, and here is the reason I didn't in the end:

For a massive pool of players, I agree, points/wins would be the best way to quickly compare them. But for 32 players, I thought just doing it case by case as I've been doing it was better because of two things. One was the one you mentioned that 2v2 games are included in total wins which deflate the person's skill, but also to the people at 80+ wins. The points tend to get saturated (this is from my own experience only). When I started, I quickly rose from 0 to 1200 points in my first 60-80 games. In the last 60 games I've won I've only increased from 1200 to 1300 or so. Obviously this is because I am now losing games as well, but also I don't get as many points for wins anymore. But according to just points/win ratio, the more games I play at this point my "skill" is supposedly decreasing which is definitely not true.

The points/wins gives the best score to the people at around 40-50 games that are winning all their easy games and getting around 20-32 points for each win. No matter who I play at this point, I cannot get more than about 18 points for a win, and so points/win is no longer relative.



Doesn't matter too much to me either way, we can let people decide here if people care one way or the other. If people want a more structured way of calculating seed then points/win makes sense. But this definitely is not a true indication of skill.

Ah. Fair enough. Point saturation didn't occur to me, though I feel that this is mostly dealt with by the fact that these people with many games are for the most part in the top 100 already. But anyway, I see how it could be an issue. As I said though, while p/w is only a partial measure of skill, raw points is no measure. Some is better than none right? =P But yea, it doesn't *really* matter so I'm fine with whatever you ultimately decide. I guess you can say, "some is better than none, but is the 'some' enough to bother with in the first place?" =)

Again though, I'm happy to do the p/w spreadsheet if we do decide to go with that so don't worry about having to deal with that (although it's probably inconsequential compared to running the whole tournament).

[RandyDogz]

To beat a dead horse, just cause I find this interesting, basic p/w is skewed down when you don't have alot of wins and really compounded if you have ranked 2x2(s). Cause you get no points for placement matches, if I remember correctly. So with ranked 1x1 and 2 2x2, I have 15 no points wins, I believe, out of 67. That brings me down ~2 p/w. If I had more wins, assuming a proportional ppg, the 15 are diluted more, and vice versa. I think. I just made my brain hurt.

And for full disclosure I reported my total points, not 1x1 only, cause that made a huge diff in p/w too. I've been watching this for a while now to try and get idea of how I stack up. It's a real rough proxy, and even rougher after some of the points raised here that I hadn't even considered, but I need something to quench my obsessive need for a way to gauge if I'm the favourite or the underdog.

With all that, I don't care how I'm ranked here - just hoping to hang around long enough to learn a little!

[Necrocat219]

Hehe, I did know that as you climb until you reach your peak your points begin to stay the same, but i suggested it because I've only played 1 vs 1 ( in 2 vs 2 it takes ages to finish a game :O)

And btw that is a neat table combat :3

[awpertunity]

Okay, the Group Play rules have been posted. I ran out of room on my initial post so I need to post an extra blurb here:

I am hoping there will be no ties beyond the 4 categories posted... If there is, and there are many games left to be played in the Group Stage I would prefer that the two players play one more match (2 games) against each other. But if not, at this point we will have to choose some way of declaring a winner to move onto the Second Stage. I am open to suggestions here, either higher seed, lower seed, or rock paper scissors on forum (I don't know, lol).





Tournament Round Rules
These will be announced in detail when we are closer to this stage.


The only things so far are that:

Maps have been pre-selected and will be displayed on the tournament bracket once I post the final Groups.

Each round will consist of one match (2 games). In the case of a tie, the player with the higher Game Point Differential (for this match only) will move on.

In the case of a tie, this is where I need suggestions! I really hate the idea of a single 3rd game as it gives whoever is Player 1 too large an advantage. I was thinking for semifinals and finals at least that they would play one additional match on the previous round's map, but if we do this every round this will be a very long tournament. The other option is to compare some sort of Group Standings result... but it may be hard to compare two players' stats in different groups. Suggestions are welcome, we got time!