Rank Dist. Charts - Weekly Update 27-Mar-08

[stringybeany]

Update: New Forum and Rank Change Day
I grabbed the first five pages of scorecard data today. I'll post it with a comparison graph in a couple weeks.

[qeee1]

That's interesting, thanks for posting that, though I reckon the 2100>2200 jump is just there by chance.

[stringybeany]

That's interesting, thanks for posting that, though I reckon the 2100>2200 jump is just there by chance.

Maybe. Take several snapshot looks at that range over the next few weeks/months. If the step never normalizes, then clearly there are influences keeping it in place. If it smooths out or moves around, then it is more likely normal distribution.

[Blind Date]

I have had a difficult time getting past 2200..it might have something to do with thte scoring. I.E. Cant win as many points, might be in the formula?

[jiminski]

I 'm not certain that it is just by chance....

could there be a correlation between that and the point handicap reaching a plateau compared to ability?

Could there be a psychological implication due to converting the 2,000 points whereby there is a large pool who teeter up and down around the 2k mark?

Seems too pronounced to just coincidence.

[qeee1]

I 'm not certain that it is just by chance....

could there be a correlation between that and the point handicap reaching a plateau compared to ability?

Could there be a psychological implication due to converting the 2,000 points whereby there is a large pool who teeter up and down around the 2k mark?

Seems too pronounced to just coincidence.

That's an interesting thought. Maybe when people get above 2000 they suddenly think they're the shit and change something due to this, maybe the start joining the 2000 points plus games or just get lazier in their games.

[stringybeany]

...
Could there be a psychological implication due to converting the 2,000 points whereby there is a large pool who teeter up and down around the 2k mark?

...

Very possibly.

Again, I think the 500 point jump required to get from captain to major has an influence here also.

[comic boy]

Thats interesting as you would expect 1900 - 2300 to reflect a more even decline. Obviously above 2000 it gets more difficult but a lot of players do tend to hit the wall around the 2200 mark and I really dont know why that is the case.

[stringybeany]

You could also be looking at "rank effect".

Up until very recently I played all open games comprised of non-officers.

Every time I achieved lieutenant rank in these games, it was like I suddenly was wearing an "attack me" button. The shift is pronounced and clear. With the rank comes a perception of threat, I guess.

The same likely occurs in the transition from captain to major.

Anyone else experience the same?

[DiM]

I 'm not certain that it is just by chance....

could there be a correlation between that and the point handicap reaching a plateau compared to ability?

Could there be a psychological implication due to converting the 2,000 points whereby there is a large pool who teeter up and down around the 2k mark?

Seems too pronounced to just coincidence.

That's an interesting thought. Maybe when people get above 2000 they suddenly think they're the shit and change something due to this, maybe the start joining the 2000 points plus games or just get lazier in their games.

or maybe that's the average score of a good player.

there are excellent players that plateau at 3000
very good at 2500
good 2000-2200
above average 1700
average 1300


PS: i too am one of the players that could not possibly reach 2200. i topped at 2196. so close and yet so far but i know i'll do it soon.

[qeee1]

or maybe that's the average score of a good player.

there are excellent players that plateau at 3000
very good at 2500
good 2000-2200
above average 1700
average 1300


PS: i too am one of the players that could not possibly reach 2200. i topped at 2196. so close and yet so far but i know i'll do it soon.

Nah, I don't think that players skillfulness can be evaluated in jumps like that, I reckon it should follow a curve.

The rank as target point is more plausible I think.

[DiM]

You could also be looking at "rank effect".

Up until very recently I played all open games comprised of non-officers.

Every time I achieved lieutenant rank in these games, it was like I suddenly was wearing an "attack me" button. The shift is pronounced and clear. With the rank comes a perception of threat, I guess.

The same likely occurs in the transition from captain to major.

Anyone else experience the same?

they solution is to play only with similar ranks then you won't be a target.

but i've seen 2 sides on this story. in some games me being a captain i was targeted right from the start by people with lower ranks. perceived as a threat or simply wanting an elimination for big points (i play a lot of terminator). but in other games i've seen a totally opposite attitude. where people simply protected me when they saw i took a lead in the game. perhaps they thought of their score and wanted to minimize their loss. or perhaps they though i'm so good they should just make it easier for me and die fast because they don't have a chance.

this happened also when i was a low ranker. i was beginning to win a game and all of a sudden some guy suicided on me to allow the better ranked player to win. he even told me he's doing it to protect his score. i don't understand this and that guy made my ignore list. he screwed my game and made me lose 150+ points just to avoid losing 30 points to me. he decided it's better to lose just 10-15 points to the other guy.

[rebelman]

it would be interesting to see the same data minus all the ? as in reality many of these will have joined and given up again I suspect as a result you will see a flattening of the distribution around the 900.

[DiM]

or maybe that's the average score of a good player.

there are excellent players that plateau at 3000
very good at 2500
good 2000-2200
above average 1700
average 1300


PS: i too am one of the players that could not possibly reach 2200. i topped at 2196. so close and yet so far but i know i'll do it soon.

Nah, I don't think that players skillfulness can be evaluated in jumps like that, I reckon it should follow a curve.

The rank as target point is more plausible I think.

could be. a great way to test this would be to make adjustments to the ranking system. add more ranks and make them equally gapped. i'm curious how the graphs would change.

it could be a psychological effect. a guy with 1800 points looks at the ranks and says "hey captain is only 200 points away i can do it." and he does but when he reaches captain and looks at the ranks he says "damn that major is 500 points away and points are even tougher to get now i won't be able to do it" and he doesn't because he lacks the motivation even though he might have the potential to do it.

perhaps adding more ranks over captain to follow the 200 point gap pattern would make people try harder and we'll see a smoother graph.

[jiminski]

hehe! what a marvelous bloody site!
With our passion for collective analysis and strategy is there some way we can actually take over the world?

As you say qeee, It could be the 2k + games having an impact and this being the first official cut off point from the rest of the lower ranked players.

So there is a huge group of players playing lots of others of similar standard.. to a large extent just swapping points.

[TipTop]

Those graphs are just 1 snapshot in time. Although the theories being offered as to why there seems to be a barrier at 2100-2200 are interesting we really can't read much into it without further samples over time.

[jiminski]

no doubt you are right.... but it does seem to ring true and we can only offer snap-shot analysis until it can be corroborated as a genuine trend.

[Uberwald]

Well i noticed i was hovering around 2k points when i was playing public games aswell. Public games tend to be played vs lower ranks. Private games when youre a captain around 2k is mostly with other captain and some higher rankers. (and i mostly play standards or term but i have my fases of playing certain settings like now im into esc mostly except some tourney games)

So while playing public i couldn't get and stay above 2k then i stopped playing publics and i was able to rise above. Now i seem to be hovering around 2300.

This might be the reason for the 2k gap normal or good players simply can't win enough games while playing public and loose too much points to lower ranks.


.... or it's just a freack occurence

[stringybeany]

Those graphs are just 1 snapshot in time. Although the theories being offered as to why there seems to be a barrier at 2100-2200 are interesting we really can't read much into it without further samples over time.

Right. I'll take a few other snapshots over the course of weeks to see if the step stays in pace.

I suspect it will.

We'll see.

others up-thread started in making suggestions on changing the ranking system.

Personally, I don't see it as a problem so there is really nothing to address. It's just interesting, is all.

[Robinette]

Great Post stringybeany...

I thought I'd be revealing something to look at that by 50's, expecting to see this heavily weighted to to the 2100-2149 group, but it was 39 below 2150, 41 at or above.

This actually makes the oddity more pronounced...

And the rank explanation doesn't sit well with me to solve this one. I don't disagree that it has an effect, but these folks are well on the way to the center point between the two ranks.

As sully once demonstrated, we can expect point inflation to effect the numbers in a slow upward manner, so this likely means we can expect to see this 'clog' slowly move over 2200.

This will be most interesting to see comparatively over time. Perhaps someone can pull up sully's graphs and overlay them to this to see what clues it reveals.

Wish I had more time to ponder this, but I've got a plane to catch...
(sigh) *real life just gets in the way sometimes*
Have a nice weekend y'all.

[MeDeFe]

maybe a snapshot every day at 24 hour intervals, and then week to week evaluation and comparisons.

[stringybeany]

....

As sully once demonstrated, we can expect point inflation to effect the numbers in a slow upward manner, so this likely means we can expect to see this 'clog' slowly move over 2200.

...

Here's a moderately interesting case study:

I've played bridge on a certain web site for the last ten years. (yes. really.) The score calculation is similar to here, but after all that time the average rank hasn't increased as you would expect.

The reason is this; in a normal system you would expect the number of players that lose points and drop out to far exceed the players that win points and drop out, thus driving the average steadily upwards.

However on this particular site, cheating is common and unmoderated. Cheaters come in, amass enough points to reach the top, then lose interest and go away.

This has occurred enough to offset those that come in and lose and go away and has therefore kept the average rating stable for ten years!

Be thankful CC is so well moderated. The average score should continue to inflate at a nice steady pace for a long time as a result!

[stringybeany]

Oh, and let's not forget the reality here: You will note that the shape of the chart is slightly skewed and the average score is 900 instead of 1000 with a very long upward tail.

Let's just say this demonstrates that the top end has become adept and gleaning points from the middle, as expected.

[sully800]

Oh, and let's not forget the reality here: You will note that the shape of the chart is slightly skewed and the average score is 900 instead of 1000 with a very long upward tail.

Let's just say this demonstrates that the top end has become adept and gleaning points from the middle, as expected.

What I really find interesting is that the highest ranks earn most of their points from the lowest ranks, just indirectly.

If we theorize that people retain a relative skill level after learning the basics (which I think is mostly true) then you'd see people hover around a certain point. Sure they go up and down a few hundred, but in the long run everyone has a level that they don't go much above (unless they figure out a new effective strategy I suppose).

Taking that into account means that the people with 1000 points may beat a bunch of players with 700-900 points. That bumps them into the 1100-1200 range where they may lose to someone who normally hovers at that level. The 1000 player in turn goes back to the base level, and the points from the novices have moved up the ladder. That pattern continues all the way up the scoreboard until the very top few get pushed even higher onto their lofty platforms. And for such a system to perpetuate it takes a massive base of poor players, and a relatively small amount of exceptional players to indirectly steal their points.

(PS- I wouldn't go making too many graphs stringy. When Robinette sees the pattern she will ask you to make graphs of everything. When you oblige she will invite you to her games and steal your points. )

[stringybeany]

...
(PS- I wouldn't go making too many graphs stringy. When Robinette sees the pattern she will ask you to make graphs of everything. When you oblige she will invite you to her games and steal your points. )

I'm sure robinette and I will have our games, eventually. I wouldn't be too quick in assuming which way the points will go.

I'll have some of yours too, eventually.