Clanner point system.

[TheKomodo]

cFc recently lost 1-2 overall, we won 5 points, and lost 141 points.

It's cool to lose to Cw, they are nice clan, and improving alot, but to lose 136 points, for 1-2, 3 games in a row, it just doesn't feel right, I am not complaining I just have an idea, why can't we change the point system for Clanners?

Like lower the middle point thing, what's it called? ELO Range or something? Is it 40? Instead of losing/winning up to 80 points, we should lower it to 25-30? So points to win/lose is 50/60?

For singles it's ok for losing 80 points because the individual is responsible and average win % are much higher in Singles than Clanners. I mean when I lose BnG, to anyone, I will lose 70+, but that's ok because my win % reflects this. But because clans have lower win % maybe we should lower the ELO Range in Clanners to reflect the average win %?

[Random00]

The overall rating should reflect the clan's skill in this scheme with the clan's average team.
The current system is exactly doing this, so its fine. There has to be a peak that you can reach as long as you dont have 100% winnings (which I doubt anyone will ever have). ;d

[TheKomodo]

The overall rating should reflect the clan's skill in this scheme with the clan's average team.
The current system is exactly doing this, so its fine. There has to be a peak that you can reach as long as you dont have 100% winnings (which I doubt anyone will ever have). ;d

Sure, but even CF complain about the amount of points lost even when you are winning games.

3 games, and winning 1, and still losing 136 points, that's a joke mate lol.

Especially when cFc really only have all these points because we play like 50% more than the rest of Wnet lol.

[darKz]

How exactly is changing that ELO value to say 30 going to fix your problem? It's the same shit with different numbers.

Imo the system is perfectly fine in that way, if you got lots of overall points every loss is painful. May be discouraging at times but it's certainly not unfair.

[TheKomodo]

Well yeah I thought of that, but balancing out the win ratio, it's still better I think in the longrun.

[Husk]

The ratings shouldn't be sky high, it is there doing its job showing the top players/clans. Like in our case it shows top clans CF and cFc just like it should work. I know its annoying to lose 50x times more than from a win, but hey if other clans are not as skilled as u guys, dun worry they wont be as high rating as u r =D

[TheKomodo]

Yeah but remember I said Cw are cool haha they deserved the wins, I was focusing soley on the swapping of points being way out of proportion for a 1-2 defeat, we win 5 points, and lost 141 points, think about it? That much of a point difference, only because cFc play literally 3/4x more than any clan, we don't have a high win %, jus cuz we play more, does this make sense?

1       CF    710    526    184    74.08%    +3    +24/-4    6712    [Elite]
2       cFc    2036    1339    697    65.77%    -2    +22/-11    6508    [Elite]
3       ps    1309    763    546    58.29%    +2    +19/-11    5250    [Field Marshall]
4       doH    576    386    190    67.01%    -5    +15/-5    4953    [Major]
5       TaG    1700    903    797    53.12%    +1    +17/-17    4590    [Highly Distinguished]


Look at the games played/win % and points, doH should be higher than us, ps (in recent 1 year) should be above us with their recent win %.

CF are pretty true, their games played win % still being higher than cFc with 2036 games played and 1339 won.

You know what i'm going nowhere with this lol, you are all right, stupid idea lol...

[franz]

The system makes your own Overall Rating and Win% not mean as much as they did in the past.
Focus on your season rating, which reflects your performance in this current season. Overall just shows how you've done historically.
Also don't focus much on Win%. It would only mean anything in an older system where everyone wins the same amount per game.

[DarkOne]

I don't think changing the points from 40 to 30 or 25 would makes a difference with the problem you're showing, Komo. Let's take the example you show here

Lose 141, win 5 is the case right now.

If we change it from 40 to 25, that's 37.5% less. (or 62.5% of the original; 40*0.625 = 25)
If we apply the same change, you'd win 0.625*5 = 3.125 points
You'd lost 141*0.625 = 88.125 points.
Overall, that's a loss of 85 points.

Yup, that's less points Still a large amount. But the problem lies somewhere else.
Let's look at the score you put up: CF has 6712, cFc has 6508
Now, the starting point is 1000. This means CF is 5712 points above it. Now these points also are different
5712*0.625 = 3570, so CF's new rating becomes 4570.
cFc's new rating becomes 1000 + (5508*0.625) = 4442.5

At first sight, the difference between the two clans looks smaller.
But let's see how big a chance it really is
Say you need 10 wins to bridge the gap. 204/10 = 20.4 points per win (current situation)

Now, we go to the new situation: the difference is now 127.5
The amount of points per win is 20.4 * 0.625 = 12.75
So even though the difference is smaller, you'll still need 10 wins.

As for the win percentage, franz pretty much sums it up in his last line win percentage only tells you something if you know who they played against in those very games. Scoring 50% against Random is a far better result than scoring 50% against me
So it's fair for there to be a better reward for scoring 50% against Random than for scoring 50% against me.

As for having more points by having played more games: I'm afraid that if we want to solve that (not sure whether it's a real problem or just cFc playing more difficult opponents here), a simple shifting of numbers isn't going to do the trick (for the reason I explained above), we'd need a new overhaul of the rating system entirely.

[TheKomodo]

As for the win percentage, franz pretty much sums it up in his last line win percentage only tells you something if you know who they played against in those very games. Scoring 50% against Random is a far better result than scoring 50% against me
So it's fair for there to be a better reward for scoring 50% against Random than for scoring 50% against me.

Maybe for specific players, but in this case it's a large variety.

In fact, the thing that gets me, is we only lose so much points to clans like Cw because they play less and we play more, they are a good clan, like I said they are improving alot, and fast, guuria is going to be a top 10 pretty much the way she's going.

That's why I feel sort of point raped, just cuz we've played over 2000 games we have so many points, and end up losing ridiculous points even though we get scores like 2-2 1-3 1-2 1-1 etc, even when we win like 3-1 we still sometimes lose 20-30+ points, If we win 3-1, we should gain points, if it's a weaker clan then fair enough, but surely we shouldn't be losing "overall" standings points, when we actually win 3-1 overall... That's the kicker for me lol.

[Random00]

Many games doesn't necessarily mean many points.

Phanton: Games: 4043 Win%: 65.64% Points: 3768
pr: Games: 323 Win%: 65.63% Points: 3867

It may be that cW still haven't reached their skill level, because they didn't play enough yet. But this imbalance will level out in the long run ;d
The current overall rating system has some problems with new players, because their start rating most likely is not there skill rating. But system's that cover these problems will most likely have some major other problems, so it fine like it is

[franz]

Focus on your season rating Komo, as that's the most important.
If you could hide overall rating, I think you would benefit from it.

Consider the points cW wins as encouragement, making them play more.
Help fellow league mates to enjoy the league instead of hogging all points ;]

If cW does deserve to be higher, then in the long run they will get there.
Encourage cW and play them so you might win their future points. if you can ;p

[DarkOne]

It may be that cW still haven't reached their skill level, because they didn't play enough yet. But this imbalance will level out in the long run ;d

I think that's the main problem here Perhaps it could be a good idea that for the overall rating (not the season rating, obviously) there's a multiplier in play; for the first 10-20 games, the amount of points won (or lost!) is multiplied by a certain amount, so they will reach their real overall rating quicker, leading to less problems with older clans losing points.

I may not be expressing myself very well, so I'll give an example

Clan PRO is a new clan, but the players that are in it are already very good. Their starting rating is 1000.
Now, clan PRO shows exactly how good they are and win the first 10 games. Each game wins them 60 points on average. So now, their overall rating is 1600 and their seasonal rating is 1600.
No problem with the seasonal rating, but their overall rating is clearly too low.
Now let's add in a multiplier I'll just a random number out of my ass: 4.
For the seasonal rating, nothing has changed, so their seasonal rating is 1600.
For the overall rating, their points are multiplied, so they win 10 times (4*60) = 2400 points. So their new overall rating is 3400 points. From that moment on, ratings are calculated as they should be.

So what has changed with regards to seasonal playoffs?
For PRO: Nothing. The amount of points they get for their seasonal rating depends only on their own seasonal rating, so no problem there.
For PRO's opponents: Well, for the first couple of clans nothing. But since the overall rating quickly goes towards the level that it should be, they lose less seasonal points.
But that seems to solve the problem Komo points out; Clan PRO is so strong that losing a lot of points for losing against them seems unreasonable.

I wonder what you guys think about this Let's discuss!

[darKz]

A new clan's initial rating could also be determined by the average singles ratings of every player of the clan. But this could also lead to other problems, I'm not sure.

[Rok]

I wonder what you guys think about this Let's discuss!

I think what your describing is the same as ratings deviation: http://en.wikipedia.org/wiki/Glicko_rating_system.

I proposed this a long time ago (and more than once), but it always gets lots in opinion-jungle

A new clan's initial rating could also be determined by the average singles ratings of every player of the clan. But this could also lead to other problems, I'm not sure.

Fishy, IMO. Singles and clans are two different leagues and not everybody plays both. So it's inaccurate from the beggining.