EL2011 - About the Scoring System

what do you think about this formulas?

BONUS: winner gets 100 points, draw gets 50 points, just same as before.

A: Won points from finished games

B: Lost points from finished games

T: Total number of games (finished or not doesn't matter)

 

SCORE = BONUS +   {CubeRoot of [A * (A+B) * T]} / 5

or; if it isn't possible to use the first formula with excel, here is another option

SCORE = BONUS +   {SquareRoot of [A* (A+B)]} / 5

---

So, even a match where A=0 and B=0 would give a bonus?

I understand your idea, but the formula may would look like this instead:

SCORE = (F/2P) * (BONUS +  (CubeRoot of (W * F * 2P)))
or
SCORE = (F/2P) * (BONUS +  (SquareRoot of (W * F)))

It's possible to calculate SquareRoots in Excel by using SQRT, but I don't know how to calculate CubeRoots in Excel.

But I believe an even better options may could be these:

SCORE = (F/2P) * (BONUS + (SquareRoot of (W * 2P))
or
SCORE = (F/2P) * (SquareRoot of (W * 2P))

These would work pretty fine.

However, I still belive that the formula SCORES = (F/2P) * (2P + W - L) would work fine and may even better?

Another option that are pretty similar, and also work fine, could be:

SCORES = (F/2P) * (BONUS + 2P + W - L)

All these - highlighted in bold - are pretty similar options.

Hi, I've covered that the Square-/CubeRoot-based formulas don't work so fine, since they awarding tied matches more than less tied matches.

Here's an example on basis of the formula SCORES = SquareRoot of (2P * W):

As you can see, the sum of scores for two teams in a match is much better in a tied match, than in less tied matches.

I believe the system must adwarding all matches (of same sizes) equaly.

A scoring system like SCORE = (F/2P) *({BONUS} + (2P + W - L)) seem to make better sense from this point-of-view, since it awarding teams that having a better winning ratios in their matches more than other ones. There're actually may not necissary applying any extra bonuses in this case, since W - L is built-in  bonus itself.

Here's almost same example but using the formula SCORE = (F/2P) *({BONUS} + (2P + W - L)):

However, I will trying to find out if there may could exist any variation of your Square-/Cube-based system that may would work better.

to be more clear, my point is that the number of players should never be used as divisor. (F/2P) ?!

otherwise the more players, the less results.  that's why i'm always trying to use multiplications and additions.

if A=0 or B=0 it can't give a bonus. how can it happen?

 

squareroot is working in excel, but i couldn't find the cuberoot. maybe it can be added manually but not so much necessary.

that formula with squareroot is also working fine in my opinion.

---

F/2P = the share of games that are finished in the match.
You have to, otherwise would the unfinished matches be too much awarded compared to the finished ones.

Example (when using the formula SCORE = SquareRoot of (2P x W)

In Case 1 is 100% of the games finished. The result is 30-20 in favour to Team A.

In Case 2 is 50% of the games finished. The result is 15-10 in favour to Team A.

In Case 1 have Team A+B get 70,35 scores together. In Case 2 have Team A+B get 49,75 scores together even if they only have finished 50% of the matches.
It seem more reasonable that the teams in Case 1 should had get a score below 35 instead of around 50.  Unfinished games are too much awarded in your model at the same moment as tied games also are awarded. Your model by that awarding unfinished tied games much more than finished games for cear winners. You've proposed to give a compensation by an extra bonus for wins. But also in that case are you proposing that teams that still may have a lot of unfinished games would get the same extra bonuses as the ones that may already have finished their matches.

So, your model too much awarding teams that are able startup large but tied matches without finish their matches, while you disawarding teams that starting up and winning smaller games that finish faster.

I agree that it's reasonable awarding larger matches more than smaller matches, but I disagree that tied and unfinished matches should be better awarded than finished and cöear victories.

******
A= 0 and B=0 when a new game is starting and no games yet have finished.

As soon as a game have finished will that change of course, but nevertheless we doesn't want to promote newly started matches soo much compared to the finished ones. The final scores for a finished match must be larger than the current scores for a unfinished match in average per participating player. It's important to ensure that.

******

As mentioned, I cannot see that the formulas with square-/cube-roots working satisfactory, since they give to much scores to tied and unfinished games compared to finished and clear victories. However, they may are able to improve!?

Ні. Мy idea. Win match - 2 points for team Draw - 1 point Lost - 0 point. % of individual results in match is the second demonstration. It is a universal system, as fom me.