Evaluation System

[RaneFire]

I'm going to add my 2cents now

I think that the normalisation margins are a great idea, but the flaw was it didn't address the types of maps and variation possibilities in the results. I didn't want to make it complicated, but it might look that way

The idea: Replace the 2nd Normalisation term with one that depends on the results of other players.
Basically the person who gets the middleman time, sets the half points margin. It doesn't have to be this, but I do think if you're going to invent a better points system, it needs to be tangible, not imaginary.

Flaws: (or considerations):

- At final rounds, smaller time gaps cause bigger point differences.
- If everyone gets standard times varying by 1sec up to the middleman, and 1st place is 2 secs faster, will result in massive point gain to 1st place, and loss of points to lower ranks.

Formula:

I'm just gonna work off arca's example as a template

Using the Middleman (Player at 50/100 position in the results):
(Mid time - Best Time)*2 = Factor

Formula: 100*(x-(b-a))/x
a = Best time
b = Time
x = Factor

Code: Select all

ROUND 0: (x = 2.688)

1. Arcaon - 17.984 - 100
3. Bazz - 18.200 - 92
5. glm - 18.312 - 87.8
46. Nt-kop4ik - 19.016 - 61.6

ROUND 1: (x = 12.960)

1. Nt-kop4ik - 32.608 - 100
3. glm - 34.280 - 87.1
4. Arcaon - 34.416 - 86
5. Bazz - 34.672 - 84.1

ROUND 2: (x = 16.272)

1. Nt-kop4ik - 36.736 - 100
2. glm - 37.888 - 92.9
4. Arcaon - 38.024 - 92.1
6. Bazz - 38.376 - 89.9

ROUND 3: (x = 6.832)

1. Arcaon - 35.464 - 100
3. Bazz - 36.440 - 85.7
5. Nt-kop4ik - 36.616 - 83.1
6. glm - 36.736 - 81.4

ROUND 4: (x = 3.168)

1. glm - 02.840 - 100
2. Nt-kop4ik - 02.992 - 95.2
12. Arcaon - 03.816 - 69.2
19. Bazz! - 04.312 - 53.5

ROUND 5: (x = 4.640)

1. Bazz! - 19.968 - 100
2. Nt-kop4ik - 20.368 - 91.4
3. Arcaon - 20.544 - 87.6

ROUND 6: (x = 5.376)

1. Nt-kop4ik - 35.768 - 100
3. Arcaon - 36.624 - 84.1

---SCORING--- (Assuming Round 0 is counted in points)

Scores After Round 3:

1. Arcaon - 378.1
2. Bazz! - 351.7
3. glm - 349.2
4. Nt-kop4ik - 344.7

Scores After Round 4:

1. glm - 449.2
2. Arcaon - 447.3
3. Nt-kop4ik - 439.9
4. Bazz! - 405.2

Scores After Round 5:

1. Arcaon - 534.9
2. Nt-kop4ik - 531.3
3. Bazz! - 505.2

Scores After Round 6:

1. Nt-kop4ik - 631.3
2. Arcaon - 619

Bad idea or?

[<)Ghost(>]

the normalization ranking system is probably the best idea here, buy as rainfire said, the problem comes from short maps such as round 4. only way to avoid this type of thing happening again, is to set a standard time frame for how long a map should be. to see the it gets mapped properly, tested properly, with no late admissions. iirc this was the problem for round 4. if the map isn't in by the due date kindly gtfo.

so i guess my point is we should use tusks system as long as we dint havv tr maps. BTW is it so hard to add a fucking wait of -1 to the start trigger??? i mean seriosly that should be mandatory to have in this years dfwc

sorry for any spelling errors i be on da phizone

[TittenIgnition]

BTW is it so hard to add a fucking wait of -1 to the start trigger??? i mean seriosly that should be mandatory to have in this years dfwc

i thought i read somewhere that that was gonna be a rule this year or something.

maybe i was dreaming?

[<hk>]

i thought i read somewhere that that was gonna be a rule this year or something.

Never heard of such a rule for mappers.
It would be sane to have a start timer with a wait of -1. Since it's not allowed to reset the timer it's not important to make this a rule.

I had this formular in mind. For computing the points.

Code: Select all

function potential = nppas( BestTime, PlayerTime, Window)
   % BestTime is always smaller or equal to PlayerTime.
   % Window is usually between 0.2 and 1.0. Default: 0.5
   % MaxPoints is the amount of poins the best Player gets.
   % n1 is the first and n2 the second normalisation point.
   n1 = BestTime;
   n2 = n1 + n1 * Window;
   
   % diff should be greater than or equal to zero.
   % potential should be between 0.0 and 1.0
   diff =      n2 - PlayerTime;
   potential = diff / (n2-n1);
end


[amt-morbus]

Since it's not allowed to reset the timer it's not important to make this a rule.

But telling mappers to put wait -1 prevents trouble. You wont get demos with a timereset. And its one fuckin key.

[<)Ghost(>]

Since it's not allowed to reset the timer it's not important to make this a rule.

But telling mappers to put wait -1 prevents trouble. You wont get demos with a timereset. And its one fuckin key.

^5

[Arcaon]

I'm going to add my 2cents now

Let's make babies.

Also, the flaws can be fixed, by for example normalizing rank 1 to give 100 points, and the middleman to have maybe 30 points. Would need a little tricky formula though, maybe I'll attempt it.

[TittenIgnition]

i thought i read somewhere that that was gonna be a rule this year or something.

Never heard of such a rule for mappers.
It would be sane to have a start timer with a wait of -1. Since it's not allowed to reset the timer it's not important to make this a rule.

but if there was wait -1, there would be less room to prerun, assuming there was room to prerun in the first place. as long as you dont COMPLETELY pass through the start timer, you wont get stuck on the player clip, and you can reset the timer as much as you want. it adds quite a bit of room for preruns.

im not really on either side, though tbh i would rather be able to do TRs, because theyre fun, but i cant so it doesnt matter. im just pointing this out.

[like@boss]

I'd just use avg rank over entire comp and be done. Obviously to be eligible for winning dfwc you must enforce a rule that all maps be completed.

otherwise it's all bullshit...gl finding a factor for map difficulty.

[<hk>]

one map is the same for all players so difficulty doesn't matter. It affects all players the same.