# Winning percentage calculation

Author: DebateArt.com

## Posts

Total: 10
Debates: 0
Posts: 1,403
3
3
8
3
3
8
Hi guys!

I wanted to ask your opinion on the winning percentage calculation. So now, I use a possibly naive approach when I simply calculate by the formula:

(won debates / all debates) * 100%
And @omar2345 made a notice about it in the platform development, suggesting that we exclude the tied debates from the calculation and I guess that's one way to go.

The other way to go is to use the formula from here, where the ties are considered as 0.5 of a win.

To provide some examples, imagine two participants that had 5 debates against each other where 2 debates are won by the first participant and the rest are a tie.

Now, the winning percentages would be like this, respectively:

1) Current approach -> 40% and 0%

2) Omar's approach -> 100% and 0%

3) Wiki's approach ->  70% and 30%

So what do you think?

My personal view is that the Wiki's approach makes sense and is better than what we have now, but maybe there even better alternatives.
Debates: 32
Posts: 10,666
4
7
11
4
7
11
-->
@DebateArt.com
wiki
Debates: 43
Posts: 2,768
6
9
10
6
9
10
-->
@DebateArt.com
It all depends on what you want to show.

I think the win percentage showing the number of debates you have won is good.

Omar’s idea, is the percentage of debates you didn’t lose (which is not the same).

I think you have two real options. Leave it alone, or make it a WLT percentage:

In your example this would show

”40% / 0% / 60%
Debates: 30
Posts: 14,911
5
8
11
5
8
11
Ties at .5 of a win
Debates: 44
Posts: 4,920
4
9
11
4
9
11
-->
@DebateArt.com
My approach.

A tie is neither a loss or a win. Instead of calling it a win percentage it should be called a win/loss percentage.
Debates: 24
Posts: 618
3
4
7
3
4
7
-->
@TheRealNihilist
The ultimate question is this - How good is someone at debating? Well, the numbers can help with that, but when there are a significant number of ties it tends to muddy things up when it comes to win percentages. Perhaps to do as omar2345 suggests, but to additionally report tie percentages right next to the win/loss percentages. If a user has, for example, a 100% win/loss percentage but an unusually high tie percentage it would be somewhat of an asterisk, but perhaps a deserved one.
Debates: 44
Posts: 4,920
4
9
11
4
9
11
-->
@Death23
The ultimate question is this - How good is someone at debating? Well, the numbers can help with that, but when there are a significant number of ties it tends to muddy things up when it comes to win percentages. Perhaps to do as omar2345 suggests, but to additionally report tie percentages right next to the win/loss percentages. If a user has, for example, a 100% win/loss percentage but an unusually high tie percentage it would be somewhat of an asterisk, but perhaps a deserved one.
Mainly the reason why I wanted to do it because when looking at a win percentage. Someone can win all their debates but have a ton of ties. When people look at they assume they are a bad debater even though that person drew that reduced their win percentage.
There is already a section dedicated to wins, losses and ties. I guess a person can hover over person's win percentage and something like the Big Issues of DDO a box opens up stating how many times they have lost, won and drew but I doubt that is necessary because if the person just looked left a bit they would see the wins, losses and draws.
Debates: 24
Posts: 618
3
4
7
3
4
7
-->
@TheRealNihilist
Hmmm I think you right. It already says the ties. The percentage there... I don't know.
Debates: 42
Posts: 2,472
5
7
11
5
7
11
-->
@DebateArt.com
Ties at .5 wins sounds better then the established method.  Either that or a strict win: loss ratio with ties not counting for either side.
Debates: 0
Posts: 1,403
3
3
8
3
3
8
I have already changed the formula to the one from Wiki :)