# Calculating a global rank from several ranks

5 replies
Goto Page
1
Nekomata
User
Offline
I'm terrible at maths. What's the approach for something like this?

Basically, consider you have 4 servers and you're ranked differently on each.

Example Dataset
Code:
1
2
3
4
Server 1 - 45 out of 500 players
Server 2 - 20 out of 7,928
Server 3 - 2 out of 32
Server 4 - 1,672 out of 17,425

How can I calculate a global rank from this data?
10.06.17 08:30:11 am
Talented Doge
User
Offline
In this case you will want to compare different aspects of the specific player. Like scores.

Rank is nothing at this point, calculate something rather meaningful and get the mean.

Like:
Code:
1
2
3
4
Server 1 - 500 scores
Server 2 - 1902 scores
Server 3 - 50 scores
Server 4 - 5555 scores

Hence the mean 2001.75. Then you compare with the rest of the players and get the rank.

Yes this approach is time wasting, but comparing rank is useless.
Talented Doge
10.06.17 08:51:02 am
eledah
User
Offline
Dump all the data sets into a bigger one. Sort them by their score. (if there are any duplicates, add up their scores in different servers to get one)
And then find the player you're looking for and extract his final rank.

Or, if you want it to be more accurate, you can add weight to each server's scores. Let's say that Server 1 is a deathmatch one, therefore getting kills will be easy and Server 2 is a standard server, so kills come by harder than usual.
In this case, you may count it like this:
Total <- SV1 + (3 * SV2)
[̲̅\$̲̅(̲̅ ͡° ͜ʖ ͡°̲̅)̲̅\$̲̅]
10.06.17 10:11:35 am
Nekomata
User
Offline
Okay, taking the mean from the score of all the servers and generating ranking on that.

Code:
1
(SV1 + SV2 + SV3 + SV4)/4 <- ranked by

And if I want to add weights, I can

Code:
1
((SV1*W1) + SV2 + SV3 + SV4)/4

Great, thanks.
10.06.17 11:14:42 am
tontonEd
User
Offline
divide by sums of weights
Code:
1
(SV1*W1 + SV2*W2 + SV3*W3+ SV4*W4)/(W1+W2+W3+W4)
10.06.17 11:39:45 am
Nekomata
User
Offline
Ah, thanks @ tontonEd
1
﻿