% won vs Diff. opp. #

[chapcrap]

http://www.conquerclub.com/forum/viewtopic.php?f=6&t=108862#p2444088
(Game-type neutral percentages)

Courtesy of Fruit's nifty spreadsheet:

Total average weighted win rate % 44.21%
Your average weighted win rate % 49.80%

5.6% difference

approx

me:

What are these percentages? I don't understand them. How do I find mine?

[Rodion]

What are these percentages? I don't understand them. How do I find mine?

Assume you only played 5 games in your CC history: 1 triples and 4 5-player games.

You'll get 4/5 x 20% + 1/5 x 50% to find out you're expected to win 26% of your games.

Now do it again, except this time you'll consider all the 7422 games you have under your belt, and they'll have to be divided between those 7 groups (12.5%, 14.3%, 16.6%, 20%, 25%, 33,3% and 50%), all weighed considering how many times they've been played.

[chapcrap]

What are these percentages? I don't understand them. How do I find mine?

Assume you only played 5 games in your CC history: 1 triples and 4 5-player games.

You'll get 4/5 x 20% + 1/5 x 50% to find out you're expected to win 26% of your games.

Now do it again, except this time you'll consider all the 7422 games you have under your belt, and they'll have to be divided between those 7 groups (12.5%, 14.3%, 16.6%, 20%, 25%, 33,3% and 50%), all weighed considering how many times they've been played.

Thanks. I understand now. Should have understood earlier.

[Robinette]

me:

That is like some kinda Bandit awesome numbers you got there DiM



Ok, my turn...

Robinette's expected weighted win rate % is 18.6%
Robinette's actual weighted win rate % is 28%

+9.4%... 51% higher than expected...

Maybe i didn't do this right
Could i really be that much higher than expected???

actually,, even if this is right, all that it proves is that i'm luckier than most,
besides... i've only played 668 games, which is a pretty small sample compared to a lot of folks,
so perhaps we should expect anomalies like this..

[Leehar]

Lol, it hasn't changed much from the last thread Robin

[DiM]

me:

That is like some kinda Bandit awesome numbers you got there DiM



Ok, my turn...

Robinette's expected weighted win rate % is 18.6%
Robinette's actual weighted win rate % is 28%

+9.4%... 51% higher than expected...

Maybe i didn't do this right
Could i really be that much higher than expected???

actually,, even if this is right, all that it proves is that i'm luckier than most,
besides... i've only played 668 games, which is a pretty small sample compared to a lot of folks,
so perhaps we should expect anomalies like this..

thanks but frankly these numbers are just for fun. they don't mean much in the great scheme of things.
without map/settings or opponent rank taken into consideration, they don't really tell too much about your value as a player. and aren't even good for comparing 2 players.

[SimplyObsessed]

Where do I find out the numbers? I installed the map rank in the first post and it's working and i'm rather confused

[Fruitcake]

I rarely come in here these days.

I have a spreadsheet, set up so all you need do is plug in your games played and wins for each possible style of game in terms of opponents etc...thus:

2 player singles
4 player doubles
6 player triples
8 player quads
3 player singles
6 player doubles
4 player singles
8 player doubles
5 player singles
6 player singles
7 player singles
8 player singles

What the spreadsheet does is works out the number of games to the overall total (so if you had played 5 games of 2 player singles and played 100 games in total the spreadsheet would 'weight' your win rate to 5% of the overall total). It then applies what should have been the win rate (in this case 50%) and totals automatically at the bottom (per DiM's picture).

From this spreadsheet you can see where your strengths are (my win rate in triples is 73%, my best) and your weaknesses are (my 7 player games has a 0 win rate, but I have played this format only rarely so it does not affect the overall win rate total much)

The spreadsheet allows for the fact that some don't play certain formats at all (looking at you Robin) by ignoring a zero played sum.

here is an easy example of how the sheet works:



I just applied an easy 10 games/5 wins to each format.

Summary of each column heading:

Style of game is self explanatory.
Number played is self explanatory.
Number of wins is self explanatory.
Average win rate shows what the average should be for that format (1v1 would be 50% whilst you should win 12.5% of 8 player games to be average)
Your win rate shows your actual % win rate
Weighted number of games. This applies the number of games to the whole. In this case they are all 8.33% of the total (see above for my earlier written example...the spreadsheet automatically adjusts to comply with totals entered, see spreadsheet example below)
weighted average win rate. This shows the weighted % as a part of the total. if you look at the above sheet this player has an 4.17% weighting for 2 player singles because he/she should have won half those games (so 8.33 divided by 2) However, as can be seen below, should the number of games (when compared to the total) change, so does the weighting.
Your weighted average win rate this does the same as the previous column, but applies YOUR win rate.

The sheet auto totals at the bottom to give you your overall reading against the norm for the number of games you have played in each different format.

here is an example of some one who obviously likes to play 7 player singles a lot. You can see from this that the weighting to that format is increased dramatically from 8.33% of the total games to a massive 36% of the total with a commensurate change in the totals at the bottom (because the average win rate would obviously drop at this point due to one not expecting to win as many 7 player games as perhaps 1v1 games!)



Sorry if this has come out as some kind of concious stream of thought but whilst numbers and my brain seem to live in harmony I often find it difficult to articulate complicated processes to others...hopefully many of you will get the picture and be able to assist others.

I am happy to pass this sheet on to anyone who wants it, just drop me a PM with an email addy and I'll flash it over.

[Fruitcake]

Actually did my own this evening.



This is a classic instance of someone who really focuses on one style of play. My triples weighting is huge because 34.5% of my total games is played in this format with a win rate of 72.8%.

[ljex]

Actually did my own this evening.



This is a classic instance of someone who really focuses on one style of play. My triples weighting is huge because 34.5% of my total games is played in this format with a win rate of 72.8%.

would you mind sending me the file? or could i send you a file with my data and you plug it in and send me the results back?

[Chariot of Fire]

Does it involve physically counting all my win/loss figures for each game type to be able to enter the info into the spreadsheet, or is there a tool that does that too? I don't use any add-ons so I'm pretty green in respect of what's needed. Cheers

[AAFitz]

Does it involve physically counting all my win/loss figures for each game type to be able to enter the info into the spreadsheet, or is there a tool that does that too? I don't use any add-ons so I'm pretty green in respect of what's needed. Cheers

you need the addon Maprank to realistically do it....or, Ill just do it for you now.

Ok, I forgot this takes some time. Its 3am here, so Im sure someone will take care of it for you, if not, Ill try to grab it in the morn.

[AAFitz]

I rarely come in here these days.

I have a spreadsheet, set up so all you need do is plug in your games played and wins for each possible style of game in terms of opponents etc...thus:

2 player singles
4 player doubles
6 player triples
8 player quads
3 player singles
6 player doubles
4 player singles
8 player doubles
5 player singles
6 player singles
7 player singles
8 player singles

What the spreadsheet does is works out the number of games to the overall total (so if you had played 5 games of 2 player singles and played 100 games in total the spreadsheet would 'weight' your win rate to 5% of the overall total). It then applies what should have been the win rate (in this case 50%) and totals automatically at the bottom (per DiM's picture).

From this spreadsheet you can see where your strengths are (my win rate in triples is 73%, my best) and your weaknesses are (my 7 player games has a 0 win rate, but I have played this format only rarely so it does not affect the overall win rate total much)

The spreadsheet allows for the fact that some don't play certain formats at all (looking at you Robin) by ignoring a zero played sum.

here is an easy example of how the sheet works:



I just applied an easy 10 games/5 wins to each format.

Summary of each column heading:

Style of game is self explanatory.
Number played is self explanatory.
Number of wins is self explanatory.
Average win rate shows what the average should be for that format (1v1 would be 50% whilst you should win 12.5% of 8 player games to be average)
Your win rate shows your actual % win rate
Weighted number of games. This applies the number of games to the whole. In this case they are all 8.33% of the total (see above for my earlier written example...the spreadsheet automatically adjusts to comply with totals entered, see spreadsheet example below)
weighted average win rate. This shows the weighted % as a part of the total. if you look at the above sheet this player has an 4.17% weighting for 2 player singles because he/she should have won half those games (so 8.33 divided by 2) However, as can be seen below, should the number of games (when compared to the total) change, so does the weighting.
Your weighted average win rate this does the same as the previous column, but applies YOUR win rate.

The sheet auto totals at the bottom to give you your overall reading against the norm for the number of games you have played in each different format.

here is an example of some one who obviously likes to play 7 player singles a lot. You can see from this that the weighting to that format is increased dramatically from 8.33% of the total games to a massive 36% of the total with a commensurate change in the totals at the bottom (because the average win rate would obviously drop at this point due to one not expecting to win as many 7 player games as perhaps 1v1 games!)



Sorry if this has come out as some kind of concious stream of thought but whilst numbers and my brain seem to live in harmony I often find it difficult to articulate complicated processes to others...hopefully many of you will get the picture and be able to assist others.

I am happy to pass this sheet on to anyone who wants it, just drop me a PM with an email addy and I'll flash it over.

Maybe you could just post the formulas for the calculations. Some or obvious, but its probably easier for you if people just make their own spreadsheets where possible.

[Leehar]

Hmm, xls can't be uploaded anyway

[AAFitz]

Ok, Let me dl it again and upload it onto the forums. We can't do it here, but I'm hoping the tech guys will be nice enough to let us keep the topic open for a bit so whoever wants to dl the Spreadsheet can do so.

Unfortunately, as I was map ranking my stats, map rank didn't differentiate between team games and regular games for some reason.

[Leehar]

Unfortunately, as I was map ranking my stats, map rank didn't differentiate between team games and regular games for some reason.

Yeah, you'll need the updated map rank I link to in the first post. GM should install sherks .js link immediately hopefully

[AAFitz]

Unfortunately, as I was map ranking my stats, map rank didn't differentiate between team games and regular games for some reason.

Yeah, you'll need the updated map rank I link to in the first post. GM should install sherks .js link immediately hopefully

Thank you for that, I was wondering why my team stats from before seemed off. I was horrified at them to be honest.

now I just need the formulas for the weighting, I may be able to figure them out, but better to use exact same ones

[AAFitz]

Game Type # Played # Wins X Rate Win Rate Weighted # AWR MWAWR
2 3179 5377 50% 59% 48.25% 24.13% 28.53%
3 273 677 33% 40% 4.14% 1.38% 1.67%
4 71 318 25% 22% 1.08% 0.27% 0.24%
5 47 242 20% 19% 0.71% 0.14% 0.14%
6 103 629 17% 16% 1.56% 0.26% 0.26%
7 3 30 14% 10% 0.05% 0.01% 0.00%
8 76 585 13% 13% 1.15% 0.14% 0.15%
D 1791 2498 50% 72% 27.19% 13.59% 19.49%
T 825 1111 50% 74% 12.52% 6.26% 9.30%
Q 220 287 50% 77% 3.34% 1.67% 2.56%
T 6588 11754 56% 100.00%



Total average win rate % 48%
Total weighted win rate% 62%

I think I did this right. Calculated the formulas backwards, but didnt double check...too tired.

I also skipped 6 and 8 player doubles.

[Leehar]

The win rate column: =IF(C5=0,"",(D5/C5*100))
Weighted number of games: =IF(C5=0,"",(C5/C18*100))
weighted average win rate: =IF(C5=0,"",(E5*G5))
Your weighted average win rate: =IF(C5=0,"",(F5*G5/100))
Total and your average weighted win rate %: =SUM(H5:H16)/100 And =SUM(I5:I16)/100

Lol k, nice work

[Fruitcake]

Sorted!

Go here and download:

http://www.filedropper.com/wincalculator

Fc

[ljex]

Sorted!

Go here and download:

http://www.filedropper.com/wincalculator

Fc

Thanks fruitcake, as a result of your file i get

Average win rate 39.48
My win rate 51.75

[nippersean]

me:

That is like some kinda Bandit awesome numbers you got there DiM



Ok, my turn...

Robinette's expected weighted win rate % is 18.6%
Robinette's actual weighted win rate % is 28%

+9.4%... 51% higher than expected...

Maybe i didn't do this right
Could i really be that much higher than expected???

actually,, even if this is right, all that it proves is that i'm luckier than most,
besides... i've only played 668 games, which is a pretty small sample compared to a lot of folks,
so perhaps we should expect anomalies like this..

Not sure you can count the 51% better Robinette. It looks correct, but could be one of those things the governments use.

So if an 8 player gamer wins 2/8 (very good for sure [and incredible if your name is Denise]), they couldn't be outscored (in your system) by a 1v1 / team player who wins 1000 out of 1000 - both get twice as many - 100% higher than expected?

What if you score 3/8?

Hopefully a maths bod (George?) can sort this out it's been bugging me all day and I doubt I can sleep.

Thanks in advance maths fellows, nipp

[chapcrap]

me:

That is like some kinda Bandit awesome numbers you got there DiM



Ok, my turn...

Robinette's expected weighted win rate % is 18.6%
Robinette's actual weighted win rate % is 28%

+9.4%... 51% higher than expected...

Maybe i didn't do this right
Could i really be that much higher than expected???

actually,, even if this is right, all that it proves is that i'm luckier than most,
besides... i've only played 668 games, which is a pretty small sample compared to a lot of folks,
so perhaps we should expect anomalies like this..

Not sure you can count the 51% better Robinette. It looks correct, but could be one of those things the governments use.

So if an 8 player gamer wins 2/8 (very good for sure [and incredible if your name is Denise]), they couldn't be outscored (in your system) by a 1v1 / team player who wins 1000 out of 1000 - both get twice as many - 100% higher than expected?

What if you score 3/8?

Hopefully a maths bod (George?) can sort this out it's been bugging me all day and I doubt I can sleep.

Thanks in advance maths fellows, nipp

3/8 is 200% better than expected. It works like this:

[[(Actual Win %) - (Expected Win %)]/(Expected Win %)]*100

In your example of 3/8 it would be:

• [[(37.5%)-(12.5%)]/(12.5%)]*100
  [25%/12.5%]*100
  2%*100
  200%

[Rodion]

So if an 8 player gamer wins 2/8 (very good for sure [and incredible if your name is Denise]), they couldn't be outscored (in your system) by a 1v1 / team player who wins 1000 out of 1000 - both get twice as many - 100% higher than expected?

Yes, the 2/8 guy can't be outscored in this stat, but the 1000 out of 1000 guy would outscore the 2/8 guy in other stats (such as the simple winning percentage).

There are so many stats that can be measured and used to support a discussion that it's not even funny. For instance, who's the best QB ever?

We have Joe Montana winning 4 Super Bowls.
We have Tom Brady passing for 50 TDs in his 2007 season and having a 9:1 TD:Int ratio in his 2010 season. Oh yeah, add 3 Super Bowls to all that.
We have Peyton Manning winning 4 regular season MVPs and 1 Super Bowl.
We have Eli Manning beating Tom Brady twice in the Super Bowl.
We have Drew Brees passing for almost 5,500 yards in the 2011 season.
We have Aaron Rodgers leading everyone else in passer rating.
We have Brett Favre with the most career touchdown passes.
Matt Ryan has the most symmetrical face of all current QBs.

And so on. There is always a stat that will make someone look better than the others.

[chapcrap]

There is always a stat that will make someone look better than the others.

I dare you to make Ryan Leaf look good.