[Archived] Why win percentage and not points per game?

[malk]

This is something that's been winding me up for a while. Why on earth are people quoting win percentage rather than points per game? Has football changed to an American style system where draws are virtually defunct? Do draws not give teams a point any more?

Of course wins are really important but, shockingly enough, that's accounted for by their weighting in points per game!

Putting aside the question of statistical relevance, Kean had a W% of 28.38 and ppg of 1.07. Appleton, at Blackpool, had a W% of 18.18 and a ppg of 1.18. Appleton at Portsmouth had a W% of 25.49 and a ppg of 0.98.

Now of course there is a valid argument that how a manager got their points/wins is important. Of the three I think it is relatively uncontroversial to suggest that Appleton at Portsmouth was the most difficult job and Kean with us was the easiest. Still, putting all that to one side, in terms of a metric for looking at managerial success win percentage is terrible.

Could we please stop using it!

[rhodie]

Yes but Kean was a kean!

[Salgado Is A Hero]

Agreed actually it baffles me as well......but there are always enough variables and mitigating circumstances to render most statistics largely meaningless imo.

[Elegia]

This is something that's been winding me up for a while. Why on earth are people quoting win percentage rather than points per game? Has football changed to an American style system where draws are virtually defunct? Do draws not give teams a point any more?

Of course wins are really important but, shockingly enough, that's accounted for by their weighting in points per game!

Putting aside the question of statistical relevance, Kean had a W% of 28.38 and ppg of 1.07. Appleton, at Blackpool, had a W% of 18.18 and a ppg of 1.18. Appleton at Portsmouth had a W% of 25.49 and a ppg of 0.98.

Now of course there is a valid argument that how a manager got their points/wins is important. Of the three I think it is relatively uncontroversial to suggest that Appleton at Portsmouth was the most difficult job and Kean with us was the easiest. Still, putting all that to one side, in terms of a metric for looking at managerial success win percentage is terrible.

Could we please stop using it!

So...... based on Appleton's ppg with Blackpool (and if that ppg average continues) we would have another 23.6 points to add to our current tally, at the end of this season. 57.6 points will not equal promotion and won't even put us into the play offs. And if we took Appleton's ppg at Portsmouth we would end up with 53.6 points.....

Over the course of a 46 game championship season that equates to 54.28 points using his Blackpool ppg average....or 45.08 points a season on his Portsmouth ppg average......

If we are to judge him based on ppg it doesn't read any better imho.

[ABBEY]

1 point per game =relegation

[Mike E]

If anything, it just goes to show the futility of statistics imo.

Appy, here's my advice:

1. If you can win the game, do it.
2. If you can't, get the draw.
3. If you can't do that, make sure you lose in a respectable manner.
4. But ffs, don't take option 3 too often.

[malk]

Agreed actually it baffles me as well......but there are always enough variables and mitigating circumstances to render most statistics largely meaningless imo.

It just rankles with me, ppg isn't exactly a difficult concept so I don't know why win % has gained traction. I suspect it has come from US sports somehow. I agree there are generally mitigating factors that should be taken into account and often heavily so.

So...... based on Appleton's ppg with Blackpool (and if that ppg average continues) we would have another 23.6 points to add to our current tally, at the end of this season. 57.6 points will not equal promotion and won't even put us into the play offs. And if we took Appleton's ppg at Portsmouth we would end up with 53.6 points.....

Over the course of a 46 game championship season that equates to 54.28 points using his Blackpool ppg average....or 45.08 points a season on his Portsmouth ppg average......

If we are to judge him based on ppg it doesn't read any better imho.

Sorry, I should have made it more clear. This isn't a comment about Appleton (or Kean for that matter), they were just a relevant recent example that showed the inadequacy of win %.

[LeChuck]

This is something that's been winding me up for a while. Why on earth are people quoting win percentage rather than points per game?

It's a good point. I guess it's because win percentage statistics are readily available on Wikipedia for pretty much any manager you can think of. Points per game statistics are very difficult to find.

I don't know why the sites that create the statistics in the first place favour win percentage, but until they change people are always going to quote them because it's a lot easier.

[yoda]

football is all about winning, so the only really valid statistic is your league position

[Steve Kean's Hypnotoad]

Know what you mean, logically it makes less sense to go off win percentage than ppg. Just a guess but it could be that win percentage is more catchy and easier for the mathematically-challenged to visualise. If you said a manager gets 1.7 ppg, that doesn't immediately mean much to everyone, you have to multiply it by 46 to get an idea of how good it is over a season. But if you say a manager wins 40% of his matches, then immediately you have a rough idea of how good that is. As an approximation its not too bad, as on average draws only make up a quarter of a teams points anyway.

[CapeTownRover]

Yeh, but if you go off win %, then you might also look at a % losing statistic. How many fewer games has somebody lost, as compared to somebody else. Eg, Appleton at about the same amount of games, has lost less games than Berg. So in theory, PPG is much better than win %. If you have been in charge of about 10 games, and won 2 games,drawn 3 draws and lost 5, you have a win % of about 20% and 9 points. However, if you are in charge of about 10 games, won 1, drawn 8 and lost 2, you only have a win% of 10%, but 11 points gained over the same amount of games. I would think that you would take more points, than win%?

Hahaha, I think I worked that out right, if not then I'm a dumbass lol

[Steve Kean's Hypnotoad]

Yeh, but if you go off win %, then you might also look at a % losing statistic.

That only makes sense if not losing games is equally as important as winning them, as would be the case if it was 2 points for a win. But with 3 points for a win, its more important to win games than it is to not lose them.

Appleton at about the same amount of games, has lost less games than Berg. So in theory, PPG is much better than win %. If you have been in charge of about 10 games, and won 2 games,drawn 3 draws and lost 5, you have a win % of about 20% and 9 points. However, if you are in charge of about 10 games, won 1, drawn 8 and lost 2, you only have a win% of 10%, but 11 points gained over the same amount of games.

Thats true, but you've picked a very unusual record in 1 win, 8 draws and 2 losses and it still only makes 2 points difference. Thats why its a decent approximation

Appleton does benefit more from ppg for Blackpool than he does win percentage, but in both cases his record is still poor. In a win percentage table, with Blackpool's record under Appleton extended to 26 games they'd be bottom, in a ppg table they'd be 18th.

[Salgado Is A Hero]

It was a hard job at Blackpool.....Holloway is over-rated.....people make a big deal of him getting a team promoted on a small budget whilst ignoring the fact Coyle did the very same and also how Holloway failed at other places.......Holloway got Blackpool promoted but McLeish got Birmingham relegated, that's how it works on here.

On a podcast I've just listened to, and posted on the podcast thread, a Blackpool fan mentions their players appalling fitness and how Holloway used to give them 3/4 days off a week....that, and the fact he doesn't even try to organise a defence, it's no wonder Appleton didn't do that well there.

First thing he did was sort out their defence.....right thing to do and what Eric Black did.....he didn't win any of his first 4 games either......then a home win against a team battling relegation when their manager was disappointed not to get anything from the game.

Results take time when you're taking over a struggling team, like Appleton was at Blackpool.......12 games isn't really enough to judge a manger......if it was, Hughes wouldn't have lasted long here.

His time at Portsmouth shouldn't be considered either....they could barely muster enough players on a Saturday.

[Pedro]

If Ferguson was in charge of Portsmouth last season his win % would be a lot lower than it is at Man Utd. Point being, a lot comes down to the league, team and circumstances you are in - it isn't always a reflection of your capabilities.

I'm going to judge Appleton on how the team fare under him, nothing else.

It was a hard job at Blackpool.....Holloway is over-rated.....people make a big deal of him getting a team promoted on a small budget whilst ignoring the fact Coyle did the very same and also how Holloway failed at other places.......Holloway got Blackpool promoted but McLeish got Birmingham relegated, that's how it works on here.

Superb point. Before Berg was appointed I was championing McLeish and McCarthy...but as you say, people only remember the relegations and not so much the promotions.

[Rover_Shaun]

At least Holloway is rated.

[CapeTownRover]

Not true Hypnotoad in terms of losing %. In theory, if you have a low losing %, then you either end up winning or drawing more than what you are losing. If you had lost say only 3 games, won 5 and drawn 20 games, would you take that lose % over that 28 games? I know drawing 20 games is a bit far fetched, but once you start to ensure that your % losses are kept down, then you stand a much better chance of winning and drawing more games. If you take Appletons Blackpool record, its 2 wins and 2 defeats, with 7 draws in his 11 games in charge. That's a win% of 18% with 13 points gained. Berg in his 10 games had 1 win, 3 draws and 6 defeats with a 10% win statistic, and 6 points gained. Would you take the win% or the points over that amount of games? It certainly is an interesting one to look at though

[Steve Kean's Hypnotoad]

Well, win percentage will get you closer to ppg than loss percentage. Loss percentage is too vague to be used as a standard indication. For example if you have a loss percentage of 10% over a season, you could have 41-124 points. If you have a win percentage of 10%, you could have 14-55 points.

Appleton's case is a good example, awful on win percentage, excellent on loss percentage, poor on ppg. Win percentage closer indication of ppg than loss percentage.

[grizfoot]

I think I understand your annoyance in the usage of win percentage instead of point per game because the latter is obviously more accurate. I do find statistics interesting but they never tell the true story, some things just aren't quantifiable. That being said I think the win percentage is preferred because (as already mentioned) its easier for people to comprehend a win percentage for example with one in every five game being around the 20% mark and one win in four around the 25% and so forth. Whereas the point per game figure doesn't really tell you anything unless you multiply by a set number of games to see how many points a season that would be. Even then the previous stats are based on how someone's performed previously and doesn't guarantee anything for the future. All in all it's probably total useless in many respects other than giving a vague general idea of how well a manager has performed in the past. It's main function is probably to make discussion of football seem more intelligent when in truth the stat is useless unless the context of which the said win percentage has been achieved in has been discussed and even then the sample size could be so small that it is totally unreliable anyway.

My personal favourite are those with and without such a player comparisons which makes it out that because a certain player is playing there's a sudden surge in the win percentage. The sample is always over a handful of games and is totally unreliable from a statistical point of view, that's not to mention the fact that opposition quality varies and said player might have been missing for the tougher games and available for the easier ones.

[DavidMailsTightPerm]

All statistics have to be viewed in context - as Rovers fans we have suffered from this where fans think we were wrong to try to get rid of Kean when we were near the top of the table - of course the context was that the performances were rubbish and we couldn't have maintained the run. For Appleton - at Portsmouth he was at a club in even more turmoil than Rovers - at Blackpool they were on a poor losing streak after a good start to the season. Any statistics - be it win ratio or points per game have to consider this - having said all that statistics for managers are largely meaningless - look at Holloway at Blackpool, I doubt his previous statistics would have foretold his later success - similarly I bet Ince had a fairly good win ratio prior to Rovers !

[bazza]

I think we will see Appleton's statistics improve here at Ewood because he has better quality players than at Pompey and Pool.

Let's give him a chance.

[BuckyRover]

Bolton win 26% (PPG 0.95)

Villa win 18% (PPG 1)

Which is more indicative of success/failure over a season?

Looking at relegations of the past. It actually happens fairly often.

I think when looking at win / loss %, win % is more relevant at top of table and loss % at the bottom.