Goals/90, Assists/90 and Points/90

[mpruitt]

Originally posted by ChrisE
Of course, another problem is that Taylor Twellman 2003 and Ante Razov 2003 (or Jason Kreis of 2000, or whoever) played radically different roles on their teams, another kind of systematic variation that there's simply no way to control for. So, while these numbers may be interesting, they are nothing like a hard and fast guide to whether A is a better forward (or playmaker) than B.


This is something that I tried to flesh out in my initial post in the thread I recently started. There has to be a way to adjust for that. I think that one way to do so would be to take a player's individual numbers and look at percentage wise how much better they were than their teamates. Because regardless of system the idea would be to see who's maximizing offensive output under any system. It would seem to me that it'd be a more important measurement than simply looking at the raw numbers because it would take into account what one's teamate's are doing. That's not to say that simply looking at output above the team average would mean that Player A would duplicate his offensive output for Team B, however isn't the whole point to find out who's most valuable to their team?

[ChrisE]

Originally posted by maxim-1
This is something that I tried to flesh out in my initial post in the thread I recently started. There has to be a way to adjust for that. I think that one way to do so would be to take a player's individual numbers and look at percentage wise how much better they were than their teamates. Because regardless of system the idea would be to see who's maximizing offensive output under any system. It would seem to me that it'd be a more important measurement than simply looking at the raw numbers because it would take into account what one's teamate's are doing. That's not to say that simply looking at output above the team average would mean that Player A would duplicate his offensive output for Team B,

however isn't the whole point to find out who's most valuable to their team?

I'd say no. The whole point is to find out who's the best, in my opinion.

If you do something like look at what percentage of a player's offense an individual accounts for, you're going to end up helping out guys like Mamadou Diallo and (reputedly) Jeff Cunningham, guys who monopolize the ball to the detriment of their team.

Furthermore, if you've got a guy who's the focal point of an offense, his goal numbers will already be inflated because the team is built around providing him the ball. If you then say he's better than his numbers because he scores a huge portion of his team's goals, you're doubly distorting things.

In the interest of fairness, this is the top 10 of players, using a very simple, un-subtle, non-PK goals/Team Goals statistic:


Ruiz, Carlos 0.477
Twellman, Taylor 0.408
Diallo, Mamadou 0.403
John, Stern 0.385
John, Stern 0.375
Lassiter, Roy 0.375
Moore, Joe-Max 0.368
Cunningham, Jeff 0.341
Pineda Chacon 0.339
Graziani, Ariel 0.333


Obviously Ruiz had a good year last year, but was it that good? I'd say no, that his numbers are that good (in my opinion, that's an amazing number to win by) simply because the Galaxy didn't have any other forwards, so they relied very heavily on him.

[mpruitt]

All of those are very fair points I think at very least it's interesting to look at these things in a who's most important to their team type thing. However, there has to be some sort of way to normalize this. If you could take the percentage of team goals and combine/compare it with their teamates' goals per shot per 90 then I think you might have one all out compelling statistic.

Sorry if I'm kind of jacking this thread by bringing this up. I don't know if you feel like it is or not, but maybe you want to respnd to that question specifically in the other thread i created?

[beineke]

Originally posted by ChrisE
Basically, when you normalize the goals/90 like I did, you're assuming that offense in MLS has remained static for the last 8 years, and defenses have been the only things that were improving.


... or we might not be assuming anything about the relative quality of offense and defense. IMO, what we've seen is mostly the evolution in strategy. Rather than getting forward to break down an opposing defense, teams now get numbers back and rely more on the counterattack. When Gansler won the 2000 championship, I think that really opened people's eyes.


Obviously, I would hope that ... offenses have been improving as well (so, in fact, Twellman, Buddle etc. deserve more credit than they get); unfortunately, I've got no idea how to separate these two, so I think it's pretty much impossible not to underestimate the more recent players' ability.

Big project, and one I've been thinking about for a while. From other, mature leagues, we can learn how a player's productivity typically evolves over time. Suppose that, on average, the productivity of 27-year-olds is 10% higher than the productivity of 28-year-olds. If 27-year-olds in MLS drop by 20% the next year, then that extra 10% can be attributed to the league getting better.

I've heard that someone actually did a careful job of this in baseball, but I've never seen the work.

[ChrisE]

From baseball, one of the better ways to decide this issue, I thought, was to use standard deviations as the measure. Take a mythical "average" rate and then checkt to see how many standard deviations the player's rate is from "average."

The problem here is that you need a number that's always between 0 and 1 to do this, and so we have to come up with an appropriate opportunity statistic with which to measure goals against. We could use minutes, though I'd probably prefer minutes X 2.

Anyway if n = opportunities, p = average goals per opportunity and q = 1-p, than the standard deviation would be:

(n*p*q)^(1/2)


Well, after almost a year, I think I finally figured out what voros was suggesting here, and so I did the calculations for the 2004 season. What I got was actually pretty boring:


Ching, Brian 0.52
Buddle, Edson 0.45
Johnson, Edward 0.43
Wolyniec, John 0.41
Eskandarian, Alecko 0.40
Noonan, Pat 0.39
Ralph, Damani 0.38
Wolff, Josh 0.36
Ruiz, Carlos 0.35
Cunningham, Jeff 0.34
Twellman, Taylor 0.32
Arnaud, Davy 0.32
Peguero, Jean Philippe 0.30
Gaven, Eddie 0.29
Glen, Cornell 0.28


None of these names are the least bit surprising, and it hardly seems like this is any better than looking at the Budweiser Scoring Leaders Chart. Maybe that's how it should be.

Nevertheless, I thought part of the problem might have been in making the scoring rate too low - because it was an average for all players, there were practically no forwards who were anything below average. So instead of the league average, I tried using the forwards' average, and got this:

Ching, Brian 0.29
Buddle, Edson 0.25
Eskandarian, Alecko 0.20
Herron, Andy 0.19
Glen, Cornell 0.18
Ruiz, Carlos 0.14
Wolyniec, John 0.12
Peguero, Jean Philippe 0.12
Ngwenya, Joseph 0.11
Gomez, Christian 0.10
Johnson, Edward 0.10
Taylor, Fabian 0.08
Cunningham, Jeff 0.06
Nhleko, Toni 0.06
Taylor, Matt 0.05


which I liked a little better. Raising the average also made it possible to pick out the worst forwards; they were:

Adu, Freddy -0.09
Delgado, Alberto -0.11
Kreis, Jason -0.17
Galvan Rey, Sergio -0.18
Moreno, Jaime -0.26
Spencer, John -0.31

Jaime obviously is badly hurt by the fact that I'm ignoring assists here; the others, not so much. I think we can conclude that Spencer, Galvan Rey, and Kreis were just plain bad this year.

[ChrisE]

I tried applying the above to career MLS stats, adjusting for both PK's and Year, using forwards' average goals/2*min, but the results were pretty disappointing. What we get is, again, essentially a list of top scorers. Voros - I wonder if the economic situation of MLS, where most of the most talented players leave after only a few seasons to play in Europe makes this approach far less useful in MLS than it is in MLB.

Anyways, below is the list; second column is standard deviations above expected goals scored, the third is minutes played:

Lassiter, Roy 1.09 13769
Kreis, Jason 1.00 20290
Razov, Ante 0.95 12482
Diaz Arce, Raul 0.95 12423
Cunningham, Jeff 0.80 11380
Serna, Diego 0.80 9668
Twellman, Taylor 0.77 6361
John, Stern 0.70 4463
Cerritos, Ronald 0.69 13595
Ruiz, Carlos 0.68 6294
Moreno, Jaime 0.66 15203
Savarese, Giovanni 0.66 8173
Washington, Dante 0.65 10205
Graziani, Ariel 0.65 7570
Diallo, Mamadou 0.59 5846

[AussieVamp2]

[QUOTE=ChrisE][/b]

I think that's pretty clear, yeah, it was just a more difficult criterion to use.


[/b]


so any idea what the percentage of the '90' minutes played by starters is, if looking at this?

[AussieVamp2]

Not hard at all (if anyone wants the file I'm using, or a full list, just ask), but I'm afraid it may bias things a little, and would exclude a lot of good candidates. Of the top 10 field players (880 total)in terms of minutes for all clubs in MLS history, only slightly more than half, 491, have played more than 2000 minutes. If you lower that bar to 1500 minutes, you include 787 players. Furthermore, you're going to be biasing this towards the early years simply because teams played more minutes previously. From 1996-1999, 267 players topped 2000 minutes, while from 2000-2003, only 224 did.


What about percentage of games (or 90 minutes played) in a particular season? (although that doesn't help with variability between different length seasons) but one possible criteria to pick players from different length seasons?

[Serie Zed]

Why? Because Carlos Ruiz takes LA's penalty kicks, and it would upset the BS faithful if he showed up higher on your lists?

A penalty kick counts just as much on the scoresheet as any other goal. If you are going to subtract penalty kicks, why not goals scored off free kicks, or tap-ins from 2 yards out with the keeper already beaten - surely that is an easier shot than a penalty kick.

Because I can find 12-year olds at the Elks club's annual free throw shooting context who shoot 24-25 every year.

Hint: they wouldn't do well in the NBA.

[ChrisE]

[QUOTE=ChrisE][/b]

I think that's pretty clear, yeah, it was just a more difficult criterion to use.


[/b]


so any idea what the percentage of the '90' minutes played by starters is, if looking at this?

I'm not sure what you're asking here. Is it "what percentage of minutes in any game are played by starters, and what by subs"? or is it "what percentage of minutes do a team's Best XI play?"

[AussieVamp2]

[QUOTE=AussieVamp2]

I'm not sure what you're asking here. Is it "what percentage of minutes in any game are played by starters, and what by subs"? or is it "what percentage of minutes do a team's Best XI play?"

Actually, I hadn't thought of it as different possibilities like that.

Guess I was thinking more along the lines of starters.

Might be different too for goalkeepers, compared to strikers etc., less likely to be subbed?

e.g. leaguewide, starting goalies play 96% of minutes, strikers 89% or whatever, that sort of thing?

Best XI (how you define, exactly?) also be interesting, too?

[ChrisE]

[QUOTE=ChrisE]

Actually, I hadn't thought of it as different possibilities like that.

Guess I was thinking more along the lines of starters.

Starters played about 95% of available minutes, with subs obviously seeing 5 (15257 out of 296394).


Might be different too for goalkeepers, compared to strikers etc., less likely to be subbed?

e.g. leaguewide, starting goalies play 96% of minutes, strikers 89% or whatever, that sort of thing?



Yeah. The below numbers are what percentage of total possible minutes starters for each position played. I'm sure that sometimes forwards were subbed for forwards, and sometimes defenders came in; I didn't bother with that part.

GK 99.4 26835 27000
D 96.3 95017.5 98685
M 93.8 107149.5 114255
F 92.3 52694 57060



Best XI (how you define, exactly?) also be interesting, too?

I'd probably just define it as the 11 players who played the most minutes.

[AussieVamp2]

[QUOTE=AussieVamp2]

Starters played about 95% of available minutes, with subs obviously seeing 5 (15257 out of 296394).



Yeah. The below numbers are what percentage of total possible minutes starters for each position played. I'm sure that sometimes forwards were subbed for forwards, and sometimes defenders came in; I didn't bother with that part.

GK 99.4 26835 27000
D 96.3 95017.5 98685
M 93.8 107149.5 114255
F 92.3 52694 57060




I'd probably just define it as the 11 players who played the most minutes.



Cool, very interesting, thanks! :)

'Best 11' be interesting again, after seeing that.

Goalkeepers come off only if dead here, it seems!

[ChrisE]

[QUOTE=ChrisE]



Cool, very interesting, thanks! :)

'Best 11' be interesting again, after seeing that.

Goalkeepers come off only if dead here, it seems!

I drew this up a few days ago, expecting to find some evidence that depth in MLS was increasing. I didn't. The first chart represents the sum minutes played by the top 1,5,10, or 16 players of each team, ranked by minutes, goalkeepers removed. The bottom chart is for individual players at each rank.

Sum
1 5 10 15
1996 9.4 43.3 75.4 92.9
1997 9.3 41.8 72.7 91.7
1998 9.3 42.3 74.7 93.0
1999 9.4 43.0 74.6 92.4
2000 9.1 42.1 73.8 92.6
2001 9.4 43.6 76.1 94.1
2002 9.5 44.2 75.8 93.9
2003 9.4 42.2 73.8 92.3
2004 9.6 43.2 75.9 94.0


Individual
1 5 10 15
1996 9.4 7.9 5.4 2.7
1997 9.3 7.5 5.3 2.7
1998 9.3 7.6 5.5 2.4
1999 9.4 7.8 5.3 2.6
2000 9.1 7.8 5.3 2.8
2001 9.4 8.0 5.2 2.7
2002 9.5 8.0 5.2 2.6
2003 9.4 7.7 5.5 2.5
2004 9.6 7.7 5.5 2.2

[AussieVamp2]

very interesting, thanks

so best 11 (included goaldude) playing close to 85% of the minutes?

have MLS team or squad sizes been the same over that period I guess is one question?


with soccer, even with more depth, does it necessarily mean that starters/best players will play less? or just that dropoff is a bit less when they do?

e.g. the +/- thing now, compared to a few years ago?

as in lose 0.02 goals with backups compared to 0.04 8 years ago?

[ChrisE]

very interesting, thanks

so best 11 (included goaldude) playing close to 85% of the minutes?

have MLS team or squad sizes been the same over that period I guess is one question?

No, actually since I excluded goalies from both numbers, it's (75+11)/(100+11)= 77.5% or so, I believe.

I think MLS squad sizes have risen marginally, but mostly in the # of young developmental players, who mostly don't make much impact.


with soccer, even with more depth, does it necessarily mean that starters/best players will play less? or just that dropoff is a bit less when they do?

e.g. the +/- thing now, compared to a few years ago?

as in lose 0.02 goals with backups compared to 0.04 8 years ago?

Good point. I didn't mean to say that I thought depth hadn't changed; I still think it's improved, just that this stuff didn't provide evidence for it.

[Nutmeg]

[quote=AussieVamp2]

I drew this up a few days ago, expecting to find some evidence that depth in MLS was increasing. I didn't. The first chart represents the sum minutes played by the top 1,5,10, or 16 players of each team, ranked by minutes, goalkeepers removed. The bottom chart is for individual players at each rank.

Sum
1 5 10 15
1996 9.4 43.3 75.4 92.9
1997 9.3 41.8 72.7 91.7
1998 9.3 42.3 74.7 93.0
1999 9.4 43.0 74.6 92.4
2000 9.1 42.1 73.8 92.6
2001 9.4 43.6 76.1 94.1
2002 9.5 44.2 75.8 93.9
2003 9.4 42.2 73.8 92.3
2004 9.6 43.2 75.9 94.0

Individual
1 5 10 15
1996 9.4 7.9 5.4 2.7
1997 9.3 7.5 5.3 2.7
1998 9.3 7.6 5.5 2.4
1999 9.4 7.8 5.3 2.6
2000 9.1 7.8 5.3 2.8
2001 9.4 8.0 5.2 2.7
2002 9.5 8.0 5.2 2.6
2003 9.4 7.7 5.5 2.5
2004 9.6 7.7 5.5 2.2



Yeah, I'm doing some old digging, but this seemed interesting to me.

One question - has anyone done any work to correlate these numbers with winning percentage?

Does anyone from the "old gang" even post here anymore?

[Oliver Anderson]

Hi guys,

This stuff is great and comparable to what I've been doing for the last four years in the UK. I've written two books on the subject of applying statisticial analysis to football (THE RED REVIEW and THE FOOTBALL REVIEW) but it is very slow going for getting other people involved. I have created and developed a number of different statistics since I started and one of the best is the goals/assists/points per-90 minute rates you are looking at here.

I also have a number of other stats to rate footballers in the Premier League. Please take a look at my site (hope the administrators don't mind - www.thefootballreview.co.uk (http://www.thefootballreview.co.uk)) and tell me what you think. I'd love to know what you guys think and whether you have anyother ideas for what could be done to better analyse football. You have by far the most analytical discussions on the web and I'd value your opinion.

Many Thanks

Oliver Anderson