PDA

View Full Version : Player performance statistics



Aim Here
15-04-2018, 12:40 AM
So, today I decided to work out how 'effective' the players were by seeing how often league goals were scored or conceded while they happened to be on the pitch.

I scraped the data from the SPFL website, did a little bit of scripting and spreadsheeting and I've put the results below. I think they might be interesting.

The main statistic I've used is goal-difference per minute - the rate at which goals are scored minus that at which goals are conceded while the player is on the pitch. I figure that's a crude indicator of how effective a player is. Note that you can't just take the best 11 players here and make them a team (well you probably can't, anyways). Teams are, well, teams and not just a collection of individual players, and the performance of a player might rely on a bunch of other factors (perhaps players only function well as part of a partnership with one or more specific players, for instance). Still, might be a fun thing to look at. Also, I'm quite mistake-prone, so bear in mind it's not gospel. If you think you've spotted a mistake, you're probably right.

Some things to note:
The effectiveness depends on how much a player played since the start of the calendar year compared to last year. Our most effective players are ones we signed in January, and the ones we ditched tend to be at the bottom half. It's a clue as to how we've improved since then. Rherras is an exception here, but he has a low sample size (only on for 30 minutes against Motherwell, during which time they scored).

Stokesy is the least effective player with significant amounts of game time. This does surprise me, since his on-the-pitch flaws were less evident than, say, Simon Murray.
Barker is also surprisingly low in the table, given how useful he looks when charging upfield with the ball, but he hasn't played since hobbling off in the Rangers game, so it might just be due to mostly playing last year. It would be interesting to see what happens when he's on the pitch with Boyle and one or more of Allan, Kamberi and Maclaren.

Efe and Paul Hanlon are slap-bang in the middle of the table because they've played in almost all the league games so they almost mathematically have to be average. John McGinn is, surprisingly, a little *below* average, with similar amounts of game time. Go figure.

We can't say much about the keepers because only Rocky has significant amounts of game time. If I was looking at them, I'd not bother about the goal scoring rate or the goal difference rate, just the goals conceded.




Player


Time on pitch (mins)
Goals scored
Goals conceded
Goals scored/minute
Goals conceded/minute
Goal difference/minute


Jamie Maclaren
574
13
5
0.022648
0.008711
0.013937


Florian Kamberi
810
17
7
0.020988
0.008642
0.012346


Scott Allan
652
13
6
0.019939
0.009202
0.010736


Oliver Shaw
525
9
5
0.017143
0.009524
0.007619


Ryan Porteous
275
7
5
0.025455
0.018182
0.007273


Daniel Swanson
291
5
3
0.017182
0.010309
0.006873


Cameron Bell
164
2
1
0.012195
0.006098
0.006098


Martin Boyle
2616
42
30
0.016055
0.011468
0.004587


Lewis Stevenson
2700
43
33
0.015926
0.012222
0.003704


Dylan McGeouch
2483
40
31
0.016110
0.012485
0.003625


Darren McGregor
1569
23
19
0.014659
0.012110
0.002549


Paul Hanlon
2790
42
35
0.015054
0.012545
0.002509


Efetobore Ambrose Emuobo
2906
44
37
0.015141
0.012732
0.002409


Ofir Marciano
2610
41
35
0.015709
0.013410
0.002299


David Gray
471
7
6
0.014862
0.012739
0.002123


John McGinn
2688
40
36
0.014881
0.013393
0.001488


Steven Whittaker
1353
23
21
0.016999
0.015521
0.001478


Marvin Bartley
1756
23
21
0.013098
0.011959
0.001139


Deivydas Matulevicius
72
3
3
0.041667
0.041667
0.000000


Simon Murray
1462
21
21
0.014364
0.014364
0.000000


Ross Laidlaw

270
3
3
0.011111
0.011111
0.000000


Fraser Murray
19
0
0
0.000000
0.000000
0.000000


Brandon Barker
1315
18
19
0.013688
0.014449
−0.000760


Vykintas Slivka
825
10
12
0.012121
0.014545
−0.002424


Anthony Stokes
1447
18
24
0.012440
0.016586
−0.004147



Faycal Rherras
31
0
1
0.000000
0.032258
−0.032258

shetlandhibee
15-04-2018, 01:28 AM
very well worked out, but you need to be a krypton factor finalist to work it all out in inside 5 mins :wink:

houstonhibbee
15-04-2018, 03:10 AM
So, today I decided to work out how 'effective' the players were by seeing how often league goals were scored or conceded while they happened to be on the pitch.

I scraped the data from the SPFL website, did a little bit of scripting and spreadsheeting and I've put the results below. I think they might be interesting.

The main statistic I've used is goal-difference per minute - the rate at which goals are scored minus that at which goals are conceded while the player is on the pitch. I figure that's a crude indicator of how effective a player is. Note that you can't just take the best 11 players here and make them a team (well you probably can't, anyways). Teams are, well, teams and not just a collection of individual players, and the performance of a player might rely on a bunch of other factors (perhaps players only function well as part of a partnership with one or more specific players, for instance). Still, might be a fun thing to look at. Also, I'm quite mistake-prone, so bear in mind it's not gospel. If you think you've spotted a mistake, you're probably right.

Some things to note:
The effectiveness depends on how much a player played since the start of the calendar year compared to last year. Our most effective players are ones we signed in January, and the ones we ditched tend to be at the bottom half. It's a clue as to how we've improved since then. Rherras is an exception here, but he has a low sample size (only on for 30 minutes against Motherwell, during which time they scored).

Stokesy is the least effective player with significant amounts of game time. This does surprise me, since his on-the-pitch flaws were less evident than, say, Simon Murray.
Barker is also surprisingly low in the table, given how useful he looks when charging upfield with the ball, but he hasn't played since hobbling off in the Rangers game, so it might just be due to mostly playing last year. It would be interesting to see what happens when he's on the pitch with Boyle and one or more of Allan, Kamberi and Maclaren.

Efe and Paul Hanlon are slap-bang in the middle of the table because they've played in almost all the league games so they almost mathematically have to be average. John McGinn is, surprisingly, a little *below* average, with similar amounts of game time. Go figure.

We can't say much about the keepers because only Rocky has significant amounts of game time. If I was looking at them, I'd not bother about the goal scoring rate or the goal difference rate, just the goals conceded.




Player

Time on pitch (mins)
Goals scored
Goals conceded
Goals scored/minute
Goals conceded/minute
Goal difference/minute


Jamie Maclaren
574
13
5
0.022648
0.008711
0.013937


Florian Kamberi
810
17
7
0.020988
0.008642
0.012346


Scott Allan
652
13
6
0.019939
0.009202
0.010736


Oliver Shaw
525
9
5
0.017143
0.009524
0.007619


Ryan Porteous
275
7
5
0.025455
0.018182
0.007273


Daniel Swanson
291
5
3
0.017182
0.010309
0.006873


Cameron Bell
164
2
1
0.012195
0.006098
0.006098


Martin Boyle
2616
42
30
0.016055
0.011468
0.004587


Lewis Stevenson
2700
43
33
0.015926
0.012222
0.003704


Dylan McGeouch
2483
40
31
0.016110
0.012485
0.003625


Darren McGregor
1569
23
19
0.014659
0.012110
0.002549


Paul Hanlon
2790
42
35
0.015054
0.012545
0.002509


Efetobore Ambrose Emuobo
2906
44
37
0.015141
0.012732
0.002409


Ofir Marciano
2610
41
35
0.015709
0.013410
0.002299


David Gray
471
7
6
0.014862
0.012739
0.002123


John McGinn
2688
40
36
0.014881
0.013393
0.001488


Steven Whittaker
1353
23
21
0.016999
0.015521
0.001478


Marvin Bartley
1756
23
21
0.013098
0.011959
0.001139


Deivydas Matulevicius
72
3
3
0.041667
0.041667
0.000000


Simon Murray
1462
21
21
0.014364
0.014364
0.000000


Ross Laidlaw
270
3
3
0.011111
0.011111
0.000000


Fraser Murray
19
0
0
0.000000
0.000000
0.000000


Brandon Barker
1315
18
19
0.013688
0.014449
−0.000760


Vykintas Slivka
825
10
12
0.012121
0.014545
−0.002424


Anthony Stokes
1447
18
24
0.012440
0.016586
−0.004147


Faycal Rherras
31
0
1
0.000000
0.032258
−0.032258



so how much is SJM worth now?

judas
15-04-2018, 07:12 AM
Nice work and interesting too.

danhibees1875
15-04-2018, 07:39 AM
So to beat Celtic, we just need to start with the following 11 and let statistics take care of the rest:

Bell
Boyle McGregor Porteous Stevenson
Allan Swanson Mcgeouch
Kamberi Shaw McLaren

Good stats :aok:

FilipinoHibs
15-04-2018, 08:41 AM
So to beat Celtic, we just need to start with the following 11 and let statistics take care of the rest:

Bell
Boyle McGregor Porteous Stevenson
Allan Swanson Mcgeouch
Kamberi Shaw McLaren

Good stats :aok:

So Rocky & Effy are mares?

hibsbollah
15-04-2018, 08:46 AM
Interesting. Although those scores don't factor in important intangibles like chemistry, character, intelligence and teamwork. You can learn a lot from data.

Don't start your sentences with 'so'and all will be well:greengrin

danhibees1875
15-04-2018, 08:56 AM
So Rocky & Effy are mares?

:agree: they spend too much time horsing about.

iwasthere1972
15-04-2018, 09:21 AM
So Rocky & Effy are mares?

Obviously good in the Derby.

iwasthere1972
15-04-2018, 09:24 AM
How long did that take to compile?

FilipinoHibs
15-04-2018, 09:25 AM
Obviously good in the Derby.

Maybe do the analysis against each team to decide on the starting line-up.

Diclonius
15-04-2018, 09:37 AM
You also have to take into account that none of the players we signed in January have lost a game yet.

Tyler Durden
15-04-2018, 09:59 AM
It’s a good effort there but they aren’t actually player performance stats. They don’t really tell us much.

There are some good accounts on Twitter with good analysis of key passing which show McGinn and Allan’s value. “The Backpass Rule” and funnily enough “The Rangers Report” are worth a look. Sure they did some analysis on Shaw and Boyle which showed they’re above the curve of majority of their counterparts in the league.

Aim Here
15-04-2018, 11:13 AM
How long did that take to compile?

Maybe an hour, once I had the notion, but I'd already written scripts to get the data from the spfl website into a relatively easy to use format, so it was just writing a program to extract the data for Hibs games from there.

Aim Here
15-04-2018, 11:14 AM
It’s a good effort there but they aren’t actually player performance stats. They don’t really tell us much.

There are some good accounts on Twitter with good analysis of key passing which show McGinn and Allan’s value. “The Backpass Rule” and funnily enough “The Rangers Report” are worth a look. Sure they did some analysis on Shaw and Boyle which showed they’re above the curve of majority of their counterparts in the league.

The trouble with that is that the data for passes and whatnot isn't publicly available. I just use whatever is thrown up for free on the internet.