• strict warning: Non-static method view::load() should not be called statically in /usr/local/www/docs/championshipatbest.com/modules/views/views.module on line 879.
  • strict warning: Declaration of views_handler_filter::options_validate() should be compatible with views_handler::options_validate($form, &$form_state) in /usr/local/www/docs/championshipatbest.com/modules/views/handlers/views_handler_filter.inc on line 589.
  • strict warning: Declaration of views_handler_filter::options_submit() should be compatible with views_handler::options_submit($form, &$form_state) in /usr/local/www/docs/championshipatbest.com/modules/views/handlers/views_handler_filter.inc on line 589.
  • strict warning: Declaration of views_handler_filter_boolean_operator::value_validate() should be compatible with views_handler_filter::value_validate($form, &$form_state) in /usr/local/www/docs/championshipatbest.com/modules/views/handlers/views_handler_filter_boolean_operator.inc on line 149.
  • strict warning: Declaration of views_plugin_style_default::options() should be compatible with views_object::options() in /usr/local/www/docs/championshipatbest.com/modules/views/plugins/views_plugin_style_default.inc on line 25.
  • strict warning: Declaration of views_plugin_row::options_validate() should be compatible with views_plugin::options_validate(&$form, &$form_state) in /usr/local/www/docs/championshipatbest.com/modules/views/plugins/views_plugin_row.inc on line 135.
  • strict warning: Declaration of views_plugin_row::options_submit() should be compatible with views_plugin::options_submit(&$form, &$form_state) in /usr/local/www/docs/championshipatbest.com/modules/views/plugins/views_plugin_row.inc on line 135.
  • strict warning: Non-static method view::load() should not be called statically in /usr/local/www/docs/championshipatbest.com/modules/views/views.module on line 879.

Far from a real site launch

but I'll post a link anyway.

http://pldb.championshipatbest.com

It's very basic at this point, mostly because I have some database issues with the more advanced statistics and features that I've been struggling with. And having it online and publicly accessible will, I hope, inspire me to fix these issues. So, have at it, let me know what you want to see, what you would change, or if it would look better in a Windows 3.1 Hot Dog Stand color scheme. Drop me an email, or post a comment here!

Side note: if anybody has emailed me at the address posted on the sidebar, I apologize for not getting back to you - I hadn't been actively checking that account in awhile. I just set it to forward to my normal account.

Yep

So, I've been working on a new football stats database site (eventually and ideally, something similar to Fangraphs or the Sports Reference family) during much of my free time this summer. I think it's almost ready to go, but since I'm not really a programmer, it has been kind of a slow process. Sadly, the real innovation over here has been put on the back burner for awhile because of this. Anyway, I intend to continue with the Fulham-specific stuff here, though the stats pages will probably end up being moved over to the new site...but who knows. More when I have it all together!

Oh, and Kagiso is going to steal 100 bases. Book it.

Kagiso Dikgacoi won't go away

I don't talk a whole lot about transfer rumors here, but this one is pretty funny. The constantly-linked South African international is apparently trying to follow us to Australia. If you remember, Roy confirmed that we had him on trial last summer, and decided not to sign him. Now, its entirely possible that we may have told him something like, "we're willing to give you a another shot next year, good luck!" I'm sure that sort of thing happens a lot. But he (or more likely, his agent), hasn't stopped leaking reports that his signing with Fulham was imminent. Even after Roy stated publically that we're not interested.

You've got to give the guy credit for trying, even if it is getting a little creepy.

A couple of changes, and my computer's view on Premier League Awards

Just wrapping up my end-of-season analysis, and I thought I'd share some findings. First off, I made a few changes to my distribution of marginal goals between goalkeepers and defenders. The biggest issue I had with the old system is that it placed too many keepers below replacement level. I set replacement level at approximately 60 goals conceded on a league average number of shots faced. A Premier League club should be able to find someone who can do this either in their reserves or on a free transfer. The order from best-to-worst is the same, but now we only have one club (Hull City) with below replacement-level goalkeeping. That seems about right. The second big change to the formula involves possession, which in turn tells us about how pressured the keeper is. Take this example:

Shots Saves Team Possession Saves/Minute of Opp. Possession
Keeper 1 10 10 60% .278
Keeper 2 10 10 40% .185

Generally speaking, Keeper 1 is under more pressure. This is almost always offset by fewer shots seen, so it shouldn't hurt teams who do a good job of keeping possession. However, if your opponents manage lots of shots in relatively little possession time, the defence deserves less credit and the keeper deserves more.

Another significant change I've made is in assigning outfield defensive contribution to individuals. This was based entirely on defensive actions, I've modified this to incorporate possession as well. Its fairly basic at this point - succesful passes are good, any giveaways are bad. But it gets the point across. Defensive actions and possession are weighed differently from team-to-team - this also depends on their total possession time.

I've also taken the Marginal Goals concept a little further, and made it into a total player contribution rating. This is calculated by multiplying a player's marginal goals by his team's "marginal goals per win" rate. I counted draws as half of a win - it causes problems if you don't. Since marginal goals have different importances to different teams, this puts each club on a level footing, so you can better compare players on different teams. You can interpret these numbers as "A team of (player X) will win Y more games than some sacrificial no-hope auto-relegation club". I guess I need a fun-sounding acronym for this, along the lines of VORP or WAR. Until then, I'll just call it Total Footballer Contribution. Using this system, I've come up with some COMPLETELY OBJECTIVE (well, almost, but no yelling at me for not watching games!) picks for several awards.

Player of the Year

Very close. It really came down to two players - Fulham GK Mark Schwarzer (56.3 TFC) and Liverpool MF Steven Gerrard (44.7). If you adjust Gerrard's score for playing time, he would edge out Schwarzer with a 58.1. For this purpose, we're going to favor actual production over what might have been.
Winner: Mark Schwarzer

Young Player of the Year

Four good candidates this year. Man City's Stephen Ireland (22.9), Everton's Marouane Fellaini (19.0), and Arsenal players Samir Nasri (22.5) and Cesc Fabregas (21.5). Time adjusted, Fabregas is the clear winner, but a full TFC point is still a lot to assume. I didn't assume anything with Gerrard, and I'm not going to do so here.
Winner: Stephen Ireland

Defender of the Year

Here's our first big surprise candidate. Stoke City's Abdoulaye Faye (22.3) has been everything you would want from a defender - good stopper, careful on the ball, good distributor, and he's even scored a few goals. And look at that, another ex-Newcastle defender going on to do well elsewhere. Nemanja Vidic (27.8) and John Terry (21.2) are a little more obvious, but less interesting.

Fulham is well-represented on the list, beginning with Brede Hangeland (18.0) at #10 overall. Others include Aaron Hughes (14.1) at #25, Paul Konchesky (13.6) at #27, John Paintsil (13.1) at #31.
Winner: Nemanja Vidic

Midfielder of the Year

Since Gerrard was second in POTY, he wins this by default. Frank Lampard (43.7) and Cristiano Ronaldo (37.2) are round out the top 3. Highest-rated player from outside the top 4 clubs was Gareth Barry (30.2) at #5.

Danny Murphy (20.1) was rated 15th best midfielder, Clint Dempsey (14.6) was #33, and Simon Davies (10.4) made the list at #56.
Winner: Steven Gerrard

Forward of the Year

Top 3 are Nicolas Anelka (36.4), Dirk Kuyt (33.2) and Wayne Rooney (29.9). The TFC system says that Anelka's pure goalscoring beats Kuyt and Rooney's balanced attacks. Gabriel Agbonlahor (25.7) came in 6th, for the best outside the top 4.

Andrew Johnson (9.5) was Fulham's best striker, though Diomansy Kamara (4.9) and Erik Nevland (3.9) were impressive in limited time.
Winner: Nicolas Anelka

I'll go into more detail on Fulham's players later on, but I wanted to share a little bit about how they compare with the rest of the league.

Depending on what the definition of the word "playmaking" is...

When allocating credit for goals to individual players in the past, I've used a fairly simple rule: each goal is 50% shooting and 50% playmaking. Playmaking was initially determined by assists, but I later changed this to assist attempts to lower the impact of a teammate's poor finishing on the playmaker.

What about things like winning penalties, corners, and fouls? Or even intercepting a pass in a dangerous area? Those should count for something, but they don't show up as assists. To determine the effect of these, I looked at every goal scored this season, and counted up the last two events leading up to the goal. Here's the final tally:

Event Number %
Pass 884 66.57%
Fouled 99 7.45%
Ball recovery 68 5.12%
Standing tackle won 61 4.59%
Shot on target 42 3.16%
Corner won 38 2.86%
Shot off target 22 1.66%
Shot blocked 21 1.58%
Ground tackle won 19 1.43%
Penalty won 17 1.28%
Turnover 15 1.13%
Aerial duel won 13 0.98%
Interception 8 0.60%
Shot hit post 7 0.53%
Player subbed on 6 0.45%
Opponent tackle lost 6 0.45%
Goalkeeper pick up 2 0.15%

I think that the 50% playmaking element can be divided amongst these to give a more accurate picture of attacking contribution. We might take a few of these out - turnovers obviously don't help you score goals, and I think its safe to assume that the off-target shots are not related to the goals (instead, these are probably mistakes on the ensuing goal kick). As for the rest, I'm sure we can all picture goals in our head that had these events leading up. Aside from open-play passes, we can see that quite a few goals are scored directly or indirectly from free-kicks, recovering loose balls, tackles, and rebounding shots.

The thing that got me started on this was brainstorming how to incorporate possession into my defensive metrics. Passing the ball to a teammate isn't generally thought of as defending, but if the other team doesn't have possession, they can't score. Similarly, tackles and interceptions aren't directly related to scoring goals, but they can lead to counter-attacks, which in turn lead to goals. Once again, credit where credit is due.

Shooting tendenies of various Championship players

...And in some cases, current Premier League players. I mentioned last week the scouting potential we had from ESPN's gamecenter app. Although the data on lower and foreign leagues isn't nearly as accurate, it still shows us something. Using similar methods as I described last week, here are the Expected Goals and Shot Quality index for some of the newer Fulham players, and also some of the Championship's top prospects:

Player Seaon G EG SQ G/EG
Zoltan Gera 2006/07 3 5.71 0.894 0.525
2007/08 8 8.46 0.989 0.945
2008/09 2 3.14 1.321 0.637
Dickson Etuhu 2005/06 2 3.12 1.293 0.640
2006/07 6 5.91 1.048 1.015
2008/09 1 0.67 0.717 1.495
Julian Gray 2007/08 3 3.25 1.232 0.924
Collins John 2007/08 2 3.14 1.282 0.638
Andrew Johnson 2005/06 13 8.84 1.012 1.471
Giles Barnes 2005/06 1 1.26 0.564 0.794
2006/07 8 5.85 0.862 1.368
Andrew Surman 2005/06 2 1.70 0.950 1.177
2006/07 4 3.58 0.782 1.118
2007/08 1 2.70 0.587 0.370
Dexter Blackstock 2005/06 6 4.70 1.446 1.276
2006/07 12 11.84 1.385 1.014
2007/08 5 7.76 1.547 0.644
Michael Kightly 2006/07 8 7.28 0.871 1.099
2007/08 4 4.40 1.280 0.908
Michael Mifsud 2006/07 4 3.27 1.067 1.223
2007/08 10 7.98 0.954 1.253
Sylvan Ebanks-Blake 2006/07 7 8.16 1.273 0.858
2007/08 18 12.90 1.239 1.396
Ben Watson 2005/06 4 3.43 0.527 1.166
2006/07 1 0.96 0.471 1.038
2007/08 3 4.10 0.538 0.732
2008/09 2 1.66 0.931 1.206
Owen Garvan 2005/06 3 1.73 0.691 1.731
2006/07 1 2.42 0.606 0.413
2007/08 2 3.22 0.588 0.621
Joe Ledley 2005/06 3 2.49 1.076 1.205
2006/07 3 2.65 0.829 1.133
2007/08 9 6.00 1.227 1.499

Most interesting to me is Zoltan Gera. We've seen him appear out of nowhere in dangerous areas this season, but according to this, he didn't do as much of that in the past. Also, his goals/expected ratio is a little low, suggesting that his questionable finishing isn't necessarily new. Dickson Etuhu looks like the exact opposite, although his 08/09 sample isn't much yet. Cardiff's Joe Ledley looks like a consistent quality shooter. So does Ben Watson, but Wigan got to him first. Giles Barnes' 06/07 season looks very encouraging. Anyway, I'll dig deeper into this over the summer, I'm sure. As for right now, I'm hard at work trying out some new ways to improve my existing stats, such as determining which events are most likely to lead to goals. I feel like we've got team performances (which really ultimately stems from goal difference) pretty well defined, but accurately distributing that amongst individuals is forever a work in progress.

From the Other Side - Premier League Shot Quality Against

Yesterday, we looked at where the clubs were taking their shots from. Today, we have the same numbers from a defensive point of view. Specifically, this tells us how good a job defences are doing at preventing quality scoring chances. My favorite part about all of this is that it shows us who is making life easy (and difficult) for their goalkeeper.

ClubShot Quality Againstsort iconGoals AgainstMGP by GKMGP by OF
Chelsea0.840197.80335.318
Man Utd0.860204.48937.632
Arsenal0.894260.84834.273
West Ham0.924344.93922.182
Stoke City0.93147-0.57515.696
Sunderland0.940432.23517.886
Fulham0.9482818.46116.660
Portsmouth0.96548-3.11918.240
Liverpool0.976176.30035.821
Everton0.976329.86719.254
Hull City0.99851-7.68515.806
Aston Villa1.019422.14718.974
Middlesbrough1.025428.6797.442
West Brom1.03757-3.7567.877
Man City1.073403.13417.987
Bolton1.076478.1905.931
Tottenham1.1103513.63413.487
Newcastle1.117523.0558.066
Wigan1.150358.51218.609
Blackburn1.16552-7.37516.496

Once again, we're not counting own goals here. I've included the adjusted Marginal Goals Prevented totals for goalkeepers and outfield players. This was done by neutralising each team's save percentage, which is used to calculate MGP. Fulham are toward the top here, as we would have expected. Our defence takes some of the credit from Mark Schwarzer, but he is still the best keeper in the league. Blackburn makes Paul Robinson look considerably less horrible, but he's still near the bottom. Side note: I have some tentative plans to re-work the goalkeeper-outfield distribution for MGP, as the goalkeeper influence "feels" high. But the best-to-worst order should remain the same. Newcastle's keepers, Chris Kirkland, and Heurelho Gomes were the other big beneficiaries of this adjustment. Petr Čech, Thomas Sørensen, Edwin van der Sar, and Manuel Almunia weren't so lucky, with their contributions dropping by 3 or more goals each.

Scouting other leagues

I've been looking at a lot of shot data lately. One of the things I've wanted to do, eventually, is identify prospects in other leagues. After I posted the expected goals and shot quality numbers for several Fulham players, I started wondering what those numbers looked like for other players.

Player G EG SQ
Ronaldo 15 9.220 0.719
Lampard 12 8.597 0.666
Gerrard 13 6.555 0.715
Ireland 8 4.492 0.913
Robinho 11 5.902 0.848
Barry 5 5.080 1.361
Rooney 9 8.337 1.013

Ronaldo, Lampard, & Gerrard (and to a lesser extent, Ireland and Robinho) fit into that "sniper" category I mentioned before. At the start of the season, I noted Gareth Barry's impressive shooting percentages - this led me to believe that he's very good at getting into dangerous areas, much like our own Zoltan Gera. The high SQ seems to confirm this - he's our anti-sniper, and his goal total tells us that he can deal with Premier League defenders. I included Rooney because I considered him to be good at both.

My theory is that the snipers are players who can more easily adjust to other leagues. If a player scored a lot of goals in the Champioinship from dangerous areas like Gareth Barry, he might have more difficulty doing so against the more talented defenders of the Premier League. Players who can score easily from further away won't have to deal with those defenders as much, making it an easier transition. How do we identify those players? Thankfully, ESPN has some very nice shot charts on their Gamecast app, which covers English leagues all the way down to the Blue Square Premier League, as well the first divisions of Spain, Italy, Germany, France, Scotland, Argentina, Mexico & the USA.

As easily as I get sidetracked, this is probably more of a transfer window thing. It might be useful in identifying some potentially good scorers.

2008/09 Premier League Shot Quality

Since the database work was a little easier for this, I did this from an attacking perspective first. The following table shows each team's relative shot quality, which can be thought of as a modifier for shot percentages; goals scored (excluding opponents' own goals); and just for fun, I invented another stat called SNIPE. Ice hockey players who are good at shooting from a distance are called snipers. Football has snipers too, players who can score from just about anywhere. These players are extremely valuable, since they don't need to charge right into the teeth of the defence in order to put the ball in the net. This number is simply the team's goals scores divided by the shot quality modifier, and it passes the Joe Morgan laugh test. It's really more about style of play anything, but higher is generally better if you're into that sort of thing.

TeamShot Qualitysort iconGoalsSNIPE
Everton1.1804638.978
Stoke City1.1793025.438
Aston Villa1.1104338.739
Newcastle1.1023632.671
Bolton1.0943834.729
Arsenal1.0615148.079
West Ham1.0423634.534
Blackburn1.0393533.700
Fulham1.0173332.455
Wigan1.0013029.977
Portsmouth0.9973333.115
Liverpool0.9955656.262
Man Utd0.9915454.509
Middlesbrough0.9442425.436
Hull City0.9213335.818
Sunderland0.9112830.719
Chelsea0.9065358.478
Man City0.8864550.787
Tottenham0.8853640.657
West Brom0.8792730.714

If you're interested, shot quality is calculated by separating each team's shot totals into zones, and multiplying those by the league-average strike rate for that zone. This gives us a number of expected goals for that team. Expected goals is then adjusted for number of shots by dividing by the league average team shot total, and then multiplied by the team's own shot total. Finally, that number is divided by the league average number of goals scored per team. Since the number is adjusted for shots taken, we can also apply it to individuals.

To be honest, I was expecting Fulham to be a lot lower on this list. We are getting into dangerous areas, but as Rich points out here, these occurrences more often than not see one striker up against an entire defence. You have to respect Andy Johnson and Bobby Zamora for trying to make things happen with little support, even if the goals aren't always there. At the same time, Zamora in particular really should have more goals. I calculated the expected goals and quality of shots for a few players:

PlayerGoalsExpected Goalssort iconShot Quality
Zamora28.01.243
Johnson77.31.344
Dempsey75.40.999
Davies23.50.908
Gera23.11.321
Hangeland12.72.113
Murphy42.10.696
Bullard21.80.419

I'll admit that I'm very surprised to see Zamora's SQ so high, but it makes sense considering the nature in which he's used. By now though, he should have scored a few more goals by accident. Johnson is actually right around where he should be, though he's generally done better with his shots in the past. Despite what the anti-Zamora crowd says Brede Hangeland wins the award for worst finisher. But that's not what he's getting paid for, so we'll give him a mulligan...or 2. Gera gets into dangerous areas, which we all knew, but hasn't scored as much as he probably should, which we also knew. Bullard's extreme low SQ is no surprise, but Danny Murphy and the unlucky Simon Davies also have that tendency. Finally, don't let anybody tell you that Clint Dempsey is wasteful.

Next up, and what I originally went through all of this trouble for, is defensive shot quality.

Passing in the box

Some more Zamora debate on TIFF lead me to dig up an interesting bit of info:

League-wide passing % in the 18-yard box: 43.6
Bobby Zamora's passing % in the 18-yard box: 56.2

That's part of what we mean when we say "defending from the front", right? You may not want your forwards playing near your own goal, but holding possession in dangerous areas does have a ton of value.

I was curious, so I pulled up the league leaders, minimum 30 attempts:

Player Team Passes Accurate Passes %
Adebayor Arsenal 44 28 0.636
Mido Wigan 53 30 0.566
Zamora Fulham 73 41 0.562
Ireland Man City 36 20 0.556
Keane Tottenham 32 17 0.531
Rooney Man Utd 40 21 0.525
Johnson Fulham 40 20 0.500
Santa Cruz Blackburn 37 18 0.486
Anelka Chelsea 44 21 0.477
Cole West Ham 42 20 0.476

Not bad company. Like many others have said, he does everything but score. If he scored a few more goals, he'd be a £20m striker.