Before we begin, a little secret:
"Advanced stats" aren't all that advanced. They are mostly percentages. Grade 8 math stuff.
What they are better thought of as is "New ways to think about old ideas while looking to place a value on them," but that's a mouthful.
So, they get called advanced stats, or new stats, or AN ATTACK ON THE ROMANCE OF THE GAME depending on your perspective.
The truth is they aren't all that new either. For generations people have tried to find an edge in evaluating team and player performance. It's only recently, however, that massive amounts of data has become available to the average hack like me.
So, hack away I do. Others will likely put things more eloquently than I (*cough* Richard Whittall *cough*), or with a greater grasp of the whole math thing (*cough* I used to spend all class playing a gambling game called 'quarters' at the back of Grade 12 algebra *cough*), but that won't stop me from jumping in with both feet this year on the stats side of things.
In the new column Jock Math, CSN will track certain things all year. The focus will be on both the league overall and the Canadian teams in detail.
Please don't mistake Jock Math as an attempt to "prove" things. Stats don't prove anything. Rather, they illustrate things and ask questions about the game. If read with an open mind they can challenge deeply held beliefs, or, yes, reinforce things you instinctively know are true.
Ultimately stats should start a conversation and collectively help us understand the game on a deeper level.
So, let's begin.
[PRBREAK][/PRBREAK]
This being week one we must be careful not to read too much into the numbers we are looking at. The n is way too small just yet to even attempt to conclude. For this week, we are more or less introducing the core concepts we'll be looking at.
Let's start with a team number. For now, it's the only team number we'll look at because the season is too young to expand on it just yet.
However, the stat is a big one and one we'll spend a lot if time looking at this year. It's Total Shots Ratio (TSR).
If you're a hockey fan you might be familiar with the stat Corsi. This is soccer's equivalent. Basically it acknowledges that the core concept of any put-the-blank-in-the-net sport is to score goals. However, because goals are so rare it uses shot attempts as a substitute for goals and looks to see what teams are directing the greatest percentage of a game's shots.
Wait, wait, wait you might be saying. How can a shot be close to the same thing as a goal? I watched Andre Lombardo play at BMO and he shot a lot but couldn't score in that infamous brothel with the familiar handful of cash (I'd have to question where Andre got the cash since he topped out at $35,000 salary, but I digress).
Well you're right, but also very, very wrong. Taken in isolation any single shot isn't the same thing as a goal. However, taken in isolation any incident is irrelevant when analyzing stats. What you're looking for is trends that happen over a long period of time and then trying to determine whether those trends are predictive.
TSR is. It's not 100% predictive because no stat is (and if it were I wouldn't be telling you people about it, I'd be on a plane to Vegas). However, it's been tested over a long period of time and has demonstrated itself to be quite reliable it predicting a general league position.
Let's look at this season's Premiership numbers as an illustration.
1. City .645
2. Chelsea .627
3. Southampton .591
4. Spurs .573
5. Liverpool .567
6. Everton .560
7. Newcastle .558
8. Arsenal .548
9. United .536
10. Swansea .522
11. WBA .499
12. Villa .470
13. Palace .461
14. Norwich .455
15. Hull .444
16. Sunderland .420
17. Stoke .411
18. Cardiff .385
19. West Ham .383
20. Fulham .376
Like I said, pretty predictive. Generally if a team's TSR position is greatly above or below their actual position then you can reasonably predict a correction.
Reasonably predict does not equal "bet the kid's education fund" because sports doesn't work that way. But, it does suggest those teams might be experiencing more or less luck than others.
Bringing it back to MLS, we will be tracking this weekly. Last year, a midsession peak at the Impact suggested they weren't truly a playoff team and if Di Vaio dropped off in form the Impact would be in trouble. You're going to have to take my word for that because I wasn't publishing Jock Math then, but we won't have the same problem this year.
It has next to no predictive value just yet, but the week one TSR numbers for MLS (Chivas v Chicago stats are not yet available):
1. Galaxy .750
2. Vancouver .636
3. Portland .600
4. Seattle .583
5. Montreal .576
6. DC United .550
7. Houston .512
8. Revs .487
9. Columbus .450
10. Dallas .423
11. Kansas City .416
12. Philly .400
13. New York .363
14. RSL .250
Drawing any conclusions from this is foolhardy. One game is not a trend, but it is a base point. As said, this will be a weekly track.
We will also be tracking individual stats. In addition to highlighting one specific area of the game each week we will provide updates to the leaders in three specific categories: The unique CSN stats of Defensive Involvement (total clearances, interceptions, recoveries and tackles) and Key Passes plus Shots.
We will also track passing percentage.
Again, with the caveat it's way early, the week one numbers.
DI
1. Jose Goncalves - 22 (this will be expressed as a per 90 min number in future weeks)
2. Austin Berry - 21
3. Norberto Paparatto - 20
4. Bobby Boswell - 19
5. Aaron Maund - 18
5. Matt Hedges - 18
7. Michael Parkhurst - 17
7. Aurelien Collin - 17
7. Jermaine Taylor - 17
10. Diego Chara - 15
S+KP
1. Landon Donovan - 10
2. Robbie Keane - 8
3. Joao Plata - 7
3. Marcelo Sarves - 7
5. Lloyd Sam - 6
5. Dominic Dwyer - 6
7. 10+ tied at 5
Passing percentage (min 50 attempts)
1. Michael Parkhurst - 92.3
2. Perry Kitchen - 90.4
3. Tony Tchani - 89.8
4. Eric Alexander - 88.7
5. Will Trapp - 88.0
6. Ibrahim Sekagya - 87.5
7. Russell Teibert - 87.5
8. Federico Higuain - 86.9
9. Dax McCarty - 82.9
10. Osvaldo Alonso - 81.3