I’m pleased to announce that I’ve joined StatDNA as Vice President of Analytics and Software Development. This is a super exciting opportunity for me as I’ll be combining my loves of software development, data analysis and soccer. What could be better? I’ll hopefully have some blog posts up for StatDNA over at their blog soon using their best in breed data. I will continue to update this site as well although not as frequently. Thanks to Jaeson and the rest of the StatDNA team for giving me this opportunity!
Archive for Uncategorized
Does Recent Performance Affect Transfers?
The midseason transfer window is about to open for MLS but trades are fair game. Today the New York Red Bulls swapped attacking midfielder Dwayne DeRosario (acquired in another trade earlier this season) for DC United’s holding midfielder Dax McCarty. This is an interesting trade on many levels (salary cap hits, contract length, etc) but what struck me most was that NY was bringing in defensive reinforcements after a spell of poor defensive performances.
Looking at the moving average of NY’s goals against and DC’s goals for, you can see a trend of a stalling offense and a leaking defense. Are these legitimate problems or just slumps? For NY, they’ve certainly been suffering from the absence of Tim Ream and Rafa Marquez, but the problems started long before they had to report to their national teams for the Gold Cup. Regardless, there are two interesting factors in decision making that might be coming in to play here. The first is recency bias — the attribution of more weight to more recent data/information than older info. Basically, you remember the recent stuff better so you consider it more important and the opposite is true for older stuff. The fact that NY kept clean sheets in 5 of their first 7 matches is long forgotten, replaced by images of Greg Sutton gifting the ball to Roger Levesque for an embarrassingly easy goal. The other issue that might be coming in to play here is prospect theory. Depending on the team’s frame of reference, swapping players, and the expected gains can look attractive or unattractive. Making the trade early in the season might not have been attractive to NY, but now that they are feeling vulnerable defensively, it looks a lot more attractive. People are also less risk adverse if they are feeling poor and view the risk as having a large upside and relatively low downside since there isn’t much to lose. This is the sort of situation NY finds itself in now. Certainly there are other factors that came into the trade decision, but I think it’s important to understand how data could have played a part and some of the caveats that come with it. The data points to NY having defensive problems. Understanding how the mind makes decisions and some of the biases and pitfalls that go along with that can help lead you to making the best decision based on that data.
Oct 7 2001
26 days after September 11th, the US was scheduled to play a World Cup Qualifier against Jamaica in Foxboro, Massachusetts. Whether or not it was actually going to be played was in doubt but eventually the game was allowed to proceed under strict security. My cousin and I had already purchased tickets and were relieved to have an excuse to leave New York for the weekend and temporarily escape the madness. The plan was for me to meet him at his place in the Bronx and we would drive up together the day before the match and stay with my folks for the weekend.
The Danger of Statistics
Now that Opta and MLS have announced a partnership, fans will be able to access a wealth of statistics similar to the Guardian’s Chalkboards. Wonderful. I’m genuinely excited about this, but at the same time I’m bracing myself for the worst. My fear is that people will be duped into a false sense of knowledge.
How Generation Adidas Affects the Salary Cap in MLS
Previously we looked at the downside of drafting Generation Adidas players and a few questions arose out of that piece, particularly around the longterm benefit of a Generation Adidas contract and what things look like on average instead of just looking at a few anecdotal cases.
To answer those questions I wanted to look at players’ salaries and how they differ between Generation Adidas and regular players as well as their effect on the salary cap. I looked at draft picks 1-22 over the last 5 years. Beyond pick #22, the frequency of Generation Adidas contracts is almost zero which is why I stopped there.
The Downside to Drafting Generation Adidas Players
When the Sounders drafted Michael Tetteh with the 20th pick overall of the 2011 MLS SuperDraft, a lot of people were asking why he wasn’t drafted sooner, especially because he is a Generation Adidas player. Generation Adidas players don’t count against the salary cap until they graduate, so on the surface they appear to be great bargains, but their inflated salaries complicate things.
Generation Adidas players make approximately twice as much as their non-Generation Adidas counterparts drafted in the same position. In some cases, it’s much more extreme. Dilly Duka, with $213,000 in guaranteed compensation, makes 5 times as much as Zach Schilawski, with $42,500. Duka was drafted one spot ahead of Schilawski in 2010, and is Generation Adidas while Schilawski is not.
- Draftee Salaries for 2009-2010 broken down by Generation Adidas versus non-Generation Adidas
So what’s the problem? These guys aren’t hitting the cap, right? In fact, Schilawski costs his team $42,500 in cap space, while Duka costs $0. The problem arises when Duka graduates. All of a sudden, Duka will start hitting the cap at well over the market price. Duka hasn’t graduated yet, so this is all hypothetical, so instead, let’s look at an example from 2009 where there were consecutive Generation Adidas/non-Generation Adidas players drafted.
Steve Zakuani (GA) was the #1 pick in 2009 and Sam Cronin (non-GA) was the #2 pick. Both have enjoyed immediate success in MLS and have been capped several times by their national teams. Zakuani graduated from Generation Adidas after a year. Let’s look at how they affect the cap.
| Year | Steve Zakuani | Sam Cronin |
|---|---|---|
| 2009 | $0 | $84,000 |
| 2010 | $178,000 | $105,500 |
| 2011 est. | $178,000 | $105,500 |
The point isn’t that Cronin is a better pick than Zakuani or vice versa. The point is that when these guys graduate from Generation Adidas, they are going to hit the cap and hit it hard. In fact, Generation Adidas salaries are higher than the average salary of a starter. That’s pretty startling if you consider the high salaries of league veterans and designated players. When people say a Generation Adidas player doesn’t hit the cap, that’s a very short term view of the world. When you’re focused more on the long term picture (2+ years out), suddenly Generation Adidas players are no more attractive than their regular counterparts and possibly less attractive.
On Data Visualizations
I’ve been meaning to play around with Tableau for a while and I thought my work with games back would be a good start. Visually it isn’t that striking, but I think it does a decent job of conveying the information.
