Plus-minus measures the impact of a player on their team’s performance. Originally invented by hockey general managers, every player on the ice is awarded a plus when their team scores a goal while every opponent player on ice gets a minus. The higher the plus-minus rating, higher the net positive of goals scored for a player’s team. In terms of plus-minus, the beste player has the highest plus-minus score, and the worst player has the lowest. Plus-minus has also been modified for use in basketball, first by 82games, and now more famously by ESPNRead More
We updated our xGoals model a few weeks ago, as well as our process for continuously updating it throughout the season. Naturally, we’ve done the same for the xPassing model, which estimates the probability of any given pass being completed based on a number of details about the pass. You can read more about the original model here, but here’s the summary of the new model:Read More
During a recent American Soccer Analysis shareholders meeting in the penthouse suite of the swanky hotel we built in Minecraft (it’s our Slack channel), we discussed our favorite ASA articles of the past year. Because it is the season of listicles and we relish every chance to talk about ourselves, we decided to put them all together in one official post. Also, our site traffic is essentially zero at this time of year, so it seemed like an easy way for us to remember where we put them.
It was a great year for MLS (though perhaps not American soccer overall) and the most successful in our five year life as a website. We added interactive tables, introduced xPG, rebooted the podcast (new episode coming out soon! …probably), and added a lot of great new writers to our existing ranks of stale old writers. They’re not all represented in the list below, but special shout out to our weekly contributors who put together content every week - Little Things (@harrisonhamm21), Lowered Expectations (@harrison_crow), Expected Narratives (@ahandleforian), and Setting the Table (@ericwsoccer) - showed us the individual plays each week that made up the whole of the MLS season. We’d also like to extend a special thank you to Neil Greenberg of the Washington Post, for including us as a part of the WaPo’s incredible World Cup coverage.Read More
Much has been written and studied about set pieces in soccer. Penalty kicks have been Bayesed multiple times, I’ve analyzed free kicks in MLS and at the World Cup, corner kicks have been rigorously studied. But what about the humble throw-in? Aside from when teams develop a long throw-in program (see Delap, Rory) they are largely ignored or even ridiculed, in the case of Liverpool hiring a throw-in coach (see the first comment here).
In the offseason we upgraded our passing model, and its outputs are now featured in our xPassing tables (both interactive and static). After a few minor tweaks this week, now is as good a time as any to explain how it works.
Much like our Expected Goals (xG) models, the purpose of this model is to estimate the probability of success. Only, in this case, a success is a pass that is completed rather than a shot that is scored. For example, if Player A is passing the towards Player B, we can assign a likelihood of that pass being completed. We do this based on a variety of factors, such as the circumstances and player postion on the field, pass type, and the direction of the pass. And in this case, we opted to use a gradient-boosted ensemble of decision trees (GBM) rather than a logistic regression model (GLM).Read More
To anyone who's watched soccer, it's obvious that all passes are not created equal. Some are routine. Some are exceptional. The usual simple statistic that divides the completed ones by the attempted ones is missing quite a lot of context. Last year, to help solve that problem, ASA debuted a passing efficiency model designed to take into account the difficulty of the pass, similar to how expected goals is developed. Over 300,000 passes from 2015 were used to build three different models, and this year those models were calibrated to match 2017 performance. Ted Knutson over at Statsbomb just revealed a similar model build on 20,000,000 passes from Opta's dataset, which calls into question whether or not our 300,000 sample size is sufficient, but alas, all the MLS passes in the history of MLS wouldn't reach a third of that larger sample, so here we are.
This year we've broken out the model by individual player, which makes things pretty interesting because you can see how different players take different levels of risk depending on which part of the field they are on. For example, Philadelphia Union right back Keegan Rosenberry has an expected pass completion percentage of 57.9 percent in his own defensive third. His main competitor Ray Gaddis has 67.7 percent in the same area. They both have actual completion percentages near their expected level. Gaddis makes higher percentage passes when controlling the ball in a defensive position. That may not tell you which player is more effective but it does indicate that Rosenberry is more likely to send the ball up field, while Ray is going to look for a closer teammate.
Here's a link to the table with the latest results, but it also has it's own tab on our menu, titled "Player xPassing". Thanks to the work of Kevin Minkus (@KevinMinkus) and Drew Olsen (@DrewJOlsen) these stats will be updated regularly, along with all our other statistics.Read More