Contact us for a free demo Contact us
for a free demo
StatsBombIQ StatsBomb Data

Introducing MiPMoP (Again)

By Mike Goodman | October 21, 2013 | Analytics


MiPMoP stands for Minutes in Possession/Minutes out of Possession, and it’s something I started keeping track of last year to help contextualize stats by adjusting them for how long a team was in possession. Basically it’s a way to look at how a team’s possession impacts their stats. Generally I use it to look at (and refer to) shooting numbers, but there’s no reason it couldn’t be applied to other stats as well.

Let’s take a quick example. Teams A and B both shoot the ball 10 times per game. Without using MiPMoP, they both average 9 minutes per shot. Now let’s say that Team A averages 50% possession while team B averages 60%.* Team A’s MiP is 4.5, while Team B’s is 5.4 (for another more completely drawn example you can check out my old site). Clearly the two teams play differently, looking at their shooting statistics through the lens of possession helps show that.

The below table is the MiPMoP for the Premier League for the first seven games (MoP is just like MiP but based on shots conceded during time without the ball). To clear up (or cause) confusion, I’ve left the units as decimals of minutes, rather than converting to seconds. We’re dealing with fine lines here, and rounding to the nearest second seemed less precise. All the data is compiled from Opta powered

Man City58.83.404.26
Man United573.703.56
Aston Villa41.93.073.94
West Brom44.13.613.96
West Ham44.33.363.77
C Palace43.33.953.50

A Random Collection of MiPMoP Thoughts

Bad Teams
I’m generally pretty hesitant to make any blanket statements about statistics, but here’s one. If a team has under 50% possession, and higher MiP than MoP you are looking at a bad team. Basically, the team doesn’t have the ball, and they shoot more slowly in possession than their opponent, they are in serious trouble. Those stats may lineup for fewer teams than you think.

This year so far only the three newly promoted teams, Fulham, and Norwich fit the bill. That seems to pick up all the major relegation battle candidates, with the exception of Sunderland, who have a surprisingly healthy looking slash line of 43/2.91/3.78. Sunderland are dire, but the reasons for it, apparently, don’t have to do with the frequency of shots taken and conceded.

It’s difficult to describe just how awful Fulham are. No single other team in the top four leagues has managed to shoot only half as often in possession as their opponents, and that’s compounded by a pretty dire possession level. Clocking in at 45.7/4.96/2.48 makes them pretty much the laughingstock of Europe’s elite leagues. In the Premier League they’ve managed to be both the slowest to shoot in possession and the fastest to concede out of possession. Conceptually you might expect that from a team that dominated the ball, and basically played defense by playing keep away. It’s pretty difficult to claim that’s your plan though with only 45.7 percent possession.

Fulham are in big big trouble.

Reverse Splits

While shooting less frequently in possession than your opponents is unambiguously bad when a team has low possession, the situation is a lot less cut and dried when a team dominates the ball. Last year both Barcelona and Manchester United’s stats looked that way, but none of the other Champions League qualifiers across the top four leagues did. And in the Premier League, only Swansea and Wigan joined United in the reverse splits category.

This year, at least so far, things are different.

In addition to United, Swansea and Everton (where Wigan manager Roberto Martinez now resides), Arsenal, Liverpool, Southampton and Stoke all have reverse splits. It’s a trend that may not mean anything, but it’s certainly one worth keeping an eye as the season progresses, especially in light of the lower goal scoring numbers rolling in. It’s not all that hard to imagine a world where cagier managers focus on killing off games with a lead, skewing MiPs higher as they defend with the ball.

Liverpool’s line of 51.7/3.58/2.95 makes sense in light of what their early season has looked like, especially before the return of Luis Suarez. They just spent so much time absorbing pressure, without counterattacking that their line almost has to be reversed, and those numbers have certainly been skewed higher given the absurd amount of time they’ve spent winning by a single goal. With the return of Suarez, however, I’d expect those numbers to start changing very quickly, as Suarez is one of the most prolific shooters and creators in the game when up a goal. I’d expect a much lower MiP than MoP for them going forward.

Those Portuguese Managers
A lot has been made about how different Andre Villas Boas is from his mentor Jose Mourinho. And while that may be true, they both excel in very similar ways when looked at through the MiPMoP looking glass.

AVB and Mourinho preside over two of the most dominant MiPMoP teams in Europe at the moment, both in attack and defense. The same was true last year (it just happens that Mourinho’s dominant side was Real Madrid). The very low MiP, very high MoP, majority position combination is an extremely rare one in Europe. Most good teams keep the ball, and are dominant at either shooting or preventing shots, but not both. Mourinho has clearly mastered it, and AVB seems to have as well, which puts them in a very small club (Jurgen Klopp and Antonio Conte also hang out there, but that’s about it).

Don’t draw too many. This is the first time I’ve tracked MiPMoP data in-season, and I expect the numbers to change a lot as the season progresses. They serve as more of a descriptor of how a team is playing stylistically than a predictor of how they will perform. But, it’s a particularly effective one-stop shop to get a basic picture of how teams look stylistically. If teams defend with the ball, it shows up here. If they emphasize counterattacking, it shows up here. Style is important, and it can be hard to reflect in numbers - this is a good way to start.


*A brief statistical note. The way Opta calculates possession is actually not based on time but based on the percentage of passes each team has played. So, using possession statistics to calculate how many minutes each team has had the ball is kind of a cheat. But, Opta has done a bunch of work showing that it’s a pretty darn good proxy for time, and I’m all for using the tools we have until improved ones come along. So, until we have a more accurate measurement of possession I’m using Opta’s, but it pays to be aware of its possible shortcomings.

Article by Mike Goodman