Derby - Moneyball at it's finest

[DavidPengilley]

two games are just so fundamentaly different it will never work. why do you think stats in football are still basic while majority of other sports it is so developed? football is just too unpredictable, too much driven by random events, uncontrollable... that stats tell you very little.

Unfortunately this reliance on random events and luck is why the statistic-movement in football is still struggling to progress into the mainstream.

[mrw072]

Unfortunately this reliance on random events and luck is why the statistic-movement in football is still struggling to progress into the mainstream.

It's an interesting idea. My worry with stats in football is whether or not the strength of the opponents is taken into consideration when compiling the stats. Because I think that could have a huge bearing on things and if it isn't considered, may give misleading results. From reading the posts above I think the ideal sport for stats like those mentioned would be cricket. I could see it being a major influence there! If it isn't already. But I'm a bit sceptical about its use in football.

[s1111]

I think some of you may be surprised how advanced statistics have become at the top level of football and how much influence they do have on decisions already. All the top Premier League clubs employ large teams of statisticians and analysts who have a big impact on all aspects of the club. Arsenal even bought out a company called StatsDNA so they could use their data exclusively.

Certainly clubs going right down to the lower levels use statistics to draw up a shortlist of potential transfer and scouting targets, even if the final decisions are made through watching individuals. It's well known the transfer policies of clubs like Liverpool, Man City and Arsenal, as well as the obvious Brentford and Midtyjlland, are heavily centered around analytics and team selection and tactics are increasingly influenced too. Arsene Wenger referenced 'Expected Goals', a very detailed statistical tool, yesterday when he talked about moving Ramsey back to the centre of midfield.

One reason you don't hear so much about how advanced it is at the top level is because clubs are desperate to protect any advantage they have and don't want to leak secrets to other clubs. This is a billion pound business and clubs will use any method they can to gain the slightest advantage. It's naive to think something like analytics, which has been proven to work across many fields, isn't already used heavily in football and really the statistics in FM are miles behind what is actually being used right now.

[benallen002]

I think some of you may be surprised how advanced statistics have become at the top level of football and how much influence they do have on decisions already. All the top Premier League clubs employ large teams of statisticians and analysts who have a big impact on all aspects of the club. Arsenal even bought out a company called StatsDNA so they could use their data exclusively.

Certainly clubs going right down to the lower levels use statistics to draw up a shortlist of potential transfer and scouting targets, even if the final decisions are made through watching individuals. It's well known the transfer policies of clubs like Liverpool, Man City and Arsenal, as well as the obvious Brentford and Midtyjlland, are heavily centered around analytics and team selection and tactics are increasingly influenced too. Arsene Wenger referenced 'Expected Goals', a very detailed statistical tool, yesterday when he talked about moving Ramsey back to the centre of midfield.

One reason you don't hear so much about how advanced it is at the top level is because clubs are desperate to protect any advantage they have and don't want to leak secrets to other clubs. This is a billion pound business and clubs will use any method they can to gain the slightest advantage. It's naive to think something like analytics, which has been proven to work across many fields, isn't already used heavily in football and really the statistics in FM are miles behind what is actually being used right now.

Just be careful though. Those clubs you mentioned might be investing a great deal in stats-based evaluations, but none of them (to my mind anyways) have sterling transfer records. And the last 3 "major" purchases by Arsenal specifically have all been "no-brainer" transfers and it's hardly as if City and Liverpool are buying obscure players for below-market prices and making them stars. Liverpool just raid S'oton every year and call it good, and even then the players don't play as well once they get there. i.e. sure, stats-based transfer policy may be on the rise, but where is the proof it's working better than the "regular" method?

The team that should be focused on is Leicester, because they are doing what everyone wants to do: winning lots of points without spending a lot of money. What led them to buy Mahrez, Vardy and Kante? Proper scouting, stats based research or a mixture of both. Do the players simply try harder at mid-level clubs because they know they are in the shop window and once they get to the top their performances drop? Is Ranieri just a tactical genius and can get more from the individuals through his system? Is their success this season really indicative of the quality of players, or is it just a run of form that will regress to the mean? etc, etc, etc.

I'm skeptical of macro-stats in baseball being worth their current value, let alone in football. Sure, the basic stuff is great (pass accuracy, headers/tackles win %, etc) but trying to spin that into some macro calculation that can equate to player value is a bit of a fools errand, IMO.

[puffascruffowitz]

I'm skeptical of macro-stats in baseball being worth their current value, let alone in football. Sure, the basic stuff is great (pass accuracy, headers/tackles win %, etc) but trying to spin that into some macro calculation that can equate to player value is a bit of a fools errand, IMO.

The same was said of sabermetrics in baseball for a long time. Granted, it's easier to apply in a sport with such a wealth of stats, and I don't ever think it would revolutionize football, but I don't see the harm. There is no holy grail of statistics, no matter how comprehensive they are. I do think though that there's always room for a different perspective.

Sabermetrics in baseball went from being an absolute joke to being almost universally accepted as a valuable tool for managers and scouts. Several general managers now, especially the younger ones, rely on them heavily in their approach to building a team. I'm not saying it would be easy or even that macro stats as you call them will be incredibly useful for evaluation, but I think it's worth a shot and at the very least would be interesting.

[BoxToBox]

Southampton have a fairly comprehensive system in place, thinking about it. Which is part of the reason they can switch managers and sell many players without too much negative impact.

[benallen002]

The same was said of sabermetrics in baseball for a long time. Granted, it's easier to apply in a sport with such a wealth of stats, and I don't ever think it would revolutionize football, but I don't see the harm. There is no holy grail of statistics, no matter how comprehensive they are. I do think though that there's always room for a different perspective.

Sabermetrics in baseball went from being an absolute joke to being almost universally accepted as a valuable tool for managers and scouts. Several general managers now, especially the younger ones, rely on them heavily in their approach to building a team. I'm not saying it would be easy or even that macro stats as you call them will be incredibly useful for evaluation, but I think it's worth a shot and at the very least would be interesting.

Absolutely. Every process or part of a process should be given a shot to prove its worth.

I live in the states and used to be a big baseball fan...and I love stats...so I am well aware of the progress that has been made in the last 10-15 regarding sabermetrics' influence in team building and management.

But there is a difference between stats-based decision making and finding value in a market. A lot of the value of advanced metrics in baseball is gone now exactly because it became "the norm".

[Shake Appeal]

The same was said of sabermetrics in baseball for a long time. Granted, it's easier to apply in a sport with such a wealth of stats, and I don't ever think it would revolutionize football, but I don't see the harm. There is no holy grail of statistics, no matter how comprehensive they are. I do think though that there's always room for a different perspective.

Sabermetrics in baseball went from being an absolute joke to being almost universally accepted as a valuable tool for managers and scouts. Several general managers now, especially the younger ones, rely on them heavily in their approach to building a team. I'm not saying it would be easy or even that macro stats as you call them will be incredibly useful for evaluation, but I think it's worth a shot and at the very least would be interesting.

Baseball has a huge advantage over football in the very nature of the game, though. It's itemized into set instances of play, and a season of 162 games generates a ton of data points that can be compared, not to mention the weight of historical data. Football is much harder to generate data with because of the sheer fluidity and chaos of it. So while stats guys have developed things like xG by comparing tens of thousands of similar shots from similar locations and moves over time, it's going to be a very long time, if ever, before they figure out how to measure a players' tactical nous or defensive positioning or even just the quality of a single through ball compared to other, similar through balls. Detecting, cataloging, and comparing like-for-like comparisons in large enough sample sizes is hard with football, and something like WAR (or even wRC+ or ERA+) may never arise.

I say that as someone who probably spends more time reading about football on stats' blogs and Twitter than actually watching it.

[puffascruffowitz]

Baseball has a huge advantage over football in the very nature of the game, though. It's itemized into set instances of play, and a season of 162 games generates a ton of data points that can be compared, not to mention the weight of historical data. Football is much harder to generate data with because of the sheer fluidity and chaos of it. So while stats guys have developed things like xG by comparing tens of thousands of similar shots from similar locations and moves over time, it's going to be a very long time, if ever, before they figure out how to measure a players' tactical nous or defensive positioning or even just the quality of a single through ball compared to other, similar through balls. Detecting, cataloging, and comparing like-for-like comparisons in large enough sample sizes is hard with football, and something like WAR (or even wRC+ or ERA+) may never arise.

I say that as someone who probably spends more time reading about football on stats' blogs and Twitter than actually watching it.

You're right it probably won't, but part of that problem is that the data just isn't there. Well maybe it is, you'd probably have a better idea. In order to better track players' defending in baseball now the landing spot of every hit ball in every game is recorded. They also have the FIELDf/x data now too which while unavailable to the public tracks the starting positions and movement of defensive players.

It's much simpler of course to track where a ball lands than to say, track a player's constantly changing defensive positioning, but I do feel there's more that can be pulled from a football game. The problem is how, not to mention what a massive undertaking it would be across countless leagues and countries.

[MBarbaric]

i bet stats can be used more in football but to make anything similar to moneyball in football is impossible. game of baseball has clear sections that are easily divided, it has moves that you can count on ten fingers while football is much more choatic. besides, not all football teams play same tactics nor players play same positions.

out of interest, has moneyball been implemented in other sports successfully and which? i know basketball has tons of stats compared to football (and its much more popular in the states than soccer so i'd guess moneyball being tried there before soccer)however it is still a "chaotic" sport even if it is less so than soccer. therefore i'd guess moneyball wouldn't work in basketball as well. or does it?

[Shake Appeal]

out of interest, has moneyball been implemented in other sports successfully and which? i know basketball has tons of stats compared to football (and its much more popular in the states than soccer so i'd guess moneyball being tried there before soccer)however it is still a "chaotic" sport even if it is less so than soccer. therefore i'd guess moneyball wouldn't work in basketball as well. or does it?

I think the idea of spotting overvalued and undervalued qualities and players applies in just about every professional sport where money is involved, but basketball and ice hockey both have strong analytics movements. I think a lot of football stuff has been borrowed or reapplied from ice hockey, since they're more similar than football and basketball or football and baseball.

[BoxToBox]

Any of you in this thread who have not read The Numbers Game should really do so. They analyze this very thing to death, particularly concerning how difficult it is to bet correctly on a game of football.

Very good book, indeed! I especially liked the sections on Roberto Martinez's Wigan side(and their abuse of shooting from distance), and the stuff on the timing of substitutions(which I have incorporated into my FM gameplay).

EDIT: The stuff about Stoke's controlling the ball by keeping it out of play was great too.

[fmfan74]

My tipp to people is go searching for young talent in which top teams are interested in. When you click on the players contract info you will see a list of teams wanting him. At times you might be successful in getting him. I do not go after young players that don't draw attention. It has worked real well profit wise. As you know young talent stats are dorment so my question is why are so many teams interested in this player and how do they know his gonna be a star.

Apart from using scouts I had to do raw scouting myself to find these players. My recent gem is mohamed saleh egyptian plaaer.

[ac13]

One of the most useful metrics they use in baseball now is WAR (win above replacement player). It's more of a cumulative stat, incorporating every aspect of the player's game. It produces a number of wins that the team has had with him on the field as opposed to a replacement player. It factors in hitting, fielding, baserunning, or pitching so is really a good overall view of the player's value to the team. Would be interesting to see WAR in football.

I wonder what a player like Messi's WAR would be? I'm guessing not terribly high since Barca seem to be able to play extremely well without him. Maybe when Luis Suarez on Liverpool would be a good example of a high football WAR since he was an absolute superstar on maybe not the best team overall.

In my second season where I struggled a lot (finished 7th which was ok but in a very bunched up table so got not many more wins than losses) I was surprised to see that 3 players in the team had hugely different wins-draws-losses records to the rest of the team. Easy to see where the weak links are if it's a position like left back where it's a like for like swap. I supposed looking at something like that for underperforming teams- seeing who when in the team raised their performances a lot could be useful for signings.