Ricky's subs - the stats

There has always been a lot of criticism of Ricky's substitution strategy even from his supporters.  Here's some stats from an article on sporting analytics on the Times that back up some of those criticisms.  Consistently bringing people on with 5 minutes to go and expecting them to change the game has always struck me as pretty unrealistic.  Interesting if you're into this sort of stuff, easily dismissed if you're not:

"Bret Myers has a theory. It is a theory that he believes can help your manager help your team win football matches. It is a theory he first established by examining almost 500 games. He then tested it against a further 1,300 fixtures, drawn from the Barclays Premier League, from Spain, from Italy and from Germany.

Myers, a former professional footballer in the United States, is now an assistant professor at Villanova Business School, in Pennsylvania. His theory is: “Heading into the second half of a game, if your side is behind, it is important to make your first substitution prior to the 58th minute.

“If you are still behind, the second change needs to come before the 73rd minute. If still behind, the third and final change should be made before the 79th minute. The studies we have done provide a clear recommendation for managers when it comes to substitutions. The rule appears to hold regardless of quality of opposition, and whether or not your team is at home.”

If this sort of thing seems remote, abstract, it is not. Myers’s pattern has a real impact. “The study assumes that a substitution strategy is successful if the goal difference is improved in the end,” he says. “It has been found that teams are roughly 40 to 45 per cent successful when following the rule and 18 to 22 per cent when not.”

To put this in layman’s terms: if your team is losing, and your manager does not follow the pattern, your team has a one in five chance of turning the result around. If he does follow the pattern, he more than doubles his odds of success. That is why Myers thinks his theory can help your team win football matches."

While I agree with you JD, he's hardly the only manager in the A-league to do it.

Stats / correlations don't mean anything. You can twist them however you like.

You could also associate that 40/45 % success down to teams that had deeper squads so they were willing to bring on players earlier. Whereas a team with no impact on the bench wouldn't bring on their bench until late.

There are probably loads more ways to interpret it

To me what the article says is correct - for the simple reason that by giving your subs more time on the field, they will have more time to do something productive.

However, you'd expect your coach to know what sort of players he has on his bench. Some players take less time to get into the game once they come on, others seem to have an impact when they come on in the final minutes....

so I guess rules are there to be broken, but following them from time to time wouldn't hurt!!

I actually thought it was weird he got stick for the subs on saturday - I didn't see the game but thought at least he tried to make changes that would change the performance rather than just pre-planned like for like subs.

Playwithfire, like I said easily dismissed if you're not taken by this sort of thing.

The problem with this kind of research is that you don't know what would have happened if earlier subs weren't made.  The starting player may have well scored himself if he stayed on.

Yep. There's way too many variables that come into play for this kind of thing to have any real meaning.

If you apply the stats to an individual game, and say that if you follow the rule you WILL do better than if you didn't, then of course there are too many variables and you are being stupid to expect better things from you team because of it. However, across many many games between many many team, it is possible, on average, for teams to do better based on things like this. 

It's just Game Theory applied to an aspect of football, there's an 'optimal solution' to the game that will cause you 'win' in the long run, but playing this optimal solution does not guarantee you a 'win' in an individual game. Here, 'win' is defined as doing better in the match than you were doing before, and note that it is not equal to actually winning the match.

There's a book called Soccernomics that analyses a whole bunch of things about football using stats and game theory, and it's a really interesting read.

Not sure I agree with that. I haven't seen this guy's research in detail, but without filters taking into account the relative strengths of teams involved, the game conditions, the reasons for making the substitution, etc, all you're coming up with is a false positive. In other words, it's common sense that the longer a player spends time on the field he has a better chance of making a positive impact, but in terms of substitution strategy that doesn't really mean much unless all of the other variables are accounted for, since they all play a much more significant factor in the outcome of the game overall.

Without seeing any of the research, I completely agree with you EG, it's just common sense. I was just saying that you can't apply the research to a single game scenario and expect it to have an effect. Guess I didn't explain it very well before.

Oh, sorry, looks like I misunderstood what you were saying. Multitasking is clearly not my forte.

Speaking of subs, why don't they change the rule so teams can have five subs on the bench including a keeper but still only use three. Would at least give coaches a few more options.

yeah ive always said an u21 spot for the bench would be awesome.

I have read Soccernomics, it is a good book - although it is not very original and refers to Moneyball far too often.  Still, I reckon every coach would learn at least one thing new from it.  Myers theory makes sense too.  If you are behind and playing worse than the other team at half time then you should make a substitution at half time.  Your team has already spent 45 minutes demonstrating that things aren't working. Waiting until the 60th (or 58th) minute just means another 15 (or 13) minutes of things not working.  Halftime tea cup tantrums and tactical tweaks hardly ever make much difference.

You don't think the beginning of the second half is like a reset though? I think it pays to see if they perform in the second game (half) before trying something else. You generally would know within 10 minutes of the half though. 

What ever happened to the old theory that it was Tim Brown's fault we lost or drew so many matches?

When did he ever play for us?

If only we'd picked up some skills and tactics from beckham and the galaxy while they were here. 

If only Boca had sent their senior side over - we could have learned a lot.

SACK RICKI!

(Hire Ian Crook!)