Sub Standard? Rafa's Subs Record
This article originally appeared on The Tomkins Times on April 12th 2010, and has since been added to the site's 'Best Of'
With thanks to Graeme Riley for providing the match statistics.
I started the research for this article three weeks ago, but after the criticism of Benitez for his decision to substitute Fernando Torres with 25 minutes to go away at Birmingham, it seems more pertinent than ever to try and assess how much success Rafa has with his substitutions.
Assessing the impact or effectiveness of substitutions can be a difficult task. Some changes that occur are enforced through injuries, or as a part of a re-shuffle following a sending off. Others are used to give a player returning from injury some game time, or to let a star performer milk the applause of the crowd and get some rest once the game is won.
But most are used tactically to try and improve or protect the result, and I am going to look at how successful Benitez is in this respect, as well as who Liverpool’s top performing sub is, and whether Andy Gray is right to criticize Benitez for not making substitutions before 65 minutes (and indeed if this is even the case).

It will probably come as little surprise to learn that Fernando Torres is Liverpool’s best performing substitute. A record of 4 goals from 14 sub appearances in all competitions gives a rate of 3.50 games per goal. But if you break it down to minutes spent on the pitch, then Torres as a sub actually has a goal rate of one every 71.5 minutes. This makes you wonder how good Liverpool could be with Torres on the pitch and a Torres equivalent to come from the bench with stats like that.
The well documented financial situation at the club makes this an impossibility. To give this some context, the Manchester United bench in the recent game with Liverpool featured the £30.75m striker Dimitar Berbatov; had United been chasing the game late on they could have brought him into play (alas they were not). Liverpool’s entire bench of seven subs that day only cost £43m, and that’s if you count Alberto Aquilani as a £20m signing, of which there is much debate as to whether this  figure is accurate or not.
So with an uneven match up like that, Rafa needs to make the most of his substitutions. My research suggests that he does. Using a simple system, applied to league games only, I can prove that Benitez does make effective use of his substitutions. I have focussed on league games only, as in 2 legged cup ties, a draw (or even a defeat) in a match, whilst not the desired result, can be suitable to get through, making the impact of a substitution harder to analyse.
When a substitute enters the field, there are three potential impacts upon the result: it can improve, stay the same, or get worse. Below is a table showing how Benitez has fared:

Put simply, when Benitez brings on a sub, the result improves 14.47% of the time, and only gets worse 5.95% of the time – nearly 2.5 times more gains than losses.  The vast majority of games (79.58%) fall into the no change category; the recent Birmingham away game being an example. Benitez made a controversial substitution, and whilst Liverpool didn’t win the game, neither did they lose.  In that instance the gamble didn’t pay off, but these figures show that the gambles pay off more than they backfire.
In six years, only once have all 3 points been lost following a substitution, and that was away at Tottenham last year, a game in which Liverpool had the chances to win comfortably, and only lost to a last minute goal. Whilst it is concerning that 80.15% of substitutions in losing games don’t result in any points being gained (and an alarming 100% for this season), the fact that Liverpool only drop points from winning positions following substitutions 5.52% of the time shows how difficult it is for teams to turn games around once that vital first goal has gone in and a lead established.
Whilst this method of analysis is simplistic, and doesn’t take in to account whether or not the team’s performance gets better without the result improving after a substitution (as is widely agreed occurred at Birmingham), it is at least factually accurate and is based upon results, the ultimate measure of success in football.
I have neither the time or the interest in carrying out this amount of research into other top teams to see how they fare, so don’t know if Benitez fares better or worse than other managers. But I’m certainly prepared to put my faith in someone with these statistics behind them.
Therefore, if Benitez is having this kind of success, does it really matter what time in the match he makes his substitutions? According to some people it does, and he should be making substitutions before the 65 minute mark when Liverpool are struggling. But is the 65 minute figure accurate?

It appears it is. I have compiled figures discounting 1st half substitutions, as these would be for injuries rather than tactical changes. Across all competitions, Benitez averages a first sub time (second half only) of 64.32 minutes. The pundits are broadly correct on this issue then. But are they right to criticise Benitez for it?
I have looked at the average first sub times for the opposition against Liverpool in league matches to see how Benitez compares to other managers.

Surprise, surprise, Benitez is criticised for something which other managers largely appear to also do. The season where Liverpool’s first tactical substitution was on average the latest after the oppositions, was the most successful league season the club has had in years. More relevant to this study, it was the year with the highest number of substitutions that turned losses to wins: six. The other five seasons combined have a total of one. So once again, these figures appear to vindicate Rafa’s decisions and methods.
It would take a lifetime to analyse the same information for all teams, but Rafa’s strategies do appear sound, successful and not dissimilar in principle to those of other managers. It will probably be too much to ask for anyone who is paid to analyse football to pay any attention to this however.
Please take a look at my other articles, a list of which can be found here.