To Rotate or Not to Rotate - That is the Question (GK)

What's this Goalkeeper Rotation Strategy All About, Anyway?

  

Curiosity killed the cat.  That should be the subtitle of this article.  Most of the FPL lessons I've learned have been learned just a little too late.  Last year, my goalkeeper strategy ended up being laughable.  If it didn't take up too much space in this article and didn't make me look like such a complete idiot, I'd describe it and give you all a nice laugh.  Instead, I will tell you that I didn't know much about a goalkeeper rotation strategy until after last season ended and the creatures of the FPL world came stumbling out of their caves and started discussing season strategy in earnest over the past three weeks or so.

I unconsciously always "kinda" had a strategy in mind, but it was always something along the lines of "Well shit, what if __________ gets hurt and doesn't play; what the hell am I gonna do!?!"  That was about the extent of it.  Then a few weeks ago, as I started to dig deeper into "real" strategies to start the season with, I saw some mention of this goalkeeper rotation strategy, mostly focused on the idea of finding two budget keepers with a favorable home and away rotation schedule.  I had two immediate reactions.  The first was, "Brilliant; why didn't I think of that before!?"  The second was, "Wait a second, could that really work?"
 

My second question was brought on by the fact that when it comes to FPL, I more or less ascribe to the chaos theory; that anything can and will happen despite our best use of logic and rationale.  How else do you explain Jonathan Walters scoring two own goals and missing a penalty in one game last year or the fact that Ali Al-Habsi scored more than twice as many points in away fixtures than he did in home fixtures during the first 20 gameweeks of last season?
 

When it comes to a strategy of rotating keepers, my mind starts going crazy, mainly because rotating keepers involves something with the goalkeeping position that I don't like to spend much time on each gameweek; deciding on one keeper vs the other.  FPL is unpredictable enough as it is.  I've seen so many games where, expecting a team at home and in good form to win handily, I've been disappointed to see them lose.  Yes, the Readings of the world can march into the Etihad and put a spanking on the City boys just as easily as Guzan can save a penalty and grab a tenner at home to United.
 

In my mind, the concept of rotating keepers and relying on the home team keeper to have a better game than the away team keeper was fraught with conflict and uncertainty.  The concept made sense, but FPL addicts have short memories, and mine are full of gameweeks last season where I played the conventional wisdom of relying on home defenders who came up short, both against strong AND weak opponents, while my mini league rival snapped up a clean sheet with their Reading defender playing away to a much stronger team.   This worried me.  A lot.  With two keepers, one playing home and one playing away, I was scared at how many points I could potentially lose by picking the wrong guy and having a bucketload of points sitting on my bench.  The pain that that would cause was almost unthinkable.

So, with this new seemingly logical concept of a home/away budget goalkeeping strategy battling it out inside my head with my intuitive skepticism, I did what most people afflicted with crazy amounts of FPL curiosity would do; I created graphs and charts and spent way too much time over the past week going back through last year's stats to put this theory to the test.  After all, the difference in many mini leagues could be 15-20 points; a spread that easily could be covered by a good goalkeeping strategy vs the preposterously stupid approach I used last year.

My Research Approach



To test things out, I started by looking at the 20 goalkeepers who started gw1 during last year's 2012-2013 EPL season.  I used the assumption that most everyone utilizing a home/away rotation strategy were doing so to save important season-starting budget money, therefore I only looked at rotation combinations that would require an outlay of 9.5m or less.  I also only focused on the first 20 gameweeks of the season, figuring that this strategy starts out as primarily a way to make it to the January wildcard, where widespread changes could be made and the strategy could be changed if it wasn't working.

The first thing I looked at were how many games each of the gw1 starters played in during the first 20 gameweeks of last season.  The results were a little surprising.  Of the 20 keepers who started last year's season, the number of games each played in the first 20 weeks were as follows:

<10        3     Given, Green, Davis

10-15     7     Szczesny, Ruddy, Vorm, Federici, Friedel, DeGea, Foster

15-19     5     Schwarzer, Reina, Krul, Jaaskelainen, Cech

20          5     Mignolet, Begovic, Al-Habsi, Hart, Howard



This scared me right off the bat.  Relying on the home/away rotation strategy looked daunting considering that only half of the keepers who started the season featured in 15 or more games during the first 20 of the season.  That already meant that 10 of the keepers would have had to have been transferred out or you'd be left with one reliable keeper, essentially throwing out the home/away strategy.

So I then focused on the 8 keepers who played in 19-20 games during the first 20 gameweeks last season.  They were Mignolet, Schwarzer, Begovic, Jaaskelainen, Al-Habsi, Cech, Hart and Howard.  Only 5 of these 8 were priced at 5.0 or less, which meant they were the only 5 whose combinations I could look at to truly see how a home/away strategy would have worked last year.

Before getting into the detail of the specific keepers and the combinations that I would review, I looked at overall 20-game statistics for all of the keepers.  What I found there made this strategy look even less likely to come up with results that would impress me.  Here are a few quick stats that jumped out at me.  When looking specifically at the 20 keepers who started gw1 last year, there were 1,005 total points accumulated amongst them from gw1 to gw20.  Of those 1,005 total points, 534 were tallied by the home keeper, with 471 scored by the visitor.  That's a 53.13% to 46.87% spread; not exactly a stat that would cause you to jump on the home/away bandwagon.  All in all, home keepers averaged 3.543 points per game played vs 3.171 points per game played for the away keepers. At a difference of a little less than 0.4 pts per game, you're looking at an 8 point swing over the course of 20 games.  Not crazy exciting, is it?

Despite how the raw data looked, I still had not looked at how any of the data applied to the specific home/away rotation combinations that involved the 5 keepers in my study.  That was my next step.  What results would I find?  Whatever they were, I sure hoped they were worth the crazy amount of time that I had spent combing through the numbers AND the crazy amount of time you're spending following along with me through this lengthy article (if in fact you are still around - and I wouldn't blame you if you weren't; but hell, I did the research and I'm gonna follow it through;).

The Testing and The Results (Please Bear With Me)



Ok, so now it was crunch time.  Time to put this home/away goalkeeping strategy to the test.  Looking at the 5 keepers who were priced 5.0 or less, who could combine to create home/away strategies priced 9.5m or less, and who included only those keepers playing in at least 19 of the first 20 games, I had 7 potential combinations to look at.  They were:

Mignolet/Begovic
Mignolet/Jaaskelainen
Schwarzer/Begovic
Schwarzer/Jaaskelainen
Al-Habsi/Begovic
Al-Habsi/Jaaskelainen
Jaaskelainen/Begovic

My approach for the test was simple.  Using a strict home/away strategy, I determined that for gameweeks where one keeper played at home and the other played away, the home keeper would be used automatically (why else would a home/away strategy be used?).  For gameweeks in which both keepers played at home or both keepers played away, I selected the keeper who had the easier fixture based on my understanding and recollection of last season's form, table, etc. 

Here is how things panned out:

Mignolet/Begovic:

# of games allowing for home/away rotation:     7/20
# of games away keeper scored more points:      1/7
Points lost:                                                            2
Total Points:                                                          87
Cash Outlay:                                                         9.5m
Value:                                                                   9.158

Mignolet/Jaaskelainen:
# of games allowing for home/away rotation:     15/20
# of games away keeper scored more points:      6/15
Points lost:                                                           18
Total Points:                                                         87
Cash Outlay:                                                         9.5
Value:                                                                   9.158

Schwarzer/Begovic:
# of games allowing for home/away rotation:     12/20
# of games away keeper scored more points:      5/12
Points lost:                                                           15
Total Points:                                                        76
Cash Outlay:                                                        9.5
Value:                                                                  8.000

Schwarzer/Jaaskelainen:
# of games allowing for home/away rotation:    7/20
# of games away keeper scored more points:     3/7
Points lost:                                                          6
Total Points:                                                       68
Cash Outlay:                                                       9.5
Value:                                                                 7.158

Al Habsi/Begovic:
# of games allowing for home/away rotation:    12/20
# of games away keeper scored more points:     7/12
Points lost:                                                          31
Total Points:                                                       61
Cash Outlay:                                                       9.5
Value:                                                                 6.421

Al Habsi/Jaaskelainen:
# of games allowing for home/away rotation:   9/20
# of games away keeper scored more points:    3/9
Points lost:                                                         10
Total Points:                                                       73
Cash Outlay:                                                       9.5
Value:                                                                 7.684

Jaaskelainen/Begovic:   
# of games allowing for home/away rotation:    20/20
# of games away keeper scored more points:     9/20
Points lost:                                                          43
Total Points:                                                       79
Cash Outlay:                                                       9.0
Value:                                                                 8.778

The Results
 

So, we've made it this far.  I now had enough data to compare a strict set of home/away rotation combinations against the "find a guy you like and stick with him for every game" strategy; or as I sometimes like to refer to as the "big guy and nobody" approach.  Based on last year's data, here are some comparisons to look at:
 

(Please excuse the formatting; I still am having a tough time with this on the Blogger app)



  Goalkeepers                              Total Points              Cash Outlay                    Value

Begovic/nobody                               91                              8.5                              10.706
Mignolet/Begovic                             87                              9.5                                9.158
Mignolet/Jaaskelainen                     87                              9.5                                9.158
Jaaskelainen/nobody                        76                              8.5                               8.941
Mignolet/nobody                             79                              9.0                                8.778
Jaaskelainen/Begovic                      79                              9.0                                8.778
Schwarzer/Begovic                         76                               9.5                               8.000
Al Habsi/Jaaskelainen                     73                               9.5                               7.684
Cech/nobody                                   77                              10.5                              7.333
Schwarzer/Jaaskelainen                   68                              9.5                               7.158
Schwarzer/nobody                           63                              9.0                                7.000
Al Habsi/Begovic                            61                              9.5                                 6.421
Hart/nobody                                     70                              11.0                               6.364

These results leave me with a number of reactions.  First and foremost, it does seem that a home/away rotation strategy does seem to work better than picking one of the elite keepers and leaving them to their own devices.  Finding a cheap guy who has an exceptional year, as was the case for Begovic, is pretty much a good stroke of luck but one that I'm sure was received joyfully by those who picked him as their sole keeper last year.  The key seems to be finding low to mid priced guys who can be effective and either rotating them or relying on them to shoulder the load and peform the way that Jaaskelainen and Mignolet did last year (at least to get you through to the January wildcard).

Sadly for me, if you can withstand the gutpunches that will come as your away keeper outscores your home guy, the rotation strategy looks like a pretty good option, and one that I wil look at more closely as I put my squad together over the next couple of weeks.

The debates surrounding this strategy will continue, and each FPL manager will probably be able to find empirical data to support the way that they want to go.  Begovic's standout start and Al-Habsi's unlikely tally of 40 away points to 19 home points during the first half of last season are two pieces of data that could easily sway you away from the home/away rotation strategy, but the fact that home/away rotation strategies accounted for 2 of the top 3 point tallies in my test also shows that the rotation can work effectively.

In another interesting twist, it was odd to me that the Jaaskelainen/Begovic rotation option allowed for every single one of the first 20 games of the season to be played with a home/away rotation approach.  It was set up PERFECTLY for the home/away rotation.  However, a total of 43 points were lost when the away keeper outscored the home keeper, and the combination also resulted in 12 less points than Begovic scored on his own and only 3 more points than Jussi scored on his own.  Interesting to say the least.

At the end of the day, we'll all probably go with our gut anyway.

Okay, time for me to finally do some tinkering.. good luck everyone!
 

Please click on this link to follow me on Twitter

P.S. Thanks for bearing with the inconsistent formatting.  I will clean that up later, along with adding some graphics. If you have any advice for me on how to fix formatting issues on Blogger.com please send them on.  It seems to have a mind of its own sometimes.... Cheers!