Given the current set up of the game, catching the captain pick correctly and netting those hauls has become an essential cog in the wheel of achieving that illusive 10k finish. I recall the heady heights of the 2018/19 season when 16 weeks into the season, I was sat just inside the top 10k OR with 14 captaincy returns at that stage thinking (naively) how easy it was. Turn the clock forward to the end of last season and I was licking my wounds and cursing my luck (or judgement) at 17 captaincy blanks.
So at the close of the season I was determined to try and find a way to untangle the birds nest of captaincy and a great Youtube video by 2019/20 FPL Winner @JoshuaAbull on mathematics and how to win at FPL whetted my appetite. This gave me some powerful insights into returns and trying to predict them.
So armed with the idea to look at this analytically and the data I had crunched in the close season I dove in and this article will share some of the inner workings of the metric.
Rules and Shortlisting
The Initial rule of no player being less than 8.5 Million in starting value produced a shortlist of around 20 players. Now prices fall and rise and new players have (and continue) to come to the forefront of our minds – for example DCL and James were not originally shortlisted but are now – and stop the press, I’ve just added Super Jack Grealish into the mix too, hardly surprising given his form.
I also have discounted defenders from the metric and I think this so far (in hindsight) was a good decision. I keep defender captaincy as a differential at the back end of a season when chasing rank for example.
Three main measurements go into the metric that I have devised.
Potential for Returns
I originally worked on the likelihood of a 10+ points return (haul) and also a single return (4+) over the last 12 games but have recently adjusted the table to a slightly higher bar, 2+ returns and 1+ return over this season now we are ~ 8 games in. For example a midfielder with 2 returns would net at least 12 points (for 2 assists as a captain), appearance points (4) and probably BPS (4-6) making a return close or just above 20 points.
I then applied a weighting 66% single return vs 33% double to calculate a captaincy potential.
Form wise I originally worked on actual returns versus expected involvement but realised that this might suggest over performance (as well as great form) I have therefore made another amendment to use expected FPL points (Credit FFHub Opta Stats) in 90 Mins (xFPL90) over the season and over the last 3 matches (again with a weighting).
For me this is the biggest factor of them all when it comes to selecting a captain so I have spent a lot of time trying to fine tune it. It is particularly important to consider match ups (@fpllens‘ domain)
Again using Data within FFHub opta stats, which are my go to reference for those interested, I have devised a new FDR rating. This continues to be built around the 1 to 7 rating of Ben Crellins FDR planner for ease and takes into account the match ups between attack and defence for Big Chances, Shots/Headers in Box and Expected Goals (for and against).
For reference this created the following FDR by team (based on the last 3 games).
So given the above, one key match up this week would be Bruno Fernandes v WBA. Man U last 3 attack rate 4.3 but WBA actually 3.9 (-0.4) so this suggests a tougher match up than many will expect based on recent results. Vardy on the other hand looks to be an interesting prospect against a weakened Liverpool (1.9 v 5.2 = +3.3).
However; this is only a small data set so only accounts for 20% of the FDR calculation with 60% of it referring to this season in total and 20% sis still drawn from post lockdown data. I hope this sets the scene as to its complexity.
Don’t forget extra points for midfielders!
Of course Midfielders will get an extra point for a goal and also for a clean sheet and this is doubled. However, now I have moved to a xPts90 metric for form this is covered and therefore I have removed a previous 10% weighting which was in place.
I hope I haven’t lost you yet but now for the big decision, how do we weight form, returns and FDR? Well in reverse FDR is, as explained earlier for me by far the most important consideration and therefore it accounts for the largest proportion of the metric. Following this I consider the return frequency and then finally XPts90. The reason for this is that not only are we looking for consistency, we are looking for that haul. I then add them all together and voila!.
The Metric – Current Performance
So how have we done so far? Not too bad at all to be honest. In 8 weeks we have returned 7 in 8 and netted 138 Points, If we continue at that run rate we will be at 655 points for our captain. I just wish I followed the Metric like some do!
And what about GW9?
Well given all the COVID woes and injuries it’s too early to call so the big reveal will be on my twitter account next Friday.
Enjoy the break and stay safe!