There is a hero within the FFS-community.
One that maybe doesn’t get the praise he deserves.
He is spoon feeding valuable information to everyone who bothers reading his weekly Hot Topic.
The hero goes by the username of G-Whizz.
The aim of the Hot Topic is to give a weekly, unbiased walkthrough of who the biggest favourites are, and which teams are most likely to keep a clean sheet. To provide the information, he uses Doosra’s stats and the bookies odds.
What I find most useful with regards to FPL is the percentages for the likelihood of clean sheets. In my opinion these numbers gives reasoning to start using “expected clean sheets” (or xCS) as a term. This article aims to discuss why.
Odds = likelihood of a certain event occurring
If you fancy betting, you should ask yourself “have I lost or made money on my betting endeavors?”. If you don’t fancy betting yourself, you should ask someone who does. Studies suggests that very few are able to constantly find valuable bets, and the inevitable outcome of not being able to do so, is that one becomes a “losing bettor”. This is the result of bookmakers being able to get the odds “right”. If they do, they will probably make money regardless of the outcome. The fact that bookies are able to “get the odds right”, makes it possible to convert their odds to percentages of probability, and this is what G-Whizz is happily helping everyone with each week.
The reason why it’s hard to make money of betting is because bookies are quite skilled at setting the odds. But after they’ve set the odds, it’s actually the bettors that make the odds a great representation of likelihood. Bookies have ways of adjusting their odds based on how bettors place their bets. If the odds they are offering sees bettors heavily favouring one side of the bet, they will adjust their odds to this information, and therefore get closer to the true likelihood (the collective wisdom of those betting), or at least changing the odds. Betting is in some ways a monetized “wisdom of the crowd”. By setting the odds well and adjusting the odds to how bettors place their money, they can trust their margins to secure profits. If bookies have reason to believe that the outcomes of an event is 50/50 they won’t offer odds that pay’s the bettor accordingly, they’ll offer odds that pay less. The difference between what the bookie actually believes is the likelihood, and the payout they offer is their “profit margin”.
If you accept that bookies odds a good tool for predicting an outcome, you should apply this knowledge to your selections in FPL. If you believe you have a better way, you should apply that method to winning money on betting. The only thing one needs to keep in mind when converting odds to probability is the effect of the bookies “profit margin”.
Why the profit margin matters
There’s various tools online that could be used to convert odds to percentages of likelihood. Running the odds of 2.0 (decimal odds) through a converter, like mentioned, would suggest a likelihood of 50%. However, the odds of 2.0 does not equal 50% likelihood if the other side of the bet only offers, lets say, 1.80. It’s fairly simple to pinpoint why, as odds of 1.80 would suggest a likelihood of 55,6%, making the two selections exceed 100%. The bookies want to make money on both sides of the bet, which would suggest that both odds are lower than what the bookies believe to be likely.
The scenario depicted in this example shows a bookie that wants to “earn” 5 cents on the dollar, the payout is only 95%. The 5% the bookies are trying to pocket are skewing the odds, and “hiding” the true percentages of what they believe is the probability for the two outcomes. When factoring in the profit margin of the bookie, one can calculate that the two odds represent 47% and 53% likelihood.
Check multiple sources
Checking the odds of one bookie gives a good indication of likelihood. Looking up multiple sources will give an even better idea, as potential outliers will have lesser effect on your own conversions. Luckily there are multiple sites that collect odds from a number of bookies, which makes it fairly simple to collect more accurate data. Some of them even convert the odds to percentages within certain selections, like Betbrain.com. If you want to use odds to make informed decisions in FPL, my suggestion would be to get familiar with that site, or similar ones.
How do we apply this to FPL?
There’s numerous ways to “exploit” this. Converting odds to percentages could give good indications of the implied chances of home-draw-away, a player scoring, a team getting a penalty, a player getting booked, indications of how many goals to expect in a match-up, or the implied chances of a team keeping a clean sheet. The last one is the one I’ve found most useful, and it’s also the one G-Whizz provides to the community each week.
My take is that the percentages in G-Whizz’s Tot Topics are fairly accurate representations of likelihood, and making selections based on the information presented in those posts should provide both more points and more peace of mind. If you’re constantly making sure the percentages are working in your favour, it should translate to a lot of points during a season. It also gives the added benefit of not beating yourself in hindsight, because you know that the decision you made, was based on making the objectively best move.
I’ve done the job of recording expected clean sheets (based on odds) for my own team so far this season. Furthermore I’ve registered the actual clean sheets gathered, and compared the two. I’d argue the results are quite convincing at this point. «Small sample size» aside, xCS seemingly translates very well to actual clean sheets. The 23 clean sheets recorded are only 1.24 more than xCS suggests.
The last couple of years I’ve used clean sheet percentages to dictate, in various degrees, the following selections:
A) Which players to fill slots (9)-10-11 in my XI.
B) How to prioritize 1st, 2nd and 3rd sub.
C) Who to transfer in (when there are glaring defensive weaknesses in the XI and great short term or good medium/long term investments are available)
A current example
This International break has been largely influenced by discussions on where to go from the highly owned Mendy, and naturally Digne and Laporte have been largely discussed by the community. Both good options, but they are only “the objectively right choice” for a number of managers, not necessarily for everyone that holds Mendy. Digne needs a companion that rotates well with his tricky away fixtures and City and Laporte’s average xCS won’t be as good going forward as it has been so far (5 of their next 8 are away).
I’m one of the current Mendy-owners, and I’m also amongst the managers most likely looking to play their WC come GW15/16. The way I’m set up, I need a replacement that comes with strong hopes of returns in gwk 13 and 14. At this point in time, the bookies market makes a case for a selection amongst Wolves or Man City, with Wolves being the slight favourite with 0.96 xCS and City 0.94 xCS over the two next fixtures. I’ll hold for pressers and more information and more accurate odds, but as it stands, Boly and an extra 1.4m looks more sensible for my team than simply opting for Laporte. If this ends up being the case and it all blows up in my face, I won’t be miserable, because I’ll simply be content with the fact that I made an informed and objectively sensible move.
Caveats:
1: Bookies use different profit margins on different markets, and unless you know their margins it’s difficult to convert this into probability. This basically means you should be wary of justifying selecting a striker/mid that has slightly better anytime goalscorer odds than a defender has for a CS.
2: All information affects bettors and odds. News in the papers, tweets on twitter, rumblings on reddit, pressers by managers, word of mouth by ITK. All this information will help inform bookies, and bettors in their selections, and the information will alter the odds. Odds are fluctuating.
3: Closing odds, the odds offered just ahead of kick off, are the best odds to use to get the “truest” representation of likelihood. this is simply because thats when all available information has reached its peak. Unfortunately for us managers, we don’t have the benefit of waiting for this, but the closer it gets to kick-off, the better the representation gets. Basically means: if you can, you should wait with making your decisions.
