Standard deviation and other perversions poker taught me about FPL

Poker mirrors FPL in many unsuspected ways.  At the heart of both games are strategies to maximise the accumulation of points.  It's just that in poker, the points are in the form of chips.  In tournament poker and FPL we have to keep in mind how our chip stack or total score is faring relative to others on our table or in our mini-league, as well as compared to those on the overall leaderboard.

Understanding the strengths and weaknesses of hands is as important in poker as it is with our squads in FPL.  Knowing whether a hand plays better in raised pots, heads up against single opponents say, rather than limped ones multi-way, is a little like weighing up whether transfers help or hinder our planned orientation to a 3-4-3 or 3-5-2 formation.

In poker, it's really important to strike the right balance between 'value betting' and bluffing.  Otherwise, we'll be pegged as easy-to-read, fit-or-fold type players, and we'll never get paid-off when we have a big hand.  This brings to mind the balance we need to strike between template players and differentials in FPL.  After all, it's difficult to make any headway in our mini-leagues if we've got pretty much the same team as those above us in the standings.  And, as pointed out to me by @kingsoflyon7, having too many differentials is akin to the rookie error of calling bets on the 'flop' and 'turn' chasing a 'runner runner' 'gutshot' straight draw, which relies on improbably catching perfect cards on both the turn and the river.

Another balance to strike in poker concerns the careful sizing of bets that keep as many options open to us as possible depending on what our opponents do next.  This translates to FPL in terms of the spread of player price ranges within our squad.  It is a balancing act we must perform well if we are to succeed, and relates to the proportion of our budgets we allocate to different positions within our squads.  Going uber-cheap on defence, for example, by starting the season with five 4.5m-or-under defenders runs the high risk of our squads requiring immediate surgery should the GW1 lineups feature few if any of our bargain basement buys.  Having a good balance of price ranges insures us against unforeseen circumstances such as bans and injuries.

An undeniable fact of FPL is that the so-called big hitters will sometimes blank, just like the best hands in poker (pocket aces, pocket kings) are often busted by lesser hands.  Having the best cards or players is no guarantee of success in either game.  The skill/luck quotient feels very similar, with our skill constantly being tested against our opponents' luck.  Both games involve numerous risk/reward decisions.  In neither game will perfect decisions always be rewarded.

On the contrary, they are often punished.  Like when we transfer out players who are in poor form and about to embark on difficult fixture runs.  Trading such players out for ones in good form facing easy fixtures seems like a no-brainer, but time and again, discarded players make a mockery out of the form book.

It's like making disciplined folds in a poker tournament with a marginal hand like Ace10 off-suit in early position on a high action table, only to suffer the torment of a King, Queen, Jack rainbow flop, meaning we'd have had the best hand possible (AKQJ10 for the nut straight) had we not folded.  And then we die a little more inside as two players go all-in and fail to overtake our nut straight, meaning we could have trebled our chip stack.   

Does this mean we were wrong to transfer/fold those players/hands?  Of course not.  It's simply 'standard deviation'.  Take flipping a coin 100 times.  On average you'd expect there to be 50 heads, but because of 'standard deviation', around 32% of the time there will be fewer than 45 heads or more than 55.  In other words, swings (up and down) are inevitable.  The right way to deal with downswings is to simply accept the fact that they happen, remain calm and keep playing your best game.

An understanding of probability is a key component of both games.   More often than not a player considered most likely to score a goal in a match is best priced by bookmakers at odds against to do so, meaning they deem it more likely he won't score than he will.  This weekend, for example, Jesus and Kane are currently both only even money (50/50) to score against newly promoted sides Brighton and Newcatsle

Last year's top points scorer in FPL, failed to score in 21 of the 38 games he played in, which is roughly 55% of the time.  What's more, Sanchez only provided an assist from those 21 games on 6 occasions, meaning in 15 games (39.5%) he scored no attacking points whatsoever.

Let us assume, however, that we 'know' Sanchez will score in half of the games he plays over the next 3 seasons.  It would be well within the normal distribution of goals predicted by 'standard deviation' for him to go 8 games in a row without scoring.  Especially if those are the gameweeks that I own him!  Actually, a sequence of 8 blanks in a row over the course of 100 games can be expected to occur around 17% of the time.  Likewise for 8 scoring games in a row, which naturally will happen when I don't own him!

'Tilt' is a poker term for a state of mental or emotional confusion 

or frustration in which a player adopts a less than optimal strategy, 

usually resulting in the player becoming over-aggressive - Wikipedia.

Going on tilt is something that FPL players are also particularly prone to, commonly in the form of making 'rage transfers' for multiple hits, especially after low scoring gameweeks.  Playing hundreds of thousands of hands, however, has reconciled me to the harsh reality of 'downswings' being an inevitable part of poker.  Poker has taught me to be a lot more accepting of bad luck, which in turn has helped me to handle the bad beats routinely administered by FPL with more serenity than most managers are seemingly able to muster.

Planning ahead is key to being a good poker player.  There are 3 streets of betting after the pre-flop action.  Namely, flop, turn and river.  In order to run an effective 'triple barrel bluff', for example, we need to size our bets on the first 2 streets (flop and turn) in such a way so as to be able to make a big enough bet on the third (river) to discourage opponents from calling.  The mistakes novice players make often arise from not thinking ahead.  As a result, they'll fire pot-sized semi-bluffs on the flop and turn say, only to discover if they miss their draws that they only have enough chips left to bet a small fraction of the pot on the river.  This makes 'crying calls' or 'hero calls' of their bluffs so much more likely.  When I act on the flop, I do so with a clear plan for how I intend to play the next 2 streets (turn and river) also.

Carrying this mindset over to FPL means I'm always thinking two or three gameweeks ahead at the very least when making transfers.  In this way, self-inflicted predicaments can be avoided.  A good example from last season was my keeping a Spurs slot free in my squad (as well as sufficient funds in the bank*), for when Harry Kane returned from his second injury spell.  That was at a time when lots of managers already had 3 Tottenham assets in their team (predominantly drawn from Walker,

Davies, Alli, Eriksen and Son), which meant a minimum of 2 transfers would be required for them to acquire Kane.   

[*NB:  Good 'bankroll management' is essential in both games: in FPL to ensure funds are available for our marquee signings; and, in poker to avoid going bankrupt!]

FPL managers are generally reluctant to take points hits for so-called 'sideways moves' no matter how preferable an alternative player from the same team might be, and my late charge up the rankings towards the end of last season probably owed much to the significant number who stuck with Alli and Eriksen, rather than twist to Kane.  Such an aversion to sideways moves was to prove especially costly, as Kane scored a massive 71 points in the last 7 games, averaging just over 10 points per match (PPM), on his way to winning a second consecutive golden boot. 

Most of the decisions I make when playing poker are informed by the concept of 'expected value' (EV).  Essentially, I'm always asking whether or not my next actions are '+EV' or '-EV', and if my lines are the ones that extract the most value on average.  My aim, therefore, is to find the sweet spot with bet sizing and actions that make me indifferent to what my opponents do next, because in the long run these lines of play will show a profit.

In the Pot-Limit Omaha cash games I play, it's not unusual to be heads-up facing an all-in pot-sized bet on the flop, when I am 50/50 to win the hand.  Effectively, each player is getting 2/1 odds on an even money shot, so it's a +EV scenario for both.  Setting aside what has already been invested into the pot hitherto, the EV of folding to the all-in shove is zero, whereas the EV of calling is half the pot-sized bet, because the amount I win is twice as much as that I lose:  

0.5*2 + 0.5*-1 = 0.5

So if facing a $60 bet, a call would on average yield $30 profit.  Naturally, this is only in the aggregate.  It's not actually possible for me to win half the pot-sized bet on an individual call, as I can only either win 2 bets or lose 1.  Calling is clearly better than folding.  Somewhat perversely perhaps, calling in the example above is still correct even if we 'know' we only have a 40% chance of winning and are more likely to lose than win:  

0.4*2 + 0.6*-1 = 0.2

No matter how many of these 'coin flips' are lost to begin with, calling will always come out ahead in the long run.  Our friend 'standard deviation', however, ensures that the long run can be much longer than most people think.

Maximising expected value is at the core of winning at poker.  And so many decisions in FPL can be thought of as EV ones too, where we weigh up the relative pros and cons of our next moves.  We are continually faced with 'coin flip' situations in FPL.  Viewing transfer options through the prism of EV leads to more optimal play in my opinion.   

By way of example, I remember an EV decision I made with my second FPL wildcard last season, concerning which two attacking assets from Spurs to own.  Namely, whether to go (for reasons of budgetary constraints) with Kane and Son, or Alli and Eriksen, both of whom were in a rich vein of form, and a popular double up at that time.  My calculation that the expected value of the former was greater than the latter proved to be vindicated in no uncertain terms, and propelled me up the overall rankings to good effect. 

Obviously, the variables in FPL are infinitely harder to quantify than is the case with the precise probabilities that apply to poker.  The best we can do is analyse available historical data, and consider metrics like points per match (PPM), or minutes per point (MPP).  Furthermore, we can cross reference these findings with big chances (BC) and shots on target (SOT), which I deem to be the reported statistics* with the strongest correlation to future goals, and FPL points.

[*NB:  xG or Expected Goals have now begun to be reported more widely this season, and are even better indicators in my opinion.]

In the case of Kane & Son vs Alli & Eriksen, I don't have any record of my estimates at that time, but for the sake of argument, let's say my expectation was that Kane would average 9 PPM, Alli 7, Eriksen 6 and Son 5.  Such a process would lead me to conclude Kane and Son were the optimal pairing as (9+5) > (7+6).   

With hindsight we know Kane and Son averaged 10.1 and 5.7 PPM respectively, for a combined 15.8 PPM, whilst Alli and Eriksen averaged 5.3 and 6.1, for just 11.4.  My EV decision averaged 4.4 PPM more than the alternative over the last 7 games, netting me an extra 31 points overall.

Unquestionably though, 10 PPM is not sustainable by any player, but I'd argue that the lower output elsewhere merely serves to highlight that 'regression to the mean' is just as much a feature of FPL, as it is in poker, and most other places besides.

 

Now you might feel this is a case of me finding facts to fit my theory.  After all, this is the same Harry Kane who incurred the wrath of managers earlier in the season by scoring appearance points only in consecutive home matches against Burnley and Hull.  Happily for me, disproving allegations of planting and rigging the evidence, doesn't fall within the remit of this article.

Meanwhile, may the FPL flops be with you.

Coley a.k.a @barCOLEYna