Sunday, July 29, 2012

Points vs Price

Two trains are set to depart at the same time for your given destination. One will take 90 minutes and cost $20 (apologies to UK readers but my keyboard doesn’t have a pound sign), the other will take just an hour but cost $70. In most cases you would opt for the cheaper train, right? The other is faster, but not sufficiently so as to justify the extra cost. Why then, do we often still talk about fantasy points in terms of absolute points earned? Every point you earn comes at cost, namely the opportunity cost of using your limited resources to strengthen other areas of your team. Getting this balance right is the name of the game, and it’s where we’ll start our look back at the 2011-12 season.

Note that in some cases, continuing the analogy, you would want to take the faster train. Perhaps there is a vital metting you need to attend or you are a millionaire and thus $50 is immaterial to you. The same goes for the fantasy world. You may wish to ‘overpay’ for a certain player given his upcoming fixture or you may have sufficient free cash as to allow you more flexibility with your high cost players. Generally these examples will be the elite players and will be discussed at a later date.

For now though let’s take a look at last season, specifically the correlation between price paid and points returned for each position:


I’ve long been an advocate of the goalkeeper rotation system, in large part because of this analysis. We can see that the correlation between price paid and points earned is almost nonexistent and every year there are ‘keepers like Michel Vorm or Tim Krul who deliver elite returns for less cash than their expensive counterparts (this doesn’t even account for the fact that significant advantages can be gained from rotating two ‘keepers whose fixtures gel together well).

Last season was no different with the top rated ‘keeper on a points per 90 minute basis (David deGea 4.57) managing to only just top the aforementioned pair of Vorm (4.19) and Krul (4.14). Simon Mignolet (4.28) was even closer though he came with some other issues as Sunderland couldn’t decide what to do between the sticks.

The really persuasive part of the data for me is the downside of those elite players. Managers who don’t favour the rotation system cite the benefit of consistency they get from their elite player and the minimal downside they bring to the table. However, with Cech (3.53) and Reina (3.6) both coming in at significantly below four points per 90 minutes you would have had a decent chance each week of rolling the dice with players like Begovic, Schwarzer or any of the aforementioned cheaper options and still finish head of the stars of the once rock solid Liverpool and Chelsea defenses.

Looking ahead to the upcoming season, there is one name that has the potential to upset the rotation theory: David de Gea. If we plot players’ P90 scores against their 2012-13 prices, deGea suddenly looks well priced:

This ignores of course the opportunities that can be gained from rotating ‘keepers, but with the potential position leader available for 6.0m it at least bears mentioning here. I will shortly be analysing the possible rotation options available so we can touch on this again then.

Logic would suggest that defender values would follow the same pattern as with ‘keepers, but the absence of saves - which act as an equalizer for the ‘keepers from weaker sides - is key.

Excluding those who played just a handful of minutes, 20 defenders managed to exceed four points per 90 minutes played last season. 11 of those players play for one of the Manchester teams while a further six play for Arsenal, Chelsea or Liverpool. That leaves only Baines, Gallas and R Taylor, and with an opening price tag of 8.0m we can hardly call Baines a surprise here. In short then, at the top end of the market, you get what you pay for (to a point), with the above group coming in with an average price tag of 6.1m (and even that is dragged down thanks to the unlikely minutes enjoyed by Johny Evans thanks to an early Vidic injury). Unlike the Vorms and Kruls of the world, who go toe-to-toe with the big names and often win outright, in the defender market you can only hope to be comparatively valuable.

The best method last season then was probably to grab one or two elite performers in the ~6m range (Kompany, Richards, P Jones) and then supplement them with a collection of budget players from good teams getting unexpected minutes (Evans, Hibbert) or rotatable options who enjoy some upside (R Taylor, N Taylor). This is hardly breaking new ground but it’s amazing how many teams you see with either four elite defenders or a total spend of around 22m for their back five, neither of which is likely to be the optimal strategy. We'll get into the mix of players for this year later, but for now I think we can conclude that you should be targeting players from top teams who come at a slight discount and exercise caution when looking at the really expensive 'attacking' players. Budget options can be found every year and should fill out the remainder of the back line.

The correlation between price and points for midfielders is the strongest among all positions (54%) with only Ben Arfa and Sigurdsson managing to join the elite group with a P90 over 6.0:

It's interesting to note that the three players valued over 10m (Malouda, Gerrard and Lampard) all failed to even exceed the break even line with Malouda of course having a particularly disappointing season. As with the defenders above, this just shows that no one is a lock to have success and so early season struggles should be taken seriously and actioned accordingly.

Again, it looks like the best option last season was to pair a couple of elite players with some cheaper options, rather than attempting to create the 'ultimate' midfield which so many managers did. Of the 12 midfielders who cost 9.0m or more, five fall below the average line with three having particularly poor seasons (Adam, Cahill and Malouda), and while it's easy to look back and suggest we 'knew' these players would fail, each at least came with the promise of another successful season.

It is, however, reasonable to conclude that in all likelihood you need to be using a substantial portion of your money in the middle of the park, given the relative reliability of the pricier options. This season, no single player tops 10.0m but we have no less than 12 midfielders costing between 9.0 and 10.0m, making it tricky to differentiate between them all. For now though, we can just conclude that a substantial number of them are likely to deliver the best value for money among the elite.

Last season was somewhat odd for forwards, with just five players topping 150 points but then three of those topping 200 (and van Persie topping 250!). P Cisse's ridiculous second half obviously stands out on a value basis, but as we will likely see in the upcoming shots vs goals analysis, it is unlikely that he can continue that kind of form, all else being equal.

One point of note is that the forwards are much closer to the line of best fit, suggesting that a reliable trend might be expected (not necessarily a strong correlation, but a consistent one nonetheless). It's also worth noting though that half of the eight forwards valued at 9.5m or more had poor seasons suggesting a lack of consistency not always associated with the league's elite hitmen (Bent in particular is someone generally considered to be untouchably consistent).

The strategy for spending money on your forwards is probably least clear among all the positions as the top performers in terms of P90 were mid-range picks who came at a price where many of their colleagues failed. Van Persie was generally considered to be a must own player for the majority of last season, but even with that historic season his value was only just above the average line. This shows how hard it is to deliver value at 12.0m and this should be a word of caution for those looking at the Arsenal forward again this year (now priced at 13.0m). As noted in the dubious analogy in the opening paragraphs, sometimes you need to chase points over value, but it's always worth remembering what you need in return to justify your lofty investments.

The above analysis isn't going to win any leagues or give you any sure things for the upcoming year, but I like to start at the bottom and make sure we know what it is we are looking for and remember that good looking prospects can sometimes fail. Oh, and it's great to be back.


job kimani said...

very interesting analysis...unfortunately it seems hard to get the real insight that would enable a manager get the players with most value, hence jump ahead of the curve.

Genevapics said...

1.) Use the "alt" key + "3/# key and that will produce the £ symbol. Using the "alt" key produces and extra set of characters/symbols from your keyboard.

2.) As always, wonderful analysis. I like the idea of rotating GK's and I do not always go for a pair that rotate home/away in sync. I sometimes go with a pair that allows me to choose the better match up, even if they are both on the road.

Podge Marley said...

Lovely stuff, keep it coming.
Looking forward to any insight you have on spending in positions.
Funny to see Bent, Carroll and Torres being some of the worst value last year are being touted strongly again this year. We'll see if they earn their high values this year.

Bazil said...

Great stuff as always. I'm wondering how you factor in the value of captaincy for premium players though? Clearly big players at big teams are liable to hit big scores at home to the minnows, whereas value players at the smaller teams are often hard to predict in terms of captaincy.

For example, you say that Van Persie at 13m is very tough to back, but surely we must somehow factor in the high expected value of being able to pick him as captain vs the weaker sides, when he was knocking out 10 point + game weeks for fun last season, compared to our inability to captain many value players due to the higher unpredictability of there returns.

doka93 said...

Great article as always.

Points vs Price has and always will be the biggest problem facing FPL managers. All in all I've found that my team seems to strike a decent balance of value which has enabled me to shift in some big hitters.

There's just one dilemma facing me however. With £13m to spend on a defender and midfielder with the view of at least one playing for an elite side, what would people recommend as the better choice.

a) Lescott and Osman
b) Williams and Y. Toure