Imagine for a moment it’s May and you’re the newly appointed director of football at a bottom half of the table Premier League football club. In fact, your club just barely escaped by the skin of its teeth this season, finishing in 17th.
You’ve inherited a team with a traditional scouting network, one in-house performance analyst, plus a stats-minded intern. While the slightly batty manager is genuinely considered adequate—if not stellar—the squad itself is both aging and shallow. Meanwhile, the one or two genuine stars you do have are outside the protected period in their contracts are are likely to move on in the summer. Though the first team is almost certainly in need of a radical overhaul, ownership has provided you a maximum transfer budget of £35 million to work with.
After questioning why you took this job in the first place, you sit down and have a think. What kind of strategy should you pursue in the summer transfer market?
Two Transfer Strategies
One approach might be to try to spread the money around evenly, casting a wide net as possible to help add depth to the squad in one fell swoop. This, of course, greatly limits the maximum transfer fee you can pay for any one target. That generally leaves you with two target groups: some older (> 32) still decentish players whom you might have a good season left in them and can help you in another relegation fight, and some potentially decent younger prospects (< 21) toiling in smaller clubs on the continent who could develop into major talents later on.
A good analyst—in this case brought on as a consultant—can obviously assist by eliminating a few definite “no-no” in whatever list you come up with. If they’re really good, they may know of one or two players in the prime of their career who are wildly undervalued for whatever reason (unicorns, basically). Though multi player spending sprees are time consuming, you may have to do a lot less wheeling and dealing—this lower end of the market will be more amenable to joining a bottom-dwelling Premier League side.
But this is a very risky approach. Your club isn’t Leicester City; it has no scouting process, no communication or analytics buy in from the manager, and you’re a brand new DoF. Now you’re coming in and in one summer are attempting to dramatically makeover the first team by investing in young players and a handful of veterans in the twilight of their careers. You risk replacing one mediocre team with another.
So you contemplate taking a different tack—instead of going for volume, you could spend the vast majority of your transfer budget on one or two expensive but ‘safer bets,’ and then use whatever tiny amount you have left—if anything—on some high risk/big reward prospects. This basically is the ‘barbell approach’ touted by Taleb:
“If you know that you are vulnerable to prediction errors, and […] accept that most “risk measures” are flawed, then your strategy is to be as hyperconservative and hyperaggressive as you can be instead of being mildly aggressive or conservative.”
This strategy comes with different drawbacks; first, £35 million in today’s football economy might allow you to buy somewhere in the “middle market” of first division players, where there is still a lot of noise.
Second, you’re not spreading the risk around—if your ‘big name’ players misfire, not only have you tied up a huge portion of the club’s amortized fees on a few duds that you will likely sell later at a loss, but your club will almost certainly face the drop barring some good luck and clever management. Uncertainty is an issue for even the safest bet in football—a club could pay Lionel Messi’s transfer fee only to have him break a leg in training.
Third, you may have to pay more for certain players than a big club because, well, you’re not a big club. Nevertheless, you could use the lure of guaranteed first team football and being a big fish in a small pond as a selling point.
Again, a good quant could help you with all this. Because this kind of strategy is far more focused, I personally would want an analytics consultant who has not only developed very good predictive metrics, but is also a genuinely good forecaster, someone who can answer basic questions like:
- what is the base rate of success in terms of first team minutes played for this type of player coming from this type of club in this position and at this age?
- What factors, be they your own metrics or other measurements, move that base rate higher in this particular case?
- What is the range of possible impacts this player could make on points/goal differentials, from best to worst, and their range of probabilities (and yes, they need to add up to 100)?
- What is the difference between the market price for this player and the price you would pay for him in a world of perfect information?
And you wouldn’t want this kind of assessment just for your preferred, ideal signings, but for an extensive range of realistic transfer targets at the same price point at different clubs (or, in the same club as your primary targets in case they’re offered up as a consolation).
Incidentally, because your consultant will be doing this work in a very short period of time, you will probably have to pay them a lot of money—see Ted Knutson’s little blurb on that from this past week at Statsbomb.
Keep in mind too—all of this legwork will come before you even pick up the phone and deal with other owners, directors, players, players’ agents, families, entourages etc. etc, things that can wreak havoc on your mathematically precise, perfectly benchmarked transfer fee. Despite all your numbers and understanding of true market value, you and the owners will almost certainly have to come to an agreement in advance as to how much you’re willing to overpay for a particular transfer.
Monchi and Sevilla
One purpose of this little hypothetical is to yet again say hey—running a football club is pretty hard you guys!
Another is to dispel the myth that the primary role of a stats analyst in soccer is to mitigate or eliminate risk in the transfer market. This is a noble goal in some cases, but it’s not a realistic one. Risk is inherent to player recruitment. The transfer fee in part reflects the amount of risk buying clubs believe a player represents—better players command higher transfer fees because they’re more likely to have a positive impact wherever they go.
It’s for this reason that, except in the most obvious cases, I’m reticent to criticise clubs for any single failed signing. Often, the probability of that failure may have been built into the terms of the deal—they were a decent bet. What obviously matters is that the club has more hits than misses.
This point about risk was brought up in Sid Lowe’s marvelous interview this past week with Sevilla’s director of football Monchi, which was the inspiration for this week’s column. Monchi spoke to Lowe about the cardinal difference between English and Spanish scouting methods:
“There are loads of off-field things in which they beat us easily,” [Monchi] says. “And on the football side, I saw very good work being done. But there’s a disconnect between that work and the advantage they glean from it. I know English clubs that are very professional, scouts everywhere, but the information they gather isn’t always applied. Why? Because they have money. That enables them to take fewer risks: ‘I’m not going to discover Keita at Lens; let Sevilla do that and then buy Keita from Sevilla.’”
But because this is football and a lot of clubs routinely make errors of judgment, there is always the potential that teams will either misjudge that risk, either positively or negatively. This is where a smart analyst can find hidden value.
In response to Low’s column, at least one analyst on twitter wondered why more English clubs didn’t save themselves a whack of money by “discovering Keita at Lens” before Sevilla, perhaps by building a similarly extensive scouting network and taking a lot of prospective bets.
But as Mike Goodman and Nikos Overheul both wisely pointed out, Sevilla can take risks on these kinds of players because they’re safe from relegation. For Sevilla—who also happen to be a very well run club with good communication between the DoF and manager—it’s a low risk, high return strategy. Southampton are the obvious English club equivalent; their scouting network, coupled with one of the few sane managerial succession strategies in European football, allows them the peace of mind to take on some higher risk, higher reward bets.
In our hypothetical situation above, with obvious similarities to Aston Villa, going for broke on an a few potentially high impact, analyst vetted targets in the middle upper range of the market makes more sense in the short term. It comes with its own risks, yes, but if you have a new staff and and a new approach you want to put in place, in addition to a manager you inherited rather than hired, you will likely save more face if things go badly than if you try a more developmental approach too quickly.
But both come with their own risks, particularly for a club in dire straits like Hypothetical FC up there. The better option is not let your club end up in that position at all; that means investing in better scouting processes now, when things appear to be going relatively well.