Moneyball Soccernomics

“Moneyball” is one of those terms that can leave people feeling a little cold.

It’s been used, misused and overused to the point where it sometimes seems to mean little more than buy cheap players who might actually be good.

And historically, the principles behind Moneyball haven’t always transferred particularly well from baseball into football.

But strip away the spreadsheets, algorithms and Brad Pitt references, and the original idea is actually incredibly simple:

Find something the market is undervaluing… and exploit it.

That is exactly what I’m attempting to do in MFL.

Welcome to MFL Soccernomics.

Obligatory Brad Pitt photo

The problem with “Moneyball” in MFL

At first glance, MFL doesn’t seem particularly suited to this kind of approach.

We don’t have dozens of detailed player attributes. There aren’t huge datasets containing progressive passes, expected assists, pressures, carries into the final third or any of the other metrics increasingly used in real-world football recruitment.

MFL gives us six simple outfield attributes:

• Pace
• Shooting
• Passing
• Dribbling
• Defending
• Physical

And from those attributes, every player is given an Overall Rating (OVR).

OVR naturally becomes one of the first things we look at when assessing a player.

An 82-rated midfielder is better than an 80-rated midfielder.

Right?

Well…

Not necessarily.

And that’s where things get interesting.

What does an 80-rated midfielder actually mean?

Overall Rating isn’t magic.

As we all know, behind the scenes, MFL needs a way of converting those six attributes into a single number.

That means each position effectively has a predefined idea of what constitutes a good player.

For a central midfielder, for example, certain attributes will contribute more heavily towards OVR than others.

That makes complete sense.

The problem is that not every manager wants their central midfielder to do the same job.

Think about how many different types of central midfielder exist in football.

You could have a box-to-box midfielder expected to contribute at both ends of the pitch.

You could have a deep-lying playmaker whose primary responsibility is progressing the ball.

An advanced creator operating close to the forwards.

A ball-winning destroyer.

A dribbling midfielder who carries the ball through the centre of the pitch.

A mezzala-style player drifting into wider areas.

Or in real terms: Gerrard, Lampard, Pirlo, De Bruyne, Petit, Scholes, Keane…..

They’re all central midfielders. But the attributes required to make each of them exceptional are completely different.

And yet in MFL, they’re all being judged by the same positional OVR calculation.

That is where the market inefficiency lies.

Stop asking: “How good is this player?”

This is the fundamental change I’m trying to make with Mayhem’s recruitment.

Instead of asking:

How good is this player?

I’m asking:

How good is this player at doing exactly what I need him to do?

Those sound like similar questions.

They’re really not.

And answering the second question requires you to temporarily forget the large font number sitting next to the player’s name.

Building players for the system

At Madero Mayhem, I play a 4-3-3 D.

More importantly, I have a pretty clear idea of what I want each position within that formation to achieve.

My three midfielders aren’t simply:

CDM – CM – CM

They perform three different jobs.

The CDM is primarily there to protect the defence, win duels and provide a reliable base behind the other midfielders.

One CM operates as our deeper creator. Passing ability is therefore significantly more valuable to me in that position.

The other CM has a more progressive role. I want someone capable of passing the ball but also carrying it forward, making dribbling much more important.

Suddenly, I’m not shopping for three generic midfielders.

I’m shopping for three completely different players.

And that changes how I value the transfer market.

I’ve looked at EVERY position and the role I want it to carry out and created my own ratings (with a little AI assistance).

Introducing the Mayhem Rating

This is where my experiment begins.

Instead of accepting MFL’s positional OVR weighting as the definitive assessment of a player, I’m creating my own.

For every position in my 4-3-3, I’ve identified which attributes matter most for the specific job that player performs.

Those attributes are then weighted accordingly.

The result is a second player rating:

The Mayhem Rating.

The Mayhem Rating isn’t intended to tell me whether a player is objectively better than another player.

It tells me something far more useful:

How suitable is this player for Madero Mayhem?

And that’s an important distinction.

Imagine two midfielders on the market:

Player A — 82 OVR
Excellent all-round attributes and therefore highly rated by MFL.

Player B — 80 OVR
Noticeably weaker in a couple of areas, but exceptionally strong in passing, dribbling and with decent physicality – exactly the attributes my progressive CM role requires.

The transfer market is likely to value Player A more highly.

But what if my model gives them:

Player A — Mayhem Rating: 79

Player B — Mayhem Rating: 83

Now things become interesting.

I’m potentially buying the better player for my system while everyone else sees me buying the worse player.

That is Soccernomics.

The attributes I’m deliberately willing to sacrifice

This is perhaps the most important part of the whole concept.

Finding undervalued players isn’t only about identifying attributes you want.

It’s about identifying attributes you’re willing to ignore.

If shooting contributes towards the OVR of a midfielder but I rarely expect my deep creator to shoot, why should I pay for it?

If one of my defenders is primarily being selected to win the ball and dominate physically, do I care if his shooting attribute is terrible or his passing is bordering on acceptable?

If my winger’s role revolves around creating rather than scoring, should his market value to me really suffer because his finishing isn’t particularly impressive?

Every point of OVR being generated by an attribute I don’t need is potentially something I’m paying for without actually using.

Conversely, a player being punished by OVR because he’s poor at something I don’t require could represent value.

That’s exactly what I’m looking for.

This doesn’t mean OVR is wrong

There’s an important caveat here.

I’m not saying MFL’s Overall Rating system is bad.

In fact, as a general measurement of players, it makes perfect sense.

OVR has to work across thousands of players, teams, formations and tactical approaches.

The Mayhem Rating only needs to work for one.

Mine.

That’s actually its greatest advantage.

MFL has to answer:

How good is this player at this position?

My model only has to answer:

How useful is this player in this particular role within this particular tactical system?

That allows me to be far more ruthless with the attributes I value.

MFL Soccernomics in Action

As midfield is an excellent example of where to find these market inefficiencies, it makes sense to gives some examples of MFL Soccernomics, and how I applied the Mayhem Rating in my recruitment approach for Season 16.

Firstly, my favourite: Ivo Dierckx.

Signed for $11, he may not raise any eyebrows.

But, when his Mayhem Rating is applied, he’s rated at 74.1 MR as a Deep Creator.

All I want my Deep Creator to do is sit in the centre of the pitch, maintain possession and when possible, ping the ball forward into dangerous areas for the front three to attack. He doesn’t need to dribble past 2 or 3 players, or get pulled out of position by trying to get into the box to score 10 goals a season.

Now when we think of a 74 rated CM for $11, suddenly we have a bargain. An exploited market inefficiency. Especially when the cheapest 74 OVR, 28-year-old CM on the market at the time of writing is $62!

That’s $51 less — an 82% saving.

MFL sees a 71-rated central midfielder. The Mayhem Rating sees a specialist creator. Six games into the season, Ivo Dierckx has created 31 chances and registered five assists, leading the division. The market sees the number beside his name; Soccernomics asks whether we’re looking at the right number.

Another example of the Deep Creator role is Walter Falcone.

Falcone is a player I packed. I’d spent two seasons trying to get rid of him because I was judging him using exactly the same OVR-led thinking I’m now trying to exploit in other managers.

Look at him. I guarantee your initial thoughts were same as mine. Too slow, can’t shoot, can’t tackle.

However, whilst MFL ratings only rate him as 68 OVR, the Mayhem Ratings make him a 72.2 Deep Creator and a 70.2 Progressor.

He was swiftly taken off the transfer market.

Finally, let’s look at the defensive midfielder role.

The MFL weightings for a Defensive Midfielder value the following, in order: Defence (40%), Passing (28%), Dribbling (17%), Physical (15%), Pace (0%), Shooting (0%).

Therefore, Henrik Westman – our new CDM – is a 74 OVR.

However, in my Disciplina y Control system, all I want my CDM to do is win the ball back by being a bit of a bastard and pass it to someone else. Shield the back four, that’s it.

So, my weightings for the CDM role are this: Defence (40%), Physicals (25%), Passing (20%), Dribbling (8%), Pace (7%), Shooting (0%).

If I apply my weightings to the role, Westman has a 76 Mayhem Rating (MR).

Want to buy a 76 OVR, 28-year-old Defensive Midfielder on the market today? That’ll be $89. I bought Westman for $28.

A massive saving.

Seven games into Season 16, Westman’s output looks exactly like I’d want from the role: 23 clearances, 13 defensive duels won and just two chances created in 630 minutes. That last number isn’t a criticism – it’s almost the point. I haven’t recruited him to create chances. I’ve recruited him to win the ball, protect the defence and give it to someone who will.

Can YOU do Soccernomics?

Absolutely you can.

Obviously, you’ll play a different formation to me, with different positions and requirements from each role.

The chances are, that when you compare each role in your formation and compare it with the official MFL weightings, you’ll find at least one which differs.

And somewhere in that difference of opinion is where transfer-market value can be found.

Players who aren’t deemed good enough by someone else because their OVR rating is a little below what they deem fit for their squad.

I’ve given three examples of it in action in my agency, but there are more I could’ve used.

Here’s how I approached it:

1. What do I actually need this position to do?
2. Which attributes make that happen?
3. Which attributes am I paying for that I don’t need?
4. Which lower-OVR players are strong in the attributes I do need?
5. Is their price significantly below a conventional player of equivalent role-fit?

Feel free to reach out on Discord or X if you’d like to chat more about this.

What’s next in MFL Soccernomics?

I need to put my hands in the air and admit a few things:

1 – I’ve only used this system for one transfer window so far, and only for the last few players I bought.

2 – As the seasons go on, and my tactic changes, so will the weightings… I expect it to be a fluid process.

3 – We’ll very likely have a new match engine in Season 17 so I’ll have to adapt accordingly.

4 – I haven’t yet delved into the player stats section but it’s something I’m looking at as the methodology evolves.

But I’m loving this approach so far.

Each player I buy enters a tactical system specifically designed to use the things he’s good at while minimising the things he’s bad at.

And that means he’s got a better chance of performing like the rating my model said he was.

That’s the theory, anyway.

Because there’s one rather important question still to answer.

Does it actually work?

That’s what I’m going to find out. I’ve already given some stats when looking at Dierckx and Westman, but we’re not yet halfway through the season. The real proof will be at the end of this season.

I’m already using the Mayhem Rating to analyse my existing squads and scout potential signings.

If it works, it could fundamentally change the way I – and hopefully you – recruit players in MFL.

But I’m increasingly convinced there’s value hiding behind that OVR number.

We just need to stop looking at the number everyone else is looking at.