Value in FPL – 2019/20 Report

*As with Talisman Theory Part One and Peter Blake’s brilliant article on predictive analysis, this is a long form article*

We all love value.

Increasingly, it’s a metric in the spotlight.

There are a plethora of views on how player value can be represented: this might be through points per game, or an analysis of the data.

For us, by ‘value for money’ (or simply ‘value’) we mean player performance in FPL proportionate to their price.

We’ve produced numerous articles on the topic.

For example, last year we looked into value as a concept and also employed this thinking to understand how new prices impacted value metrics.

I’ve also written a (happily) well received THREAD (in the style of Neil Murray) on this subject:

Another way of expressing this is that value is long term measure of player worth. 

As such, it’s very useful in situations such as when you’re choosing your Gameweek 1 squad, or as part of transfer decision making during the season. Luckily, this article encompasses both.

In the evaluation section, I’ll give some reasons for why this is useful – and how to weigh this and the other long piece on Talisman Theory in making your Gameweek 1 decisions.

In the absence of creativity once more on my part, this article will replicate both of last year’s analyses of value in one piece, and also add some additional areas of inquiry to enhance our findings further.

The Data

As with Talisman, we scraped the FPL API and had our way with it.

We break value analysis down into 4 steps:

1/ Decide your data universe

2/ Compare last year’s final price and final points with this year’s initial pricing and last year’s final points by position

3/ Interrogate how we could make this even better by tweaking

4/ Team by team analysis

But before we dive in:

UNIVERSAL CAVEATS:

  • There are so many ways you can approach this complex, vexed and often differently perceived subject.  I’ve tried to give a fair representation to everything and preemptively answer any questions I can think of.
  • Value is relative. I understand that, Tim. If you have a player for just one Gameweek and he scores 1 point, compared to someone who held for all the highs and lows, the player may have different value to you compared to the general view. The solution (as always) is to engage brain to apply the findings to your own context.
  • Throughout, “Basic” is used as a noun rather than as an adjective. 

Right, let’s go.

Step 1: Data Universe

Alright, straight up – this is the least interesting step but also very important as we need to know what we’re actually analysing. Bear with, I’ll make it quick…

The first step in value analysis anyone can do and, without intending to disparage the efforts of others, this’ll be where some stabs at this analysis will end up.

This basic approach involves a simple equation:

Player Points / Price = Value.

We’ve also used this equation ourselves on occasions where time is short and we can’t get to our preferred data set.

But that equation throws up two questions:

  1. Do we look at player points in total?
  2. Do we look at Starting Price or Exit Price for historic data?

Usually, the answers are as follows:

  1. Yes, we’ll use it all
  2. Starting Price for historic data

It’ll come as no surprise that I dislike both of these answers.

From a researcher’s perspective, I’m never satisfied with an unfiltered, mass data set.

This is because it’s best to optimise your data sample before you analyse it – otherwise known as “data cleaning” – if you can.

Doing this allows you to home in on what’s really important, and reduce the impact of “noise” on your eventual conclusions.

For me, appearance points are “noise”.

Why I don’t use appearance points

I don’t think there’s much merit in including appearance in our analysis – for me, it’s a hygiene factor that a player is starting games. 

Encouragingly, it’s not just WGTA ploughing a lone furrow in this regard.

Peter Blake says in his piece that:

There is one metric that is universal in FPL and that is minutes played. The points accumulated from appearances (1 point for up to 60 mins; 2 points for over 60 mins) have been removed from the analysis to keep the view of the on-pitch actions as clean as possible. Therefore, this analysis refers to ‘adjusted points,’ which is a direct reference to the removal of appearance points.

(my emphases)

To add to this, I think appearance isn’t what we’re interested in looking at in terms of value analysis either: though appearance is important for obvious reasons, the points that are received for simply showing up aren’t earned, and gives you “noise” driven outcomes when analysing data the basic way.

Example of a nonsensical outcome from the basic method

The classic case we use here is someone like Jack Cork who typically plays every minute of the season.

He’ll get you 76 points by default, which skews the data.

Coady: last time he took a strike was when the Miners did (photocred: PL)

Last campaign, a good example of this was Wolves’ Conor Coady (4.5m last season), who played every one of the 3,420 minutes available and scored 94 points (despite not having a single shot!).

By the basic equation, Coady would get 94/4.5 = 20.8 as a value score.

By this metric, Coady comes out as better value than the likes of Wilfried Zaha and Mohamed Salah.

This nonsensical outcome reflects the inadequacies in the basic analysis method.

Of course, this is a very specific example chosen for effect, but nonetheless I’ve done the work behind it (promise) to check and concluded that this doesn’t show you anything – except the fact that this analysis isn’t effective.

So, points excluding appearance is our way forward. Peter calls the remaining number “adjusted points” in his piece; we refer to them as “non-appearance” points here, but it’s exactly the same thing.

Before you raise your pitchfork, I have made provisions to include appearance as a factor in how we understand value that I’ll discuss later, but appearance points in of themselves aren’t of interest to me.

NB: this analysis slightly differs from Talisman Theory in that “group effort” points such as clean sheets are included in value – it’s only appearance points removed, whereas for Talisman it was solely focused on attacking end product

Why I don’t use Start Price for historic data

I don’t think using last season’s start price to assess player value through their historic data is a good approach.

This is because (and this is slightly more qualitative) the Exit Price is probably a better feel for the true “end-value” of a player on the prices that previously existed, if we agree that value and price diverge over the course of the season (which I think they do):

Obviously, it’s not perfect, but my argument here is that Exit Price feels a better reflection of a player’s value in terms of how they performed during the season, rather than the prices we’re unilaterally handed by FPL Towers when the game relaunches every summer.

This means I will be taking Exit Price (i.e. Gameweek 38 price points) in any analysis of 2018/19, alongside the new prices for 2019/20.

Please revisit the universal caveats at this point if triggered.

I’m not going to spend any time on charting the basic analysis.

Let’s move on, though, by summarising this bit:

  1. Do we look at player points in total? WGTA: No, remove appearance.
  2. Do we look at Starting Price or Exit Price for historic data? WGTA: Exit Price.

This means our qualified equation for working out value is:

Player Points (Excluding Appearance) / Exit Price (if assessing 2018/19) OR Initial Price (if assessing 2019/20) = Value.

Now that’s sorted and you’re no doubt thoroughly convinced by my arguments (no, Tim, leave it!!), we’ll move to the fun part – analysis:

Step 2 : Old and New Prices versus Points – Value Per Position

Coolio, time for some of our designer friend Matt Brewer’s charts now, as seen on Talisman Theory (if you like his work and want to hire him, his behance is here).

Let’s use his lovely work to illustrate some of the data we create through applying the above equation to old and new prices.

We’re also indebted once more to Mitchell Stirling for his expertise in filtering the data.

First up, it’s the Goalkeepers.

2018/19 price on the left, 2019/20 price on the right – n.b. you can enlarge this image without leaving the page:

Some notes on this:

  • Jordan Pickford emerged as the top value keeper last season, as his 161 points from an Exit Price of 4.9m (after 14 clean sheets and 94 saves) saw him hurtle to the top of our rankings – however, the price increase to 5.5m has moved him down to joint 2nd by this year’s pricing.
  • Instead, Alisson emerges as the value victor with this year’s prices included, shading Pickford and fellow Brazilian custodian Ederson. Between the Brazilians, both priced at 6.0, it’s fine margins: 76 Alisson v 58 Ederson saves (Ederson has had the same number of saves two years running, by the way) and 9 bonus v 6 seems to have swayed it the Liverpool No1’s way, as he wins out by just 7 non-appearance points (100 to 93).
  • However, it’s worth noting that Ederson has kept a clean sheet on an astonishing 49% of games he’s appeared in (36/74, discounting the two games at the end of 2017/18 he missed) since rocking up at City in 2017/18. It remains to be seen if Alisson can keep the pace with him, and it also counts against the Liverpool keeper that the Reds’ defenders are so highly coveted whereas rotation fears in the City defence may mean Ederson is the port in the storm as the best way into their backline.
  • Hugo Lloris and Kepa at 5.5m may still provide decent value, as their unchanged pricing still looks to be about right in terms of what we’d expect in paying for a premium goalkeeper.
  • Lukasz Fabianski’s 5.0m price may not be so mad after all. Despite a 0.5m elevation from last year’s start price, the Hammers keeper still finished last season with the most saves (148, taking the mantle for the second year in a row) and 143 points – despite the Claret and Blue defence only recording 7 shut outs. Flappy is an excellent goalkeeper, it must be said, shown by how he prevented a chart-topping 10.9 expected goals last season – 3.6 more than his nearest competitor Lloris.
  • Martin Dubravka, Ben Foster, Kasper Schmeichel and Rui Patricio look much of a muchness – all priced at 5.0, all with OK numbers. Any of them will do an OK job for you.
  • If anything, the key here may be finding the next Neil Etheridge – the Cardiff stopper ended the season on 154 points, which saw him pip premium rivals Alisson and Ederson to second on last year’s Exit Prices. Of course, he doesn’t appear in this year’s ratings, but imagining Cardiff stayed up and he got the 5.0 price, he’d be second to Alisson on 15.6. In past years, the likes of Tom Heaton and Nick Pope have helped us by exceeding their initial 4.5m price tags – could either do it again this year, now back at that bargain bucket price?

Next, let’s move on to the Defenders:

2018/19 price on the left, 2019/20 price on the right – n.b. you can enlarge this image without leaving the page:

Lots to talk about:

  • Despite price rises across the board, inflation doesn’t diminish value as much for the defenders as it does for other positions. 
  • It’s unsurprising hegemony for the Liverpool boys, with the ridiculous-in-hindsight initial price of just 5.0 – leading to a 5.8 Exit Price – seeing Trent Alexander-Arnold firmly ensconced in top spot for last season. Looking at that, it’s all the more testament to the dynamic consistency of Andrew Robertson (Exit Price: 6.7) to remain on his coattails thanks to a 141 non-appearance point haul. TAA scored slightly fewer non-appearance points – 129 – but wins out on value due to his Exit Price.
  • However, Virgil Van Dijk ascends to the top of the defender value charts with the new pricing: VVD being 0.5m lower priced than the other duo means the Dutch destroyer sits top of the value charts for defenders. And for all players, in fact. Hold that thought for now.
  • Aymeric Laporte’s slight pump to 6.5m sees him stand still, but he remains good value.
  • Lucas Digne’s receipt of a 6.0 price tag actually sees his value slightly diluted as he drops from 5th to 7th. I really rate Digne, but this is not the forum for it. Hear our A-L preview podcast or read this top, top piece by FPL Chef for more info.
  • Chelsea’s frizzy grizzled centre-back and long ball extraordinaire David Luiz at 6.0 is the final occupant of a top five berth on this year’s prices. Chelsea’s defence has actually been pretty mean in recent years, recording 16 clean sheets on average across the last three seasons. Funnily enough, Luiz was outscored in terms of non-appearance points by Marcos Alonso, who recorded 99 points to the Brazilian’s 92; however, most of those points were scored during the RAM period and his 6.5 price tag has counted against him by this metric. Fun fact: 11 of the 17 chances David Luiz created last campaign were Big Chances (65%).
  • Matt Doherty’s heavy handed price rise to 6.0 sees him drop out of the top 10; instead, by this year’s prices it’s another Chelsea stalwart in Cesar Azpilicueta (82 non-appearance points) who appears in the 2019/20 chart at the Doc’s expense.

Next, it’s midfielders:

2018/19 price on the left, 2019/20 price on the right – n.b. you can enlarge this image without leaving the page:

Hot diggity!

  • Mo Salah is good value. A mammoth 184 personal points, plus a downgrade in price to 12.5m, combine as factors to see him move above Ryan Fraser. The Egyptian King is a monster on both the Value and Talisman metrics, showing what a perfect FPL asset he is, offering huge points potential from his Talismanic status and providing value whilst doing so.
  • Let’s talk about Weeman a bit, though, with Fraser proving absolutely absurd value last season at the 5.5m start price. A 2.0 hike for the Cherry sees his value diminish a little, but he’s still in the top 2, which shows that that price rise still doesn’t dampen his appeal on this front. I had almost forgotten about Fraser in the hubbub surrounding players such as Ayoze Perez; however, a player who both created more Big Chances than anyone else last season and also provides goal threat should not be underestimated, especially with two newly promoted sides first up.
  • What we do see overall is that inflation diminishes value for midfielders much more than it did for defenders: the likes of Richarlison drop out of the top 10, as a 1.5m price rise from his 153 point haul last season sees him drop to 14th here; the likes of Felipe Anderson (no price rise) and Pogba (modest 0.5m rise) enter the top 10 almost by default. The departures of Eden Hazard from the Premier League and Gerard Deulofeu from the midfield category also has an impact.
  • Gylfi Sigurdsson’s relatively modest 0.5m price rise sees him take a slightly different path from his Brazilian team mate and retain his place in the top 10 midfield value rankings; I’ve long banged on about Siggy, who scored his most points ever (including heroics at Swansea) last season and made the FPL team of the year. I’m glad to see the value data matches my expectations.
  • Leroy Sane and Heung-Min Son both retain their places in the top 10; Sane receives no change in price, and the South Korean takes a 1.0 jump. What’s interesting about both of these is that they are quite situational picks who are explosive. Our friend James at Planet FPL Podcast, for instance, has been rightly highlighting Sane’s potential since the prices were announced, pointing out if the German winger had played in more games last season he could have smashed the 300 point threshold like Salah did 2 years ago at the rate he was producing. Son is a similar kind of case, playing 22.5 games in terms of mins played/90, meaning he would have hauled ~265 points if he had played all 38 games at the rate he was returning. Both could be great dark horses in the value stakes.

Finally, the Forwards:

Placeholder for 2018/19 data as chart needs updating….

2018/19 price on the left, 2019/20 price on the right – n.b. you can enlarge this image without leaving the page.

  • It’s all about the mid priced forwards, as Raul Jimenez (7.5m), Callum Wilson (8.0m) and reclassified 6.5ms Jota and Gerard Deulofeu find themselves in the top 5.
  • Wilson again looks excellent value – though his starting price hike means that Jimenez wins out among the strikers.
  • Watford’s Gerry D does see a slight drop-off in value with the reclassification; throughout this analysis, we have refrained from recalculating the points of newly reclassified players but Nick will do something on this in the near future (tbc – thread or pod!)
  • What I notice looking at this comparatively with other positions is that the forwards see little depreciation in value due to inflation; instead their fairly consistent pricing (most of the forwards in situ last season received similar prices this) means that they remain fairly static.
  • Some of those who were the same price to start with, such as Josh King, see slight drops due to the fact we focused on Exit Price.
  • Staying with King, the difference between him and Wilson isn’t too big given their 1.5m price difference; 2.5 points per million difference is still appreciable, but means there is a case on the value front to consider the beta King over alpha Wilson if you want to make upgrades elsewhere.
  • Talisman hero Jamie Vardy is nowhere near as standout here compared to fellow hero Jimmy – at 9.0m, he may offer Talismanic appeal but on Value his appeal is slightly muted.

OK so… interesting. We’ve got a good look at value here.

Most people would stop there.

But like Chris Tarrant used to say on Who Wants To Be A Millionaire… 

We don’t want to give you that!

Step 3 – Tweak

OK, so I’ve rubbished appearance points in terms of analysis.

But appearance as a concept is still important; I just don’t think including it in data that we’re analysing is useful.

The fact that a player who is on the field for 59 mins gets rewarded the same as someone who shows up for a 5 second cameo should tell you what you need to know.

However, I wouldn’t pretend that appearance as a factor doesn’t exist.

After all, it goes without saying that being more likely to start is pretty important (but in this day and age you sadly have to).

But I didn’t want appearance points to be part of the value equation because it’s just “noise” in the actual data we analyse.

So I’ve done this another way which is, after excluding them for Step Two, including a modifier in Step Three to reflect appearance, which gives us a final value metric.

The cleanest way to do this was to create a modifier multiplier by dividing a player’s minutes played by the 3420 total available minutes (i.e. 90*38).

So, if a player played 2565 minutes, that works out at them playing 75% of the available game time. That means that whatever score they got by the method discussed in Step Two would be multiplied by 0.75 to reflect appearance in the value metric.

Yes of course there’s extra time etc. But I’m going to go with the minutes OFPL said they played and approach this as if games were 90 mins long – it’s the best approximation I have and a couple of extra mins here and there won’t change the story.

The intent behind this is to calibrate our data so we end up with a finesse data set.

Example for why this is important

Photocred: PL

A good example of how this impacts the data is very clear in the likes of Mo Salah.

Salah scored 259 points total last season, of which 184 were non-appearance points.

Let’s use all three ways of calculating value to illustrate why I think our method is optimal.

1/ Using the basic calculation  (i.e. all points including appearance/new price = value), Salah was the 52nd most valuable player in FPL last season. *insert shrug emoji*

2/ Taking out appearance to judge the actual value he added above stepping over the white line (non-appearance points/new price = value), Salah rises to 14th most valuable.

That looks OK – he is the most expensive player in the game after all at 12.5m.

However, the fact that Salah played 95% of the available minutes needs to be represented in how we complete the analysis in the final step:

3/ Multiply his output to Step Two by 0.95 (to reflect 3254 mins played / 3420 available) to get to the final stage, which puts Salah 8th overall.

I think that makes sense in the wider sense of our investigation of value: with all factors included, Salah is the top non-defender in our final data, which reflects the value that he represents to any FPL manager.

Top 20s – modifier

Unfortunately, we exhausted our time with the designer at this point, so I’ve had to just stick to top 20s. 

Here’s how this creates the top 20s – both in 2018/19 and 2019/20 prices:

2018/19 price on the left, 2019/20 price on the right – n.b. you can enlarge this image without leaving the page:

Key points here:

  • Defensive assets reign supreme. 
  • 6 keepers and 8 defenders make up the top 20 in the 2019/20 value rankings, compared to 5 midfielders and 1, lone, striker in Jimenez.
  • This is actually fairly similar year on year – the key impact comparing them is merely to see how far price changes for TAA and Robertson have pushed VVD to the top of the rankings almost by default. I’ll expand on this more in the podium analysis below.
  • Mo Salah is amazing – he stands tall in both Talisman and Value as a serious must-own based on both of these measures.
  • TAA is let down by his absences at times, as his *0.72 modifier, plus a 2.0m price rise from last season’s paltry 5.0m start price, sees him fall slightly down the rankings
  • Jimenez’s talismanic performances, plus his cheap price, means that he is the only forward in the top 20.
  • For those wondering where he is, Raheem Sterling doesn’t make it into the top 20: with a score of 11.41, he finds himself in 24th overall.

So, yeah, value is all about the defensive players.

Here’s the top 10s per positions, as well:

n.b. you can enlarge this image without leaving the page:

Check out the variances in the conditional formatting (colour scale) – the top 10 forwards cannot compete with the defensive assets.

  • For those who will ask, Harry Kane does not make the top 10 here for forwards: he’s brought down by a low modifier (0.7) and less points per value than the other forwards who do make it.

The universal finding I don’t think is new news to anyone, and in some ways it’s just confirmed what we’ve been arguing from the beginning of WGTA: Big At The Back Is The Best Value.

FPL has changed because we’re now far more receptive to Big At The Back as a concept, which has become more recognised as a value-driven way to help think about your squad structure.

It was simply unheard of two to three seasons ago to spend big in the defence: we used to pour all of our money into the big frontline.

A final thing to look at is inter-team comparisons of value, which I’ve termed a “Podium Analysis” in the absence of thinking of something smarter.

Step 4: Podium Analysis

As with Sloppy Seconds in the Talisman Analysis, let’s have a look, per team, at value & modifier numbers.

This might be helpful in answering a few intra-team disputes further, as well as shedding light on what’s going on beneath the upper echelons of value.

n.b. you can enlarge this image without leaving the page:

As you can see, I’ve shown the 1st, 2nd and 3rd ranking players per team in terms of value and shown the differences between these rankings as percentages.

This may be helpful to you in deciding who to buy from each side using a value perspective as you make your Gameweek 1 choices.

Some notes:

  • This should reinforce to you 1/ what good value defenders are and 2/ just how much value the Liverpool defenders in particular are. VVD’s smaller price rise of 0.5m from last year’s starting price compared to 1.0m for Robertson and 2.0m for TAA means that the difference between the Dutchman and TAA in 3rd is the highest between players from any club.
  • Fraser is the best value Bournemouth asset, followed by Wilson. What is interesting here is the different dimension this adds to the Wilson v King dilemma also discussed in Talisman Theory; there, Wilson won out comfortably, but in terms of this analysis its interesting to see that, though the Englishman still out-values King, it’s actually not by much using the multiplier condition. A 0.47 difference in terms of value might show that there is an argument to get King – especially if the 1.5m you can save there helps you turn a sub-par option into a good one elsewhere.
  • Without Eden Hazard, the Chelsea team really looks to be all about that mean defence. David Luiz, Azpi and Marcos Alonso are the top 3 assets there for a defence that has averaged 16 clean sheets per season over the last 3 years. Food for thought.
  • Similarly, the focus for City is also all about the defence, with the forwards supplanted by the power of Ederson, Laporte and Walker. Again, this might help you with your Gameweek 1 decision making.
  • Aaron Wan Bissaka‘s move to Man United looks a good one – the young right back, despite being 5.5m, finds himself in second in the value rankings for them, and could still provide a decent choice from this perspective if we are to need a Red Devils defensive asset
  • On movers, I was similarly impressed by Ayoze Perez clambering onto the podium for Leicester following his move to the Foxes, beating out James Maddison; though “Maddy” may provide interest through his chance creation numbers (top overall last season with 99) that 0.5m saving to buy Perez may look a viable pick if you base it on value. Nonetheless, there’s acclimatisation and also the signing of Youri Tielemans to bear in mind as contextual factors
  • Spurs had a tough year last season – it’s hard to assert that any of their players offered value apart from Lloris. Harry Kane makes it on to the podium despite a stop-start season by dint of Kieran Trippier’s departure to Atleti which says it all really.

Evaluation

Caveats first:

  • There are lots of ways to analyse this kind of data. You may disagree, and have a different approach. I make no apologies for showing what I believe to be a better way to do this value analysis, but this shouldn’t be taken as a slight on anyone who decides to take the basic path.
  • Context is key. Signing VVD, Robbo and TAA over a Salah because of value analysis alone is probably not optimal as Salah is a standout captaincy option for Gameweek 1. I know that, you know that. But this analysis remains valid, and if you apply brain we’re not saying that.

…Right, so what have we learnt?

Pricing significantly correlates with performance to a huge degree across the players currently in the game. I say this because inflation doesn’t radically alter the complexion of the rankings. Reassuringly, this shows on a quantitative level that prices are mostly set by objective means, rather than human ones.

Big up the defensive assets – the goalkeepers and defenders are the place to look if you’re looking to optimise value for money. They dominate the value rankings, with 14 of the top 20 value players being defensive. I think what’s important, as per the caveat, is to recognise that defenders are really powerful FPL assets – but week to week you might find that a more offensive outlook has a higher points potential.

  • As an example of this, although VVD is the Value King, I’m favouring going with TAA and Robbo to start with. This is because their week to week ceiling is higher, whereas VVD will provide you a constant drip feed throughout the season (plus some flukey assists!). The fact that it’s easy to go down 0.5m to VVD from them also compounds this.

Mo Salah is a monster. Both this and Talisman analysis have shown how the Egyptian King, despite being the most expensive player in the game, has both the influence on his team and the output to provide value. The fact he’s the only non-defender in the top 10 only underlines what we already know: he’s an elite FPL asset who really should be in your team most, if not all, of the time. Sadio Mane does appear in the top 20, though, but might need some time to rest prior to GW1 having played a 356 day season!

City haven’t really starred in this analysis. The fact that Raheem Sterling and Kun Aguero have barely been mentioned in this value analysis shows two things:

  1. The City defence is slightly underappreciated. Plainly, the consistency of clean sheets produced by City – if the opposition don’t have the ball, they can’t score – is really notable in how it’s overlooked in favour of their embarrassment of riches going forward. Nonetheless, it’s that defence, quietly grinding out the points, that could be the best value of all – Ederson truly is “steady Eddie” in that he will get you 6 points basically half the time, which is a really great proposition.
  2. But remember this analysis provides part of, not the full, picture. I’d never say “go without one of the main City boys” because their obvious merits are a captaincy option and explosive potential. Yes, they aren’t really value for money if we’re looking at FPL through this lens, but this analysis isn’t seeking to be a panacea to answer all of FPL’s questions.

Raul Jimenez now needs to produce the difficult second album. Now 7.5m and with the Europa League to think about, can “Jimmy” – who smashed both the Talisman data and was the lone forward in the top 20 for value – do it again? If you were to solely base your choices on analysis such as this, Jimmy is a shoo-in. However, it’s actually a tough judgement, with Wolves’ small squad surely to be impacted by the European involvement plus the links with a litany of new strikers (such as Patrick Cutrone) which may eat into Jimmy’s game time. This isn’t to take away from Jimmy’s achievement, and I’m sure many will own the Mexican come Gameweek 1. It’ll remain to be seen if he can hit the heights of last season.

Ryan Fraser is the value pocket rocket. One thing that really stood out to me doing the analysis is how Fraser, much like Vardy in Talisman, emerged as a “beta story” to the defenders (assuming that we take Salah’s ridiculous data as a given). The wee Scotsman is an absolute gem looking at this analysis, with last season proving to be a big breakthrough for him. Could he actually be a good “middle ground” between owning Wilson or King?

  • King v Wilson can also be helped with a combination of Talisman Theory and Value: Talisman (and the stats) point to Wilson being the best option, but King isn’t actually too far off the Englishman in terms of value – as per the Podium Analysis, barely half a point per million separates them. There’s a perfectly valid option to consider either depending on your situation, though the fact that 50% of King’s output last season came from dead balls (such as pens) might be something to think about.
  • Could doubling up on Fraser (Value) and Wilson (Talisman) be a good calculated risk?

Value v Talisman – which is more important?

What’s really interesting here is comparing Value with Talisman Theory, especially in terms of how we should weigh each when making judgements on our Gameweek 1 sides.

I would say that Value is more important than Talisman to start with.

Because of the individual nature of the focus of Talisman Theory, it’s well suited to deciding who to prioritise for Gameweek 1 and who might be a good transfer each week during the season proper.

However, it’s less useful in terms of helping us make trade-offs ahead of finalising our Gameweek 1 side.

Like this value analysis, Talisman Theory was a retrospective analysis, but the point of that was how to use that knowledge going forward – we don’t have the key form-based information yet as the season has not kicked off.

In contrast, value helps inform our judgements using a large cohort of data to discern which choices are better because we do have the relevant data – pricing – and last season’s final scores that the new prices seem to be based on.

Conclusion

So there you have it. All hail the Defence, Mo, Fraser and Jimmy.

I know, I know. Some people may dislike that totally and think it’s boring that defenders are the value men. I completely see where that opinion stems from – but because you feel that way doesn’t change the findings.

In this vein, it’s maybe a case of weighing up being analytical against the need to have fun with the game that is FPL: could having a solid, value-based spine with a couple of Talismanic/calculated risk picks peppered in be a good way forward?

A good finding is that this echoes what we said on our Unwritten Rules podcast with Neil Murray. In that, during our discussion around the first rule – The Game Has Evolved, Change With It – we said that it might perhaps be time to step away from the importance of positions and instead view players by price point. 

This feeds in to the Big At The Back topic, and I think a subtle shift has begun through the fact setting up with Premium defenders in the shape of Robertson and TAA seems fairly common now.

Of course, captaincy is one thing I haven’t really touched on. This is because it’s something that is hard to evaluate in the data, but it goes without saying that value is amplified by putting your armband on a player.

As we progress through the season, we’ll see how this develops: team performances start to see some players start to make a mockery of their initial price – which helps us spot value. The early season tends to be an exercise in finding value players, adding grist to many a social media conversation and inspiration for analyses such as this one.

Identifying those players and investing early is what we all want to do.

Observing the trends at the end of the season (and showing awareness of the limitations of season-long analyses) is key to helping interpret value beyond the metrics – remember that this is near enough objective value measurement, but you apply your interpretation of value when choosing your team.

Value is in the eye of the beholder.

Many thanks to Mitchell Sterling for his data expertise, Matt Brewer for design work, Tim Shelbourne for inspiring caveats, and ConAdam, FPL Chef, Goops, Alex Ball and Neil Murray for proof reading.