Talisman Theory – 2019/20 Report

Talisman Theory is the concept that you should always favour buying in the key individual – the Talisman – over their other team mates as they are the most likely source of FPL points.

That’s bloody obvious, you may think.

Well yes, yes it is.

However, unlike in answers heard to eternal FPL debates such as “should we favour form or fixtures?”, or received wisdom such as “you should always favour captaining the home player“, the difference with Talisman Theory is we can prove the concept, using evidence rather than anecdote. 

Anyway, over the last couple of years (2017/18; 2018/19) I’ve delved deep into the data and done the analysis to help bring Talisman Theory to life for FPL managers.

It was my pleasure to do so again this summer for the following report.

Just to warn you, this is a bit longer than your average weekly churn piece.

I make no apology for that, and hope that a mixture of intriguing data and my idiosyncratic penmanship will see you stick with me ’til the end.

The data

As usual, we scraped the FPL API for last season’s data and had our way with it to form the basis of our analysis.

Talisman analysis breaks down into 3 steps:

  1. Collate team points scored
  2. Find out who scored the most points within each team (“Lesser Talisman”)
  3. Remove non-individual points like appearance and clean sheets to generate a “Talisman Points” metric, which enables us to see who was truly the individual to own per side (“Greater Talisman”)

I also throw in a couple of extra analyses to cover off any further areas of interest, which I’ve christened Exit Velocity and Sloppy Seconds.

This year, Exit Velocity will focus on pre and post lockdown data in order to include the unique circumstances of this season. I won’t be breaking down everything down into pre and post lockdown, however, or it’ll just get too long.

Step 1: Team Points

The first step of our analysis is to check out team data to understand how many FPL points were scored per team. We’ve juxtaposed this with their final Premier League position in order to provide a point of evaluation:

[Link to IMAGE ~ GIF]

An intriguing observation at first glance is the total points scored:

An additional note is that City’s high score of 2100 points is the lowest high score over the past 3 years.

That would have placed 3rd last season.

Further observations on the data:

  • The top 4 scored 7727 points between them – 25% of all points in FPL.
  • City supplant last year’s top scorers Liverpool at the top to give Abu Dhabi something to cheer about but, as usual, it’s a close run thing – only 6 points separate the two. It was a similar story last year, when the Reds pipped the Citizens by a small (9 point) margin.
    • If anything, this illustrates the massive strides made by Klopp’s men: in 2017/18, City beat Liverpool (who finished 4th in the real PL table but 2nd again in the FPL numbers) by a stonking 172 points.
  • Looking at the numbers this way underlines how FPL is becoming a bit like the SPL in terms of the gap between the top two and the rest. A huge 311 point gap separates Liverpool in 2nd from Leicester in 3rd.
    • However, this gap is smaller than last year when 341 points separated City in 2nd and then 3rd place Chelsea.
  • Leicester are new entrants to the top 4, after scoring a massive 260 extra points more than last year’s total. Their early season prowess gave way to late season malaise of course, but they still did enough to clinch 3rd.
  • In a similar vein to the Foxes, Man Utd also turned it round big time after a poor prior campaign. In 2018/19, they scored a paltry 1544 points and were outscored by the likes of Crystal Palace. By contrast in 2019/20, driven by a 344 point restart haul, they managed to accue 206 points more on their way to finishing 4th in the rankings.
  • Elsewhere, it was certainly a year for stoic defences. Wolves, Sheffield United & Burnley occupy 6-8th in the table (Burnley were joint 8th with Spurs, which we’ll get on to/laugh at shortly), reflecting their defensive success in recording 13, 13 and 15 clean sheets respectively.
    • The 15 clean sheets for Burnley saw them =2 with Liverpool on clean sheets registered, just behind Man City’s 17. This defensive solidity contrasts with a Europa League energy sapped 8 from the season before, which meant the Clarets rocket up the rankings from 2018/19’s 17th place to 8th, after racking up 222 extra points.
  • Last year Spurs finished in the top 4 on 1828 points, meaning that in 2019/20 they scored an astonishing 284 points less than in the season before. Ouch.
  • This lightens the misery that a Gooner may feel about Arsenal’s lowly post here – though a 1535 and 10th place isn’t where we want to be, at least our year-on-year of decrease of 109 points less isn’t as bad as Tottenham’s. [Insert “clutching at straws.gif”].
  • Every year, we see around 10 teams fail to make the 1500 points mark and it was no different in 2019/20. However, these sorts of teams are often the source of a Talisman…
  • Bournemouth went down having scored 197 fewer points (ie ~28 Fraser/Wilson combinations) than they did the year before.
  • You may think LOL @ Norwich, but actually they’re not as bad as last year’s big losers Huddersfield (remember them?) who recorded just 1068 points – 55 less than what the Canaries managed.

Having collated this data, we move forward to the next step.

Step Two: Lesser Talisman

Next we move to a simple look at which player scored the most points from each team, which I call “Lesser Talisman”.

I refer to this as “Lesser Talisman” because I think it’s nothing more than a vanilla look at the Talisman data; it gives a good idea of things, yet could be taken further.

For now, here’s 2019/20’s player points / team points graphic:

[Link to IMAGE ~ GIF]

Summary notes;

  • Colour me shocked that Danny Ings is the Lesser Talisman King, as a season free from the dark shadow of injury sees him scampering and scurrying to the top of the table after contributing 13.5% of Southampton’s overall points during his 198 point annus mirabilis.
  • FA Cup dropper Pierre-Emerick Aubameyang comes second by just a 0.1% margin. His consistency is remarkable – this is the second season in a row he’s returned 205 points, and the second season in a row he’s obtained the silver medal in the Lesser Talisman category (he lost out to Eden Hazard last time round). More on him to follow…
  • Early season darling Teemu Pukki parties into third, ousting no1 point scorer Kevin de Bruyne from the podium in terms of % of team points scored.
  • 6 teams’ Talismans were goalkeepers.
  • Michail Antonio‘s 7.8% is the lowest Lesser Talisman outcome ever seen. More on him later, too…

As I say every year, I find this analysis really basic: this is demonstrable in the mush of players between 11-12%, which is too bunched up to tell us anything useful.

With this critique in mind, we should ask: how can we improve this analysis?

Well, my answer is to remember the insight we’re looking for from doing this – in the context of evidence-based Talisman Theory, the focus is on identifying the most potent FPL point scoring individuals.

Therefore, to optimise the data set above, our next step is to remove “team based points” so we can home in on individual measures like goals scored, assists and bonus points to create our “Greater Talisman” metric, tallying up what we’ll call “Talisman Points”.

This means things like appearance points and clean sheet points (including negative points incurred for conceding) are removed from the data.

Why?

  • On removing appearance: for the purposes of this analysis, understanding what players added beyond this hygiene measure – the “added value” – is what’s important. For example, if a player plays 38/38 games for 90 mins (Declan Rice this year, for example), they score 76 points by default as a base line which skews the data.
    • I don’t deny players getting the most minutes is an important – if not the most important – part of FPL. However, it’s not as relevant to this analysis as evidenced by the above chart.
  • On removing clean sheets: I am well aware that this penalises defenders and goalkeepers at first glance. However, the point of this analysis is not to uncover the players who provide value for money in terms of their overall performance (which will, shockingly, be covered in the upcoming Value analysis) but those players who make the difference for their teams in terms of end product.

Right, now we’re all agreed (lol), let’s elevate this “Lesser” analysis into something better…

Step 3: Greater Talisman

Happy days, we’re finally at the meat of the report!

Let’s take out all those “team points” and see what we’re finally left with in terms of Talisman Points, ranked by %:

[Link to IMAGE ~ GIF]

Ohh look at that lovely spread – far better than the Lesser analysis.

Anyway, Teemu Pukki surprisingly takes the Greater Talisman crown.

This is actually the second year in a row a departing player has topped the Greater Talisman hierarchy – last year, Eden Hazard won out on this measure.

It’s also not unheard of to find a newly promoted (and subsequently relegated) Talisman performing well – last year, FPL villain Aleksander Mitrovic came 5th in the Greater Talisman ranking.

In any case, this outcome is a great example of where we can apply brain to contextualise the numbers; as with any stats you’ll ever see, you need to think critically about them (even if we are sometimes too mentally lazy to do so).

Basically, this is easily explained: note Norwich’s paltry Talisman points team tally of just 207, and think to yourself “who would have scored the majority of those?” Yup – Pukki.

So it makes sense: the numbers don’t lie. In fact, the Finn’s victory here is the epitome of why Talisman Theory is a provable concept – it shows, using evidenced/data based analysis, how important a Talisman can be to teams as their primary source of points. This is perhaps especially true in teams like Norwich who are reliant on one individual to really perform in order for them to prosper.

Of course, we also know that Norwich are relegated and that Pukki (at time of writing) doesn’t look likely to return to the Prem anytime soon, so let’s say “grats, Teemz” and now move down the list to some more points of interest:

  • If we isolated “still in Prem”, Danny Ings’ 30.1% Talisman points scored would see him top the list. It’s an outstanding performance that makes a mockery of his initial 6.0 pricetag at the start of the 2019/20 season.
    • To underline what a breakthrough year this was for Ings, last year Southampton’s Greater Talisman was, almost laughably, Nathan Redmond. He scored the lowest % of any Greater Talisman in 2018/19 and was outscored by the likes of Felipe Anderson and, damningly, Neil Etheridge and Jonas Lossl.
  • Aubameyang occupies the bronze spot here, but has increased his Talisman point contribution year on year – in the 2018/19 season he scored 22.3% of Arsenal’s points (despite sharing the golden boot) compared to 2019/20’s amassing of 28.8%. It’s worth noting that if he was a midfielder – as is now the case – he’d have surpassed Ings’ total. 
    • Wait up. It’s important digression time

Tell me more about if Aubameyang was a midfielder last season…

If Auba was a midfielder last season, he’d have scored additional 32 points (22 goals, 10 clean sheets) as a basic measure, but according to my recalculations, lost 7 bonus points (strikers score 24 bps per goal, whereas midfielders only get 18). This means he’d have gotten 230 points as a midfielder (3rd highest scoring overall). Calculations by FPL Poker Player also corroborate my maths.

In terms of Talisman analysis, the upshot of the above is adding 15 extra points (+1 point for a midfielder scoring vs a striker and then subtract the 7 lost bonus) to his tally.

If we add the 15 “Talisman points” to his total, he ends up with an amended score of 149 Talisman points. I’ll also add these Talisman points to Arsenal’s overall score.

Here’s what this does to the top 4:

Looking at this, I’m excited by Aubameyang’s reclassification because he becomes the most potent Talisman remaining in FPL based on last season’s numbers. The new 12.0 price tag is expensive, yes. But it could well be worth it.

Anyway, we digress…

Further notes on Greater Talisman (click/tap here for a reminder of the chart)

  • 2018/19’s runner up Raul Jimenez‘s excellent consistency is worth noting – this is the second year in a row he’s scored ~a quarter of Wolves’ FPL points as the Mexican continues to show he merits a spot in our teams.
  • Anthony Martial just about shades out Marcus Rashford (see Sloppy Seconds below) as United’s Talisman, and also shades out last year’s third place Jamie Vardy, due to an end of season surge that saw him score the 3rd most FPL points post restart.
    • It’s also worth mentioning that the United attackers’ places would be reversed had their positions been classified as they are now (in terms of FPL points, my rough calculation is that Martial would have scored 177 points as a forward, Rashford 198 – again, FPL Poker Player concurs with me here).
  • Removing team points bashes Kevin de Bruyne (4th in “Lesser Talisman”) down the rankings as it heightens the spread of points City scored – his scoring of 1 in 5 of the attacking points of the most potent attack in England is still worth shouting about, though.
  • It’s also worth mentioning Mo Salah here, who is often (and to some, surprisingly) absent from these discussions. This is because Liverpool score lots of points, however the Egyptian king actually upped his Talisman point accrual this season – up from 11.6% last season to 18.5% this.
  • The likes of Richarlison (now a Forward), Neal MaupayChris Wood and Michail Antonio (also now a (kindly priced) Forward) being anointed as their team’s Talismans is again validation for the theory – if any goal was scored by those teams, your first thought would have been involvement by the likes of that foursome. Proving that’s the case is what this is all about.
  • Team point removal also sees Wood (rather than Nick Pope – don’t worry, he’ll get his comeuppance in Value), Jordan Ayew (rather than Vicente Guaita), Troy Deeney (rather than Ben Foster), Jonjo Shelvey (rather than Martin Dubravka) and 2019/20’s darling Lord John Lundstram (rather than Dean Henderson) installed as true Greater Talisman for their respective clubs
    • But let’s spare a lol for Bournemouth, for whom Aaron Ramsdale remains Talisman purely based on a pen save and save points. Woeful.

Inevitably, Talisman Theory underlines the importance of those at the sharp end – 15/20 of those on the list can be described primarily as goalscorers. The only exceptions are KdB, Grealish and Shelvey – more seen as creators – and outliers in Lundstram and Ramsdale.

In other parishes, you may expect the analysis to end here.

But to me, if we’re being good researchers, it’s always worth exploring some extra angles to try to couch this data in further context and therefore provide more insight.

This is especially true of quantifying the impact of the elephant in the room: the pause between GW1-29 and the restart post-lockdown of GW30-8.

Exit Velocity – or, this year, Pre v Post Lockdown Comparison

In usual times, this analysis charts how things developed between a certain snapshot (usually around March) and the end of the season to determine Exit Velocity and therefore who may be worth a look for the upcoming season.

This season, though, it’s my way to include this weirdest of season contexts within the Talisman analysis.

This is done through comparing pre and post lockdown points scored and identifying if the Talisman changed between these two periods, as well as calculating the Exit Velocity where I can (ie where a Talisman was the same both times).

This looks like so:

[Apologies if this is a bit cramped on mobile – link to Image]

Note that if there was a new Talisman after restart who was not a Season Talisman then the “exit velocity” isn’t calculated for ease of reference.

Key points:

  • The headline finding is that Pre and post lockdown, no less than 13 different Talismen emerged.
    • The identity of some of these won’t surprise anyone, as the likes of Bruno Fernandes, Raheem Sterling and, most explosively thanks to his 4 goal haul against Norwich, Michail Antonio emerged as Talismen supplanting the pre-lockdown incumbent.
    • Bruno’s emergence as an option of course dilutes the United points concentration still further. His performance after the restart was exemplary, with his 56 Talisman (ie non appearance/clean sheet/etc) points over the 8 games a crazy score – that number beats the season score for the Talismen of 3 clubs (Newcastle, Sheffield United, Bournemouth) and is equal to Watford’s Deeney. He would undoubtedly have been in with a huge shout of being United’s overall Talisman if he’d been around all season given that momentum.
    • Sterling pushing De Bruyne close also underlines what I mentioned above – goals > assists in FPL. The England winger underperformed in terms of points pre-lockdown but managed to close the gap considerably by the end. As both are now 11.5m, it’ll be interesting to see which performs best according to that generous price tag.
    • Josip Drmic. LOL. This is just because of how terrible Norwich were post lockdown. Played 8. Lost 8. Oof.
    • A note here on Harry Kane, who became the only player – fuelled by his restart consistency – to reach the 150 points landmark despite playing less than 30 games (h/t Tom Campbell). One to watch next year potentially, especially at 10.5m.
  • All of this means only 7 kept up their Talisman status between Pre and Post.
  • 2 players – Deeney and Antonio – won the mantle of season-long Talisman overall based on their restart performances.
  • The numbers for Antonio in particular are just astonishing 71% of his Talisman points were scored after the restart, along with 59% of his actual points (66/111). He blows everyone around him out of the water in terms of lockdown gain.
  • This being said, Ings’ consistency was simply brilliant, even outdoing “the Postman” Jimenez. The Saints striker scored 30% of his already impressive Talisman points haul post lockdown, though the Wolves frontman’s 24.4% is not to be sniffed at either.
  • Elsewhere, Aubameyang cemented his Talisman status by scoring just over a quarter of his Talisman points post lockdown (admittedly a few were gifted by the Norwich defence), as did the likes of Maupay (to Bernd Leno’s chagrin) and, to a lesser extent, the likes of Salah, Vardy and the reclassified Richarlison.
  • Willian’s set play proficiency saw him catch and match Tammy Abraham to jointly occupy the role of Chelsea’s Talisman – the second year in a row a departing player has held the Blues’ Talisman position.
    • Keep an eye on Chelsea this season. The advent of big signings Timo Werner, Hakim Ziyech and, maybe, Kai Havertz are due to shake things up big time in the Talisman stakes, plus Captain America himself Christian Pulisic will be hoping to find more consistency after a stop-start, injury pockmarked season within which he sometimes burned bright.

Let’s bring it back to the course of the season with our final step: Sloppy Seconds.

Extra Analysis: Sloppy Seconds – which teams are the most Talismanic?

This elegantly named analysis answers the question of who was the nearly man in terms of being the Talisman per team, and then asks: what can that tell us?

I’ve made some references to this already, but here it is in its entirety:

[Apologies if this is a bit cramped on mobile – link to Image]

A big benefit of this further step is that we can identify which teams were truly Talismanic, versus those for whom FPL points scoring was more a group effort. We’ve already done some of this by looking at low % Talisman scorers, but we can also do this by looking at the difference between the Talisman and the Sloppy Second.

I’ve again been lazy and just done (x-y) to accentuate these differences (hey this is free, you get what you’re given 😉 ), but it shows what we need to see.

Key notes:

  • Danny Ings is the player whose team are the most reliant on him for FPL points – a massive 86 point difference, or 20% margin, between him and Stuart Armstrong shows how important his newfound ironclad fitness is to Ralph Hasenhuttl’s side.
    • Big note – last year, Ings himself was a Sloppy Second. In 2018/19, as alluded to above, he was just 2% behind then Southampton Talisman Nathan Redmond – which just goes to show a Talisman can emerge from anywhere.
  • Elsewhere, Arsenal, Leicester and Wolves are Talisman-friendly teams as the usual suspects in Aubameyang, Vardy and Jimenez record 10%+ margins over their respective silver medallists.
    • Back to Aubameyang again. His reclassification heightens his appeal on the Sloppy Seconds measure, too. This is because applying the extra points he’d have scored if a midfielder, the margin between him and Alex Lacazette goes from 13.9% to 18.3% – just 1.7% behind the gap Ings posted vs his Sloppy Second, Armstrong. 
  • We shouldn’t discount Norwich as an instructive picture of what can happen with newly promoted teams. Pukki had a clear lead over Krul, which may lead to assessments of the newly promoted sides and which figures may be Talismanic for them. The likes of Mateus Pereira at WBA and the aforementioned Mitrovic at Fulham spring to mind as candidates (depending on price).
  • Any team with <4% difference between Talisman and Sloppy Second I’d be happy to characterise as non-Talismanic
    • As mentioned, Chelsea epitomise this through Willian and Abraham matching each other
    • Elsewhere, Salah and Mane have only 6 points between them and a similarly small gap was also seen between Son and Kane (4)
    • Furthering the big club theme, for Man Utd we only saw 7 points between Martial and Rashford – but this would be reversed if their positions were. The introduction of Bruno Fernandes may also see United share the points even more next campaign.
    • KdB and Antonio are just outside of this 4% threshold conveniently (yeah, I know, sue me) because of events after the restart – but for different reasons. As said above, the Belgian creator found that the gap lessen between himself and Sterling as the restart wore on due to Raz finding his scoring boots, whereas for Antonio it was more a case of a last ditch sprint past a non-starting Haller (especially thanks to his 4 goal haul versus Norwich).
  • There are a few teams in the middle (Talisman 4-8% better than Sloppy Second), which are perhaps in transition (eg Brighton, Everton), and could see a Talisman emerge next year. As we saw with the Ings example, could the likes of a Maupay or Richarlison – or even the Brazilian’s Sloppy Second Dominic Calvert-Lewin – have a breakout season if they both hit the ground running and consistently exceed expectations?

Evaluation

Caveats first (sorry Marz):

  • There are lots of ways to analyse data. You may disagree with the method or the findings, or indeed think a different approach is better. Perfectly reasonable – but, if so, you do the work and show how you can do it better (but please credit and don’t brazenly steal others’ work or ideas!)
  • Context is key. Pukki is the Greater Talisman king according to the numbers. But if you apply brain, I’m not saying that Pukki would have been a better pick than, say, KdB throughout the season.
  • Talisman Theory is part of – not the full -picture. This article and analysis is not claiming to have solved FPL nor be the most important bit of information you should use when picking your team. Hopefully the fact it’s evidence based means it should be worthy of consideration, but please weigh this analysis alongside others you may have read, listened to or made yourself before making decisions.

So what have we learnt?

Aubameyang: a goalscoring midfield Talisman looks irresistible. Like Chelsea in the Hazard days, whatever your views on them, Arsenal’s recent underachievement belies that fact that they are a team with lofty goals, and good standard of players, that have a Talisman. As we’ve seen through reformatting Auba’s data to account for his new classification as a midfielder, he emerges as a hero of this article and one I am definitely having in my FPL side (pending fixtures) to begin with – especially if events transpire that mean he’s shunted back to Centre Forward (e.g. Alexandre Lacazette departs). The arrival of Willian may also mean (assuming professional Fortnite player Mesut Ozil remains ostracised) Aubameyang has some decent chance creation servicing him as well this season. As a midfielder, he’ll be a solid captain pick most Gameweeks – especially at the Emirates – and, provided he remains in North London, is heartily recommended via this analysis for a spot on your team from the off.

Pukki a good example of how newly promoted teams can be Talismanic – as long as you time it right. I’m not saying buy Mitrovic or anything like that, but – at the risk of being cancelled on social media – I am saying that this kind of individual is a fairly obvious candidate to be a Talisman for their team. This was a secondary point last year when the Serbian came 5th in the rankings but, with Pukki coming first this year to amplify that point, it now suddenly seems far more relevant. As always, it’ll be about context – getting the player in when he’s hot. Pukki outscored everyone in the first 5 gameweeks of last season before his point returns began to tumble downward. Getting the timing right is always key.

Universal Ings ownership means the search for the next Value Talisman is on. Danny Ings is likely to be in many teams in Gameweek 1 2020/21, but he can be viewed now as a bit of a hygiene factor. At the start of 2019/20, he was nowhere in the stats, 2nd in the data to Nathan Redmond and, generally, a bit of an injury risk at the start of the season – the legacy of which caused many of us to overlook him despite the now ridiculous looking 6.0 price tag (to our detriment). He’s now an insta-pick for many, meaning that finding the next one is a great way to get ahead. As mentioned, the likes of Maupay at Brighton or Richarlison at Everton could be some candidates to monitor.

Only 4 teams had true Talisman reliance in 2019/20. The yawning chasm between City, Liverpool and the rest that was documented at the start of this report highlights how the challenge (as always) is in picking the points scorers you can afford after the inevitable majority budget allocation to those premiums. Unfortunately, Vardy and Aubameyang were expensive and remain so, and Jimenez (then 7.5) and Ings (then 6.0) have also seen price increases. Feasibly, however, this quad should still continue offering us a source of Talisman and, indeed, FPL points for our money. This means they should be viewed as the next level down priorities to consider fitting in after we’ve picked our premiums – especially Aubameyang.

The traditional big clubs are again non-Talismanic. Though Anthony Martial was the Talisman at United, points were shared far more given the breadth of quality in the frontline that could cannibalise points, as you can see in the Sloppy Seconds data. This is especially accentuated by the signing of Bruno and also the emergence of Mason Greenwood (any potential new signing aside). It’s common that we see this greater spread of Talisman points for clubs like United, City and Liverpool. You don’t need a specialised analysis like this one to know that picking players from those clubs is probably a good idea. But, in the light of this analysis, this cannibalisation at the “bigger” clubs underpins the importance of a case-by-case approach to their key men: where some may be scary to not own due to mass ownership and/or performance, in other instances (think Pep roulette) it’s who you judge will score you the most points in the current context.

Conclusion

From doing this analysis, this year’s TL;DR is pretty much as follows:

  • Aubameyang is strongly recommended as a buy for your teams (cue a missed pen and red in Gameweek 1, eh)
  • Keep an eye out for the next Ings

The key takeaway, however, remains that Talismen should be high on our agenda when deciding who to buy in at any point because the data shows that focusing on these guys in how we build our teams should help maximise your FPL point gain throughout the season.

Many thanks to Mitchell Stirling for his data expertise, Tomasz Kucharski for the design work (find him here) and then Tom Campbell and Karam Tayser for proof reading.