FPL DataMonkey (Gibby)
@fpl.datamonkeys.app
#FPLDraft with a bit of #FPL, and I quite like data
Data scientist when I’m not playing FPL
🌍 London
Data scientist when I’m not playing FPL
🌍 London
Pinned
--- Expected Fantasy Assists (xFA) ---
Ok the results are in, how does xFA v1.0 perform in comparison to xA?
Well I looked at 2023-24 #FPL season data and xFA is 3.6% off the FPL assist total, compared to 29.8% off for xA - that's an 8x improvement (!!)
🧵
Ok the results are in, how does xFA v1.0 perform in comparison to xA?
Well I looked at 2023-24 #FPL season data and xFA is 3.6% off the FPL assist total, compared to 29.8% off for xA - that's an 8x improvement (!!)
🧵
Reposted by FPL DataMonkey (Gibby)
We can use xG to compute post-hoc win-draw-loss probabilities.
Our new blogpost shows MCMC's estimates of these probabilities are unstable, computing them exactly is faster than approximating them, and MCMC's runtime strongly depends on its implementation.
dtai.cs.kuleuven.be/sports/blog/...
Our new blogpost shows MCMC's estimates of these probabilities are unstable, computing them exactly is faster than approximating them, and MCMC's runtime strongly depends on its implementation.
dtai.cs.kuleuven.be/sports/blog/...
Expected goals and the Monte Carlo Trap
On a team level, xG is often used to assess the quality of the goal scoring opportunities that a team creates and concedes. That is, it…
dtai.cs.kuleuven.be
September 11, 2025 at 11:34 AM
We can use xG to compute post-hoc win-draw-loss probabilities.
Our new blogpost shows MCMC's estimates of these probabilities are unstable, computing them exactly is faster than approximating them, and MCMC's runtime strongly depends on its implementation.
dtai.cs.kuleuven.be/sports/blog/...
Our new blogpost shows MCMC's estimates of these probabilities are unstable, computing them exactly is faster than approximating them, and MCMC's runtime strongly depends on its implementation.
dtai.cs.kuleuven.be/sports/blog/...
I find it interesting how much weight/trust people put on #FPL model predictions - especially at the start of the season
I remember debating people about Review putting Eze to be the top scorer gw1-6 last year 🤦♂️
Review had him down for 230pts last year which is, and was, insane
I remember debating people about Review putting Eze to be the top scorer gw1-6 last year 🤦♂️
Review had him down for 230pts last year which is, and was, insane
August 25, 2025 at 4:59 PM
I find it interesting how much weight/trust people put on #FPL model predictions - especially at the start of the season
I remember debating people about Review putting Eze to be the top scorer gw1-6 last year 🤦♂️
Review had him down for 230pts last year which is, and was, insane
I remember debating people about Review putting Eze to be the top scorer gw1-6 last year 🤦♂️
Review had him down for 230pts last year which is, and was, insane
Has anyone seen any FPL models or predictions that include uncertainty (or error)? Or anyone who discusses it particularly
It really frustrates me that predictions are shown as absolute values, when we’re playing a game with so much variation.
It really frustrates me that predictions are shown as absolute values, when we’re playing a game with so much variation.
August 1, 2025 at 9:18 AM
Has anyone seen any FPL models or predictions that include uncertainty (or error)? Or anyone who discusses it particularly
It really frustrates me that predictions are shown as absolute values, when we’re playing a game with so much variation.
It really frustrates me that predictions are shown as absolute values, when we’re playing a game with so much variation.
Reposted by FPL DataMonkey (Gibby)
Yearly reminder that pre-season is an unreliable indicator of any 'meaningful' season trends.
If there is any signal, it is in fitness, squad availability, tactics/set-up/formations.
Individual performances, goals, assists, and especially results are almost entirely noise.
If there is any signal, it is in fitness, squad availability, tactics/set-up/formations.
Individual performances, goals, assists, and especially results are almost entirely noise.
July 31, 2025 at 10:19 AM
Yearly reminder that pre-season is an unreliable indicator of any 'meaningful' season trends.
If there is any signal, it is in fitness, squad availability, tactics/set-up/formations.
Individual performances, goals, assists, and especially results are almost entirely noise.
If there is any signal, it is in fitness, squad availability, tactics/set-up/formations.
Individual performances, goals, assists, and especially results are almost entirely noise.
Reposted by FPL DataMonkey (Gibby)
{THREAD} Introducing xDC (Expected DEFCON)
- Chance of a defender to get to 10 CBIT in #FPL - RT appreciated
All data from @ffscoutfpl.bsky.social
I will post a few plots with a brief description.
1. DC (DEFCON) points per 90 from 24-25 season vs Expected DEFCON points (2*xDC) per 90
- Chance of a defender to get to 10 CBIT in #FPL - RT appreciated
All data from @ffscoutfpl.bsky.social
I will post a few plots with a brief description.
1. DC (DEFCON) points per 90 from 24-25 season vs Expected DEFCON points (2*xDC) per 90
July 30, 2025 at 8:11 PM
{THREAD} Introducing xDC (Expected DEFCON)
- Chance of a defender to get to 10 CBIT in #FPL - RT appreciated
All data from @ffscoutfpl.bsky.social
I will post a few plots with a brief description.
1. DC (DEFCON) points per 90 from 24-25 season vs Expected DEFCON points (2*xDC) per 90
- Chance of a defender to get to 10 CBIT in #FPL - RT appreciated
All data from @ffscoutfpl.bsky.social
I will post a few plots with a brief description.
1. DC (DEFCON) points per 90 from 24-25 season vs Expected DEFCON points (2*xDC) per 90
Anyone seen any analysis on the new BPS system for tackles?
Old: 2 BPS per net tkl (tkls won - tkls lost)
New: 2 BPS per tkl won
For an example:
Munoz attempted 123 tkls, won 80, lost 43 (net 37)
Previous BPS: 74
New BPS: 160
#FPL
Disclaimer - Fbref data not Opta, so not identically comparable
Old: 2 BPS per net tkl (tkls won - tkls lost)
New: 2 BPS per tkl won
For an example:
Munoz attempted 123 tkls, won 80, lost 43 (net 37)
Previous BPS: 74
New BPS: 160
#FPL
Disclaimer - Fbref data not Opta, so not identically comparable
July 23, 2025 at 9:16 AM
Anyone seen any analysis on the new BPS system for tackles?
Old: 2 BPS per net tkl (tkls won - tkls lost)
New: 2 BPS per tkl won
For an example:
Munoz attempted 123 tkls, won 80, lost 43 (net 37)
Previous BPS: 74
New BPS: 160
#FPL
Disclaimer - Fbref data not Opta, so not identically comparable
Old: 2 BPS per net tkl (tkls won - tkls lost)
New: 2 BPS per tkl won
For an example:
Munoz attempted 123 tkls, won 80, lost 43 (net 37)
Previous BPS: 74
New BPS: 160
#FPL
Disclaimer - Fbref data not Opta, so not identically comparable
Alongside all the issues with the #FPL site/app so far - not populating the DefCon points feels like a big miss.
We know they're calculated for last season (really they could easily calculate back a lot further), why not include them to help managers gauge the change?
We know they're calculated for last season (really they could easily calculate back a lot further), why not include them to help managers gauge the change?
July 22, 2025 at 2:24 PM
Alongside all the issues with the #FPL site/app so far - not populating the DefCon points feels like a big miss.
We know they're calculated for last season (really they could easily calculate back a lot further), why not include them to help managers gauge the change?
We know they're calculated for last season (really they could easily calculate back a lot further), why not include them to help managers gauge the change?
Don't love the fact that FPL seem to have forgotten to click GO on FPL draft as well as the main game
July 21, 2025 at 12:35 PM
Don't love the fact that FPL seem to have forgotten to click GO on FPL draft as well as the main game
Reposted by FPL DataMonkey (Gibby)
elevenify.com is now free for all.
📈 Team Ratings
⚽️ Predicted Goals + Assists
🧤 Predicted Clean Sheets
⚖️ Predicted Win%
🖥️ Simulated Table
⭐️ Fantasy Predictions
💡 Decision Frameworks
📊 Team Style Profiles
Enjoy 💜
📈 Team Ratings
⚽️ Predicted Goals + Assists
🧤 Predicted Clean Sheets
⚖️ Predicted Win%
🖥️ Simulated Table
⭐️ Fantasy Predictions
💡 Decision Frameworks
📊 Team Style Profiles
Enjoy 💜
March 20, 2025 at 8:54 PM
elevenify.com is now free for all.
📈 Team Ratings
⚽️ Predicted Goals + Assists
🧤 Predicted Clean Sheets
⚖️ Predicted Win%
🖥️ Simulated Table
⭐️ Fantasy Predictions
💡 Decision Frameworks
📊 Team Style Profiles
Enjoy 💜
📈 Team Ratings
⚽️ Predicted Goals + Assists
🧤 Predicted Clean Sheets
⚖️ Predicted Win%
🖥️ Simulated Table
⭐️ Fantasy Predictions
💡 Decision Frameworks
📊 Team Style Profiles
Enjoy 💜
Reposted by FPL DataMonkey (Gibby)
❗Data Deep Dive Week 6❗
Moyes' Everton look reborn, fresh and full of quality after his hiring, whereas Amorim's Man United looks even more grim then in the Erik ten Hag era. For this week, let's dive into the famous concept of the 'New Manager Bounce'.
A thread ⬇️
Moyes' Everton look reborn, fresh and full of quality after his hiring, whereas Amorim's Man United looks even more grim then in the Erik ten Hag era. For this week, let's dive into the famous concept of the 'New Manager Bounce'.
A thread ⬇️
February 6, 2025 at 4:18 PM
❗Data Deep Dive Week 6❗
Moyes' Everton look reborn, fresh and full of quality after his hiring, whereas Amorim's Man United looks even more grim then in the Erik ten Hag era. For this week, let's dive into the famous concept of the 'New Manager Bounce'.
A thread ⬇️
Moyes' Everton look reborn, fresh and full of quality after his hiring, whereas Amorim's Man United looks even more grim then in the Erik ten Hag era. For this week, let's dive into the famous concept of the 'New Manager Bounce'.
A thread ⬇️
Reposted by FPL DataMonkey (Gibby)
I went back and calculated #FPL Assistant Manager points for every PL game over the last 8 seasons (and combined with pre-game market WDL odds)
These are the top 100 AM scores for games played in GW24-38 between 2016-7 and 2023-24 👇
And a few quick observations below 🧵
These are the top 100 AM scores for games played in GW24-38 between 2016-7 and 2023-24 👇
And a few quick observations below 🧵
January 16, 2025 at 12:57 PM
I went back and calculated #FPL Assistant Manager points for every PL game over the last 8 seasons (and combined with pre-game market WDL odds)
These are the top 100 AM scores for games played in GW24-38 between 2016-7 and 2023-24 👇
And a few quick observations below 🧵
These are the top 100 AM scores for games played in GW24-38 between 2016-7 and 2023-24 👇
And a few quick observations below 🧵
Seen a bit of chat around Cunha's ban length, so going to drop in my two cents and the closest comparable bans I've found
Overall (my opinion): minimum 8 games, my expectation is 10-12, but could be rest of the season if the FA want to send a message
🧵
Overall (my opinion): minimum 8 games, my expectation is 10-12, but could be rest of the season if the FA want to send a message
🧵
December 18, 2024 at 1:22 PM
Seen a bit of chat around Cunha's ban length, so going to drop in my two cents and the closest comparable bans I've found
Overall (my opinion): minimum 8 games, my expectation is 10-12, but could be rest of the season if the FA want to send a message
🧵
Overall (my opinion): minimum 8 games, my expectation is 10-12, but could be rest of the season if the FA want to send a message
🧵
Finally the club are starting to trickle in over here - any I've missed of just these 5 so far?
December 16, 2024 at 9:56 AM
Finally the club are starting to trickle in over here - any I've missed of just these 5 so far?
--- GW15 xFA top 10 ---
A lot of underperformance in xFA at the top of the table this week - and a few interesting scenarios I wanted to look into further
Top 3:
🐓Son: 0.47xA -> 1.27xFA (0 FPL assists)
🍒 Kerkez: 0.12xA -> 1.07xFA (0 FPL assists)
🔴Havertz: 0.01xA -> 0.86xFA (1 FPL assist)
🧵
A lot of underperformance in xFA at the top of the table this week - and a few interesting scenarios I wanted to look into further
Top 3:
🐓Son: 0.47xA -> 1.27xFA (0 FPL assists)
🍒 Kerkez: 0.12xA -> 1.07xFA (0 FPL assists)
🔴Havertz: 0.01xA -> 0.86xFA (1 FPL assist)
🧵
December 12, 2024 at 2:14 PM
--- GW15 xFA top 10 ---
A lot of underperformance in xFA at the top of the table this week - and a few interesting scenarios I wanted to look into further
Top 3:
🐓Son: 0.47xA -> 1.27xFA (0 FPL assists)
🍒 Kerkez: 0.12xA -> 1.07xFA (0 FPL assists)
🔴Havertz: 0.01xA -> 0.86xFA (1 FPL assist)
🧵
A lot of underperformance in xFA at the top of the table this week - and a few interesting scenarios I wanted to look into further
Top 3:
🐓Son: 0.47xA -> 1.27xFA (0 FPL assists)
🍒 Kerkez: 0.12xA -> 1.07xFA (0 FPL assists)
🔴Havertz: 0.01xA -> 0.86xFA (1 FPL assist)
🧵
98% red arrow and Palmer (C) scored more than my entire free hit team combined 🤩
December 8, 2024 at 8:29 PM
98% red arrow and Palmer (C) scored more than my entire free hit team combined 🤩
Well this didn’t go to plan….
Nothing like waking up to a storm calling off a game - 25 mins later the free hit is played and I've gone full chaos and captained KDB 😍
December 7, 2024 at 8:00 PM
Well this didn’t go to plan….
Nothing like waking up to a storm calling off a game - 25 mins later the free hit is played and I've gone full chaos and captained KDB 😍
December 7, 2024 at 11:02 AM
Nothing like waking up to a storm calling off a game - 25 mins later the free hit is played and I've gone full chaos and captained KDB 😍
Reposted by FPL DataMonkey (Gibby)
Visualising the strength of the Premier League "Big Six" with a long-term moving average of their xG dominance: a (logged) ratio of their xG relative to their xGA.
For the first time in the last 6 years Man City are not 1st by this measure, but in fact 3rd behind Arsenal (1st) & Liverpool (2nd).
For the first time in the last 6 years Man City are not 1st by this measure, but in fact 3rd behind Arsenal (1st) & Liverpool (2nd).
December 3, 2024 at 9:35 PM
Visualising the strength of the Premier League "Big Six" with a long-term moving average of their xG dominance: a (logged) ratio of their xG relative to their xGA.
For the first time in the last 6 years Man City are not 1st by this measure, but in fact 3rd behind Arsenal (1st) & Liverpool (2nd).
For the first time in the last 6 years Man City are not 1st by this measure, but in fact 3rd behind Arsenal (1st) & Liverpool (2nd).
Reposted by FPL DataMonkey (Gibby)
--- GW13 xFA top 12 ---
Some HUGE xFA performance this week, with Evanilson probably having the highest difference between xFA and xA that we will ever see (239x higher!)
Top 3:
🍒 Evanilson: 0.01xA -> 2.39xFA thanks to 3 penalties won
🔴 Saka: 1.11xA -> 1.77xFA
🎄Jota Silva: 0.54xA -> 1.07xFA
Some HUGE xFA performance this week, with Evanilson probably having the highest difference between xFA and xA that we will ever see (239x higher!)
Top 3:
🍒 Evanilson: 0.01xA -> 2.39xFA thanks to 3 penalties won
🔴 Saka: 1.11xA -> 1.77xFA
🎄Jota Silva: 0.54xA -> 1.07xFA
December 2, 2024 at 3:34 PM
--- GW13 xFA top 12 ---
Some HUGE xFA performance this week, with Evanilson probably having the highest difference between xFA and xA that we will ever see (239x higher!)
Top 3:
🍒 Evanilson: 0.01xA -> 2.39xFA thanks to 3 penalties won
🔴 Saka: 1.11xA -> 1.77xFA
🎄Jota Silva: 0.54xA -> 1.07xFA
Some HUGE xFA performance this week, with Evanilson probably having the highest difference between xFA and xA that we will ever see (239x higher!)
Top 3:
🍒 Evanilson: 0.01xA -> 2.39xFA thanks to 3 penalties won
🔴 Saka: 1.11xA -> 1.77xFA
🎄Jota Silva: 0.54xA -> 1.07xFA
I know he’s not fully nailed but really tempted by Rutter instead of Rogers this week
(Bruno & Mbeumo -> Saka + Rogers was the original plan for this week)
(Bruno & Mbeumo -> Saka + Rogers was the original plan for this week)
December 1, 2024 at 7:04 PM
I know he’s not fully nailed but really tempted by Rutter instead of Rogers this week
(Bruno & Mbeumo -> Saka + Rogers was the original plan for this week)
(Bruno & Mbeumo -> Saka + Rogers was the original plan for this week)