Last updated on September 25th, 2024
Written by FPL Frano
Projected points, AI team ratings, xGI, and so on, it’s everywhere. You will hardly find an FPL-related tweet or article where one of these words is not mentioned (I could have added fixture difficulty, but unlike projected points, this one I consider relevant). If you are an engaged FPL Manager, like it or not, it will affect your decision-making to a certain extent. But is it really helpful and what consequences does it have on our teams?
1. Questionable algorithms
If we go back to late July or the beginning of August when draft creations were in full swing and take a quick look at projected points, we could be surprised.
Ben Brereton Diaz is the most obvious example. He was projected to score 25.93 points in the opening six game weeks, higher than Smith Rowe’s 21.08 or any other midfielder in this price range. He was also projected to outscore some of the most popular midfield picks, Gordon (23.32) and Mbeumo (24.82) for example, not even too far from Eze (27.03). It’s fair to say that he failed to back those expectations and while we are still waiting for his first attacking return of the season, his price has already fallen from 5.5 to 5.3.
Morgan Rogers is another example, whose projections were around 16.5 points for the first six games (currently on 18 with Ipswich game to follow). He already outscored his projection and should have registered at least double these points, considering chances created and missed. However, Adam Wharton, Will Smallbone, and even Carlos Alcaraz (yes, the one who signed for Flamengo) were all expected to outscore him in this period, based on various algorithms.
2. Eye-Test > Numbers
It’s impossible to watch and scout all games but I would strongly suggest watching at least those teams whose players you are targeting or relying on for the upcoming period. For example, if you didn’t see how dangerous Rogers was in the final third in the opening four games where he somehow dodged points, you would at least consider selling him and missing out on his 10-pointer in the last game. Similar could be said for Eze, although he is yet to justify his price with points returns. Nailed for 90 minutes, team’s best player, constant presence in the finishing zones, ticks most of the boxes needed to get into my FPL team. Palace’s poor form is a slight concern, to be honest.
3. Conclusion
Data models and AI are everywhere nowadays and don’t get me wrong, I am not against it. Personally, love data analytics, stats, numbers et cetera but do we really need it to this extent to play FPL? Don’t think so. Most of the players can get the AI Team rating of 90+, but what good does it bring you in the long term, is the rotation and transfer plan calculated in this rating, will you ask AI to suggest transfers every weekend? In the end, this is a game of decisions and in my opinion, getting the right differential player based on your expectations is a lot more fun than just filling the team with players based on expected points and AI-generated feedback. The same can be applied vice versa, if your player ends up being a miss and disappointing pick, you have only one person to blame – yourself.
