Overnight FPL (@OvernightFPL) header
Overnight FPL (@OvernightFPL) profile picture

FPL insights, for everyone. Overnight moves, DefCon analytics, a model with its own team. Want something dug into? Ask.

Joined July 2026
Photo by Overnight FPL (@OvernightFPL)Photo by Overnight FPL (@OvernightFPL)Photo by Overnight FPL (@OvernightFPL)Photo by Overnight FPL (@OvernightFPL)Photo by Overnight FPL (@OvernightFPL)Photo by Overnight FPL (@OvernightFPL)
Everyone is loading up on DEFCON defenders this season. Last year only 15 of 93 regular defenders took the two defensive points in half their games. These are the 15. Look who is not on it. Van Dijk, Gabriel, Saliba and Gvardiol all missed. Their value was clean sheets, not defensive actions. Who are you choosing? #FPL
1
2
155
Next up in the Speak FPL series, the scoring engine! The three letters behind every projection we publish. xG and xA measure the chances, xP turns them into expected points under this season's rules.
1
2
74
Why xP matters: The number the squad is somewhat picked on. This is more of a forecast, than a promise of points.
1
13
Every term we use lives in this series, explained once, properly. Reply with the acronym that annoys you most and it goes in the next card.
12
Why xG matters: A player who takes great chances but misses is usually unlucky, not bad. xG highlights that before the goals convert.
2
18
Why xA matters: Separates creators having a quiet run from creators who have actually stopped creating.
12
Starting the Speak FPL mini series. FPL is played by all, so let’s be grounded in all the acronyms and definitions. Starting with the absolute basics I’m sure everyone knows! #FPL #FPLCommunity
68
Nine players who turned up every GW. Anderson is the outlier though with 13.8 defensive actions per start against the 12 a midfielder needs, clearing the bar in 61% of starts. Everyone else here is a defender clearing 10. City impact? #FPL follow for more graphics like these!
1
82
The bar is his average actions per start last season, the line is the threshold. The % is not last season’s rate, it is what we expect next season, which is why it is lower. Senesi cleared it 70 percent of the time. We print 63 because rates that high almost always come down.
16
Some info on what this account is. One manager, one model, a real squad picked in public, and its record published - let’s see what happens!
1
61
Why this exists: most FPL content is confident and unsourced, and most FPL data is accurate and unreadable. There is a gap in the middle for numbers that are both checkable and easy to follow.
10
29,757. That’s every player-gameweek from last season, re-run through the new scoring including DefCon, the defensive contribution points system. No shortcuts. A model manages a real team, with human touch and posts every result, good and bad, in public.
1
66
What you will get here: the overnight price and points corrections most managers sleep through, defensive contribution analytics with measured hit rates, and what the top 50 managers in the world are doing.
1
17
A model also picks a real squad in public every week, with human touch for any amends. Its calls are published before the deadline and scored after it, right or wrong. Every number traces to public data. #fpl
12