Update (Friday 28 August, before the deadline). Thursday’s press conferences changed a few things. Elliot Anderson (City, 6.3% ownership) and Ferdi Kadıoğlu (Brighton) have lost their flags — both are back to a full 100% chance of playing, so yesterday’s warning no longer applies to them. Still flagged: Gibbs-White (8.7%, knee) and Caicedo (3.1%), plus a new one — Jack Hinshelwood (Brighton, 2.6%, 75%), the same player who scored 16 points on opening weekend. Ismaïla Sarr remains out. Prices have now moved for seventeen players (ten yesterday), and six of the falls are Arsenal players ahead of Monday’s trip to Villa Park. One practical note: Ayyoub Bouaddi is now in the game as a Manchester City player at £5.5m — buyable for the first time. Wan-Bissaka (Villa), Di Gregorio (Bournemouth) and Hadjam (Brighton) have also been added. Delap, Marmoush and Barcola are all still listed at their old clubs.
There are 24 hours left before the Gameweek 2 deadline — it closes at 6.30pm on Friday 28 August. Two things have happened in the last day that are worth checking before your final transfers: the price market has properly woken up, and in several places the crowd is doing the opposite of what our model suggests.
Prices: from four moves to ten in a single day
On Wednesday we reported that prices had moved for just four players. Today it is ten — the market has finally started.
Risers: Riccardo Calafiori (Arsenal, up to £5.6m), Michael Kayode (Brentford, £4.6m), Mamadou Sangaré (Brentford, £5.6m), Maxim De Cuyper (Brighton, £4.6m) and João Pedro (Chelsea, £7.6m). Fallers include Ollie Watkins (down to £7.9m), Viktor Gyökeres (£7.4m) and Gabriel Martinelli.
The same point as Wednesday still stands: £0.1m is not a reason to make a transfer. If a player belongs in your squad, he belongs there at the higher price. If he does not, a tenth off changes nothing.
Where the crowd is right
The three most-bought players this week look good on our numbers too. De Cuyper — a £4.6m defender with 17 points in Gameweek 1 — rates 80. Kayode, also a defender, 13 points and a rating of 85. Cole Palmer, 13 points and a rating of 83.
This is the case where the market and the model agree and the decision is easy.
Three places where we disagree with the crowd
Sangaré: 14 points and a rating of 52
Mamadou Sangaré is the second most-bought player of the week — more than 320,000 transfers in. He scored 14 points in the opening round and has just risen in price. Our rating is 52 with a sell verdict.
Why the gap? The model does not rate one match, it rates repeatability: minutes, underlying numbers, role in the side and fixtures. Fourteen points on opening weekend is a fine return, but it does not change the profile of a player without a season of regular football at this level behind him. Buying him now means paying for what has already happened.
To be fair: if Sangaré does it again next week, his rating will climb sharply and we will be the ones pointing at him. Today the data does not support it.
Bruno Fernandes: sold despite a rating of 89
Bruno Fernandes is among the six most-sold players — over 165,000 managers moved him on after two points in Gameweek 1. He still holds a rating of 89, the second highest in the game, at 48.9% ownership.
In Gameweek 2 Manchester United host Ipswich, one of the lowest-difficulty fixtures of the round. Selling him the day before that, on the back of one quiet performance, is reacting to the last result rather than to the data.
Calafiori: bought heavily at a rating of 76
Calafiori is the fourth most-bought player of the week and is now owned by 41.4% of managers. Our rating: 76, verdict hold. He is not a bad pick — he is a more popular pick than the data justifies. In the same Arsenal defence we rate Gabriel higher, and on Monday Arsenal travel to Villa Park, the only fixture of the round rated 4 for both sides.
The case where our model is the blind one
Ollie Watkins carries a rating of 84 and a buy verdict, while being the fourth most-sold player of the week and having just dropped in price. Here the crowd is right and the model is not.
The reason is simple: Watkins is in advanced talks with Al Hilal and played no minutes in Gameweek 1. Our model reads football data, not transfer reports. That 84 describes an Aston Villa striker, not a player who may be about to leave the league.
It is the same blind spot we described with Delap and Marmoush, only reversed: the model can be too harsh on players who are not playing, and too generous to players who are packing their bags.
What changed in the flags
One change matters: Ismaïla Sarr (Crystal Palace, 3.1% ownership) has gone from “75% chance of playing” to a full groin injury with no return date. If he is in your squad, this is no longer a risk call — he will not play.
Unchanged are the two flags affecting more managers: Morgan Gibbs-White (Forest, 9.3%) with a knee problem and Elliot Anderson (City, 6.3%) with a knock, both at 75%. Gibbs-White is also the second most-sold player of the week. The full list is in our Injury Centre.
Captaincy is unchanged
Our ranking is led by Szoboszlai (rating 90, the only Strong Buy in the game), but the best risk-and-upside profile belongs to Morgan Rogers ahead of Chelsea’s home game with Brighton. We set all of that out on Wednesday in our full deadline guide, along with differentials, budget picks and the whole fixture list.
Before you hit confirm
- Check Sarr — if you own him, he is not playing.
- Do not sell a highly rated player after one quiet match.
- Do not buy anyone because he rose by £0.1m.
- Captain and vice-captain from different fixtures.
- Remember Crystal Palace v City kicks off at 8pm on Friday, after the deadline.
Every rating is in our AI Player Rankings, and the captaincy shortlist with risk and upside is in Captain Picks. You can also join our mini-league — code 3ix1cw, currently 63 teams.
Our ratings are generated by an AI model from official FPL data, refreshed hourly. The numbers here are from Thursday 27 August in the morning — ownership, prices and flags move during the day, so check again close to the deadline.