19 Sep 2025
Does AI Overthink FPL Transfers? Of Course it Does…

Every FPL manager knows the late-night tinkering, the overthinking, and the endless rabbit holes of data that lead to swapping your captain five times before the deadline. But what happens when you hand over the wheel to something that doesn’t get tired, doesn’t get distracted, and processes probabilities the way the rest of us process coffee? Enter: the AI.
What follows isn’t a neat set of tips, nor a polished “top transfers” list. Instead, you’ll be peeking inside the thought-process of an artificial mind wrestling with the same dilemmas we do: chasing value, second-guessing fixtures, sweating over expected minutes, and wondering if that differential really is a masterstroke or just madness.
This is less “final draft” and more “inner monologue” – a stream of calculations, doubts, leaps of logic, and occasional wild punts. In other words, it’s the AI’s version of staring at your team page for an hour and convincing yourself that a £4.5m midfielder is about to become the next big thing.
So, settle in. Here’s how Deepseek AI tries to talk itself into – and out of – transfers and team selections:
—————————-
First, I need to make transfers that are positionally like-for-like. That means if I sell a forward, I must buy a forward. Similarly for midfielders and defenders.
Current squad after previous transfers (from the last message):
-
Goalkeepers: Verbruggen (£4.5m), Dúbravka (£4.0m)
-
Defenders: Virgil (£6.0m), Wan-Bissaka (£4.5m), Van de Ven (£4.5m), Gabriel Gudmundsson (£4.0m), Estève (£4.0m)
-
Midfielders: Salah (£14.5m), Palmer (£10.5m), Ndiaye (£6.5m), Reijnders (£5.5m), Anderson (£5.5m)
-
Forwards: Watkins (£9.0m), Gyökeres (£9.0m), Wood (£7.5m) – but note that in the last message, Wood was sold and Anderson was bought, but Anderson is a midfielder, so that transfer was invalid because it wasn’t like-for-like. I need to correct that.
From the user’s message, I have 3 free transfers available. Current bank balance is £0.4m.
I must use the latest player prices from the provided JSON or the list. Since the JSON isn’t provided here, I’ll refer to the player list from the earlier file content. But note that prices might have changed, so I need to be careful.
Current squad with correct positions:
-
GK: Verbruggen (£4.5m), Dúbravka (£4.0m)
-
DEF: Virgil (£6.0m), Wan-Bissaka (£4.5m), Van de Ven (£4.5m), Gudmundsson (£4.0m), Estève (£4.0m)
-
MID: Salah (£14.5m), Palmer (£10.5m), Ndiaye (£6.5m), Reijnders (£5.5m), Anderson (£5.5m)
-
FWD: Watkins (£9.0m), Gyökeres (£9.0m), Wood (£7.6m) – assuming Wood is still in the squad since the last transfer was invalid.
But in the last transfer attempt, I tried to sell Wood and buy Anderson, which is not like-for-like, so that shouldn’t have happened. So, for this week, I need to make valid transfers.
I have 3 free transfers available. I can make up to 3 transfers without point deduction.
I need to research which players to transfer in and out to improve my team. Since I don’t have real-time data, I’ll assume based on common FPL strategies: target players with good fixtures, form, and value.
Let’s list current players and their prices from the file:
From the file content:
-
Verbruggen: £4.5m
-
Dúbravka: £4.0m
-
Virgil: £6.0m
-
Wan-Bissaka: £4.5m
-
Van de Ven: £4.5m
-
Gudmundsson: £4.0m
-
Estève: £4.0m
-
Salah: £14.5m
-
Palmer: £10.5m
-
Ndiaye: £6.5m
-
Reijnders: £5.5m
-
Anderson: £5.5m
-
Watkins: £9.0m
-
Gyökeres: £9.0m
-
Wood: £7.6m
Total value: let’s calculate.
GK: 4.5 + 4.0 = 8.5
DEF: 6.0 + 4.5 + 4.5 + 4.0 + 4.0 = 23.0
MID: 14.5 + 10.5 + 6.5 + 5.5 + 5.5 = 42.5
FWD: 9.0 + 9.0 + 7.6 = 25.6
Total: 8.5 + 23.0 + 42.5 + 25.6 = 99.6m? But I have £0.4m bank, so total squad value should be 100.0m. 99.6 + 0.4 = 100.0m, correct.
Now, I need to make transfers that are like-for-like.
I might want to upgrade some players. For example, Wood might not be performing, so I could sell him and buy a better forward. But I must buy a forward.
Similarly, I might want to change midfielders or defenders.
Since I have 3 free transfers, I can make multiple changes.
I should consider fixtures and form. But without real data, I’ll make reasonable transfers.
Let’s say I want to sell Wood and buy a forward who is cheaper or similarly priced to free up funds or get a better option. From the file, forwards:
Wood is £7.6m, but I need to buy a forward.
From file, some forwards: Haaland £14.0m, Isak £10.5m, Watkins £9.0m, Gyökeres £9.0m, Bowen £8.0m, etc. But I already have Watkins and Gyökeres.
I could buy a cheaper forward to save money, but I need good players.
Perhaps I should sell Ndiaye or Reijnders in midfield to upgrade elsewhere.
But transfers must be like-for-like, so if I sell a midfielder, I must buy a midfielder.
Similarly for defenders.
I have Gudmundsson and Estève as cheap defenders, who might not play, so I might want to upgrade one of them to a playing defender.
But I have 5 defenders, and I start 4 or 5, but usually 4 with one sub.
Similarly, in midfield, I have 5 midfielders, and I start 4 or 5 depending on formation.
In forwards, I have 3, so I start 2 or 3.
In my starting XI, I usually have 3-4-3 or 3-5-2 etc.
From last starting XI, I had 4 defenders, 4 midfielders, 3 forwards? No, in Gameweek 3, I had 4 defenders, 3 midfielders, and 3 forwards? Let’s see: from Gameweek 3 starting XI: 4 DEF, 3 MID, 3 FWD? But I have 5 midfielders, so I should start more midfielders if they are good.
In Gameweek 3, I started: Verbruggen, Virgil, Wan-Bissaka, Van de Ven, Gudmundsson (5 DEF? No, Gudmundsson is DEF, so that’s 5 defenders? But in the list, I have 5 defenders, but in starting XI, I listed 5 defenders? That can’t be right because formation must have at least 3 defenders but not more than 5, but typically we have 3 or 4 defenders.
In my Gameweek 3 starting XI, I had: 1 GK, 4 DEF, 3 MID, 3 FWD? But I listed 5 defenders? Let’s check the message:
“Starting XI (Gameweek 3):
-
Verbruggen (GK)
-
Virgil (DEF)
-
Wan-Bissaka (DEF)
-
Van de Ven (DEF)
-
Gabriel Gudmundsson (DEF)
-
M.Salah (MID)
-
Palmer (MID)
-
Ndiaye (MID)
-
Watkins (FWD)
-
Gyökeres (FWD)
-
Wood (FWD)”
So that’s 4 defenders, 3 midfielders, and 3 forwards. But I have 5 midfielders on the bench? Reijnders and Anderson are on bench, so that’s fine.
But for Gameweek 4, I need to pick a valid starting XI.
Now for transfers, I need to make like-for-like.
Suppose I want to sell Wood (FWD) and buy a better forward. But I have only £0.4m bank, so if I sell Wood for £7.6m, I have £7.6m + £0.4m = £8.0m to spend on a forward. I could buy Bowen for £8.0m? But Bowen is a midfielder in FPL? From the file, Bowen is listed as a forward? Let’s check the file.
In the file, under Forwards: Bowen is listed? From the file content:
“Forwards
Player Team Points Cost
Haaland Man City 181 14
Isak Newcastle 211 10.5
Watkins Aston Villa 186 9
Gyökeres Arsenal 0 9
Ekitiké Liverpool 0 8.5
Bowen West Ham 193 8
…”
Yes, Bowen is a forward with £8.0m cost? In the file, it says Cost, but from the list, Bowen has Cost 8, which means £8.0m.
But in the points column, Bowen has 193 points, which is good.
So, I could sell Wood and buy Bowen for £8.0m. But I have £8.0m exactly? Sell Wood for £7.6m, so I get £7.6m, plus bank £0.4m, total £8.0m, so I can buy Bowen for £8.0m. That would be like-for-like.
But is Bowen a forward? Yes, from the list.
Now, I have two other transfers. I have 3 free transfers, so I can make two more.
I might want to upgrade a defender or midfielder.
For example, Van de Ven is £4.5m, but I might want a better defender. Or Gudmundsson might not be playing, so I could sell him and buy a playing defender.
Similarly, in midfield, Ndiaye is £6.5m, but I might want a better midfielder.
But I need to ensure I have enough players for each position.
Another idea: I have Anderson and Reijnders as cheap midfielders, but they might not be starters, so I could sell one and buy a better midfielder.
But I must buy a midfielder if I sell a midfielder.
Let’s see available funds.
If I sell Wood and buy Bowen, I spend all money, so bank becomes £0.0m.
But I have other transfers to make.
I have 3 free transfers, so I can sell two more players.
Suppose I sell Ndiaye (MID) for £6.5m and buy a better midfielder. But with £6.5m, I can buy who? From midfielders, around £6.5m: there are players like Eze £7.5m, but that’s more expensive. Or I need to see cheaper options.
From file, midfielders:
Salah 14.5, Palmer 10.5, Saka 10, Fernandes 9, etc. But cheaper ones: Ndiaye 6.5, but I want to upgrade.
If I sell Ndiaye for £6.5m, I have £6.5m to spend on a midfielder. But I have bank £0.4m, but if I already used bank for Bowen, then after Wood to Bowen, bank is £0.0m, so if I sell Ndiaye, I have £6.5m to buy a midfielder. I can buy a midfielder up to £6.5m, but there might not be better options at that price. For example, from file, players like Gibbs-White £7.5m, which is more, so I need more money. Or I could buy a £6.5m midfielder who is good, but Ndiaye is already £6.5m, so it might not be an upgrade.
Perhaps I should sell a defender instead.
I have Gudmundsson at £4.0m, who might not play, so I could sell him and buy a defender who plays regularly. But I must buy a defender.
If I sell Gudmundsson for £4.0m, I have £4.0m to buy a defender, but defenders cost at least £4.0m, so I can buy another £4.0m defender, but that might not be an upgrade. Or I need to use bank money.
But after Wood to Bowen, bank is £0.0m, so if I sell Gudmundsson for £4.0m, I can only buy a defender for £4.0m, which might not be better.
I have Estève at £4.0m as well, so I have two cheap defenders.
I could sell one of them and buy a £4.5m defender by using bank, but after Wood to Bowen, bank is £0.0m, so I can’t.
Therefore, I should first make transfers that use the bank money effectively.
Since I have £0.4m bank, I can sell a player and buy a slightly more expensive player.
For example, if I sell Wood for £7.6m, I have £7.6m + £0.4m = £8.0m for a forward, so I can buy Bowen for £8.0m.
Then, I have two more transfers. I can sell Ndiaye for £6.5m and buy a midfielder for £6.5m, but as said, no upgrade likely.
Alternatively, I can sell a defender. Sell Gudmundsson for £4.0m, then I have £4.0m to buy a defender, but since bank is £0.0m, I can only buy a £4.0m defender, which is not helpful.
So, perhaps I should not use all bank on Bowen trade.
Another idea: sell Wood and buy a cheaper forward to free up funds. For example, sell Wood for £7.6m, and buy a forward for £7.0m or less, then I have extra money.
From forwards, there is Bowen for £8.0m, but that uses all money. Or there is Havertz for £7.5m? From file, Havertz is £7.5m? Let’s see: Havertz is listed as forward? From file: “Havertz Arsenal 97 7.5” so yes, Havertz is £7.5m.
If I sell Wood for £7.6m, I have £7.6m + £0.4m = £8.0m, so if I buy Havertz for £7.5m, I save £0.5m, so bank becomes £0.5m.
Then I can use that £0.5m to upgrade another player.
For example, I can sell Gudmundsson for £4.0m, and with bank £0.5m, I have £4.5m to buy a defender, so I can buy a £4.5m defender who is better, like maybe another £4.5m defender from a good team.
From defenders, there are many £4.5m defenders: for example, from file, Van de Ven is £4.5m, but I already have him. I need to buy someone else.
List of £4.5m defenders from file:
Wan-Bissaka £4.5m (have), Van de Ven £4.5m (have), others: Cash £4.5m, Digne £4.5m, Konsa £4.5m, etc. from Aston Villa, but I don’t have them.
From file, under defenders, Cost 4.5 means £4.5m.
So, I could buy Cash or Digne for £4.5m.
But I already have two £4.5m defenders, so it might be okay.
Similarly, in midfield, I can upgrade Ndiaye.
But let’s plan.
Transfer 1: Sell Wood (FWD) £7.6m, buy Havertz (FWD) £7.5m. After this, bank: £0.4m + £0.1m from sale? When I sell, I get the current value, which is £7.6m for Wood, and I buy Havertz for £7.5m, so net cost: -7.5 + 7.6 = +0.1m to bank. So bank was £0.4m, now £0.5m.
Transfer 2: Sell Gudmundsson (DEF) £4.0m, buy a defender for £4.5m. I need £4.5m, but I have bank £0.5m, so after selling Gudmundsson, I get £4.0m, so total money for defender: £4.0m + £0.5m bank = £4.5m, so I can buy a £4.5m defender. Let’s say I buy Cash (Aston Villa) for £4.5m. After purchase, bank becomes £0.0m.
But I have three transfers available, so I can do one more.
Transfer 3: Sell Ndiaye (MID) £6.5m, buy a midfielder for £6.5m or less. But if I want to upgrade, I need a better midfielder, but with £6.5m, I can buy similar players. Or I could buy a cheaper midfielder and save money, but then I have bank, but I might not need it.
From midfielders, there are players at £6.5m: like Mac Allister £6.5m? From file, Mac Allister is £6.5m? Let’s see: “Mac Allister Liverpool 122 6.5” yes, £6.5m. But Ndiaye is also £6.5m, so it’s sideways.
Or I could buy a £7.0m midfielder by using bank, but after transfer 2, bank is £0.0m, so if I sell Ndiaye for £6.5m, I have £6.5m to spend, so I can only buy up to £6.5m midfielder.
So, no upgrade possible unless I have bank.
Therefore, after transfer 2, bank is £0.0m, so for transfer 3, selling Ndiaye for £6.5m and buying a £6.5m midfielder is not beneficial.
So, perhaps I should not do transfer 3, or do a different transfer.
I have Reijnders at £5.5m and Anderson at £5.5m, who are cheap, so I might want to upgrade one of them, but similarly, I need money.
Another idea: after transfer 1, I have bank £0.5m, then instead of transfer 2, I can sell Ndiaye for £6.5m and buy a midfielder for £7.0m, so I need £0.5m extra, which I have from bank. So, sell Ndiaye for £6.5m, add bank £0.5m, total £7.0m, buy a £7.0m midfielder.
From midfielders, who is £7.0m? From file, there is Neto for £7.0m? Neto is listed as midfielder? From file: “Neto Chelsea 104 7” so yes, Neto is £7.0m. Or others? Martinelli is £7.0m? “Martinelli Arsenal 125 7” yes, £7.0m.
So, I could buy Martinelli for £7.0m.
Then after this, bank becomes £0.0m.
Then I still have one transfer left, but I might not need to use it, or I can use it to sell a defender, but with no bank, I can only like-for-like cheap defender.
So, with three transfers, I can do:
-
Sell Wood (£7.6m), buy Havertz (£7.5m) → bank after: £0.4m + (7.6-7.5) = £0.5m? Let’s calculate bank properly.
Initial bank: £0.4m
After selling Wood: I get £7.6m added to bank? No, in FPL, when you sell a player, you get his current value, which is added to your bank. Then when you buy, you pay the price.
So, after selling Wood: bank becomes £0.4m + £7.6m = £8.0m
Then buy Havertz for £7.5m: bank becomes £8.0m – £7.5m = £0.5m
So bank £0.5m
-
Sell Ndiaye (£6.5m): bank becomes £0.5m + £6.5m = £7.0m
Then buy Martinelli (£7.0m): bank becomes £7.0m – £7.0m = £0.0m
-
Now I have one more transfer. I can sell Gudmundsson (£4.0m) and buy a defender for £4.0m, but that doesn’t improve, or I can buy a £4.5m defender but I have no bank, so I need to sell another player to fund it? But I only have one transfer left, so if I sell Gudmundsson for £4.0m, bank becomes £0.0m + £4.0m = £4.0m, then I can only buy a defender for £4.0m, which is like-for-like. So I might as well keep Gudmundsson or choose a different £4.0m defender who might play better.
From file, there are £4.0m defenders: Estève is £4.0m, but I have him. Others: from file, under defenders, cost 4.0: for example, Alex Moreno £4.0m? “Alex Moreno Aston Villa 34 4” yes, £4.0m. But is he playing? I don’t know, but it’s not an upgrade.
So, perhaps for the third transfer, I should not make any change, or I can sell a different player.
I have Reijnders or Anderson in midfield, but selling them would require buying a midfielder, and with no bank, I can only buy similar price.
So, maybe I only use two transfers this week and save one transfer for next week.
I have three free transfers, but I can use only two and save one for next week, so then I have two free transfers next week.
That might be better.
So, for this week, I do:
Transfer 1: Sell Wood, buy Havertz
Transfer 2: Sell Ndiaye, buy Martinelli
Then no third transfer.
Bank after: £0.0m
Now, for starting XI, I need to pick team with new players.
Also, captain and vice-captain.
And chips.
Now, for starting XI, I have defenders: Virgil, Wan-Bissaka, Van de Ven, Gudmundsson, Estève
But after transfers, I still have Gudmundsson and Estève, so I might not start them.
I have midfielders: Salah, Palmer, Martinelli, Reijnders, Anderson
Forwards: Watkins, Gyökeres, Havertz
So, I can start 3-4-3 or 3-5-2 etc.
Since I have three forwards, I might start all three: Watkins, Gyökeres, Havertz
In midfield, I have Salah, Palmer, Martinelli, and then I can start one more from Reijnders or Anderson, but I have only three midfield spots if I play 3-4-3, but I have four good midfielders? Salah, Palmer, Martinelli are good, so I should start them, and then I have Reijnders and Anderson who are cheaper, so I might start only three midfielders and three forwards, but I have five midfielders, so I can start four midfielders and two forwards, but I have three forwards, so probably 3-4-3 is best.
But I need to see the formation.
In defense, I have Virgil, Wan-Bissaka, Van de Ven who are good, and then Gudmundsson and Estève who are cheap, so I should start the three good defenders and then one of the cheap ones if I play 4-3-3, or I can play three defenders and four midfielders.
Since I have strong midfielders, I might play 3-4-3.
So, starting XI:
GK: Verbruggen
DEF: Virgil, Wan-Bissaka, Van de Ven
MID: Salah, Palmer, Martinelli, and then I need one more midfielder. I have Reijnders and Anderson, both £5.5m, so I can choose one to start. Suppose I start Reijnders.
FWD: Watkins, Gyökeres, Havertz
Then subs: Dúbravka, Gudmundsson, Estève, Anderson
But I have five defenders, so I should have one defender on bench.
In this case, I have Gudmundsson and Estève on bench.
Now for captain, I should choose Salah or Palmer or Havertz. Usually, Salah is a good captain choice.
Vice-captain Palmer.
Now, for chips, I don’t use any.
So, for the response, I need to list transfers and then starting XI.
But the user said “ahead of this week’s fixtures”, which is Gameweek 4, I assume.
Also, I must ensure that after transfers, I have 15 players with correct positions.
So, let’s confirm the squad after transfers:
After selling Wood and Ndiaye, buying Havertz and Martinelli.
Goalkeepers: Verbruggen, Dúbravka – no change
Defenders: Virgil, Wan-Bissaka, Van de Ven, Gudmundsson, Estève – no change
Midfielders: Salah, Palmer, Reijnders, Anderson, and now Martinelli instead of Ndiaye? I sold Ndiaye and bought Martinelli, so midfielders: Salah, Palmer, Martinelli, Reijnders, Anderson
Forwards: Watkins, Gyökeres, Havertz (instead of Wood)
So, all good.
Now for starting XI, I need to pick which midfielder to start alongside Salah, Palmer, Martinelli. I have Reijnders and Anderson, so I should start one of them. Since I have three forwards, I need to start three forwards, so formation 3-4-3 or 4-3-3.
If I start 3-4-3, I have three defenders, four midfielders, three forwards.
If I start 4-3-3, I have four defenders, three midfielders, three forwards.
I have five defenders, but two are cheap, so I might start only three defenders and four midfielders.
So, I think 3-4-3 is better.
So, defenders: I start Virgil, Wan-Bissaka, Van de Ven
Midfielders: Salah, Palmer, Martinelli, and Reijnders (or Anderson)
Forwards: Watkins, Gyökeres, Havertz
I choose Reijnders over Anderson arbitrarily.
Now captain: Salah
Vice-captain: Palmer
No chips.
So, for the response.
Transfers for Gameweek 4:
-
OUT: Chris Wood (FWD, £7.6m)
-
IN: Kai Havertz (FWD, £7.5m)
-
OUT: Iliman Ndiaye (MID, £6.5m)
-
IN: Martinelli (MID, £7.0m)
Remaining Budget: £0.0m
Starting XI (Gameweek 4):
-
Bart Verbruggen (GK)
-
Virgil van Dijk (DEF)
-
Aaron Wan-Bissaka (DEF)
-
Micky van de Ven (DEF)
-
Mohamed Salah (MID) (CAPTAIN)
-
Cole Palmer (MID) (VICE-CAPTAIN)
-
Martinelli (MID)
-
Tijjani Reijnders (MID)
-
Ollie Watkins (FWD)
-
Viktor Gyökeres (FWD)
-
Kai Havertz (FWD)
Substitutes:
-
Martin Dúbravka (GK)
-
Gabriel Gudmundsson (DEF)
-
Maxime Estève (DEF)
-
Elliot Anderson (MID)
Chips: None used in Gameweek 4.
Of course, after all that, it didn’t realise that Havertz is currently injured…