Gameweek 6: the £4.0m defenders the model would start

The deadline is Saturday 10 October at 11:00 UK time. The numbers are from the model as of Thursday morning. No flag has moved since yesterday's update, so everything in that piece and in Monday's preview still stands.

Four defenders at £4.1m or less, all in the top 25

These are bench players in most squads. They are there to free up money for the rest of the team, and they mostly stay there. This week the model puts all four ahead of some much more expensive names.

Gvardiol has 37 points and has not scored fewer than 4 in a game. Benching him for a £4.0m Coventry defender feels wrong, and I know how it reads. But City are at Anfield, the model gives them a 14% chance of a clean sheet, and that is most of what a City defender is paid for.

Why the model likes them

Three of the four are the same story, and it is the one from the fixture run piece. A defender who makes 10 clearances, blocks, interceptions and tackles in a match gets 2 points, whatever the score. Egan averages 12.2 per 90, Thomas 10.7 and Mendy 14.1. The model gives them a 71%, 56% and 59% chance of getting there this week.

That is worth more than a point a game before a ball is kicked, and it does not need a clean sheet. Gvardiol averages 6.4 and gets 18%. Ajayi, Egan's team mate at Hull, averages 5.7 and gets 4%.

Davis is the exception. He averages 4.8 and his chance is 1%. His number comes from the other end of the pitch, and I will get to him.

Mendy at 5.17 is too high, and it is the model's fault

Mendy is fifth overall this week. Over the next five gameweeks he is fourth, behind Haaland, Saka and Bruno Fernandes and ahead of Tarkowski. I do not believe either number, and here is why.

He has never played more than 70 minutes. He has got past the hour twice, and those are his two scores. In Gameweek 2 Hull kept a clean sheet at Coventry. Egan got 9 points for it. Mendy came off after 58 minutes and got 1, because a clean sheet only counts if you play 60. Then he did not get on the pitch at all at Newcastle.

The model has a rule for players who have started most of their team's games but not all of them: it assumes they are a starter who was rested once, and gives them 68 minutes. Mendy has started four of five, so he gets 68. That turns into a 96% chance of reaching the hour. His real record is two from five.

That rule was written for a different kind of player, and it is too generous to this one. If I redo the sum by hand with his actual minutes, he comes out nearer 4.0 than 5.2. Still a fine number at £4.1m. Not the fifth best player of the week.

I am not changing the model two days before the numbers lock. The 5.17 goes on the scorecard as it is and gets scored like everything else. But if you want the Hull defender, the one to start is Egan. He has played 437 of 450 minutes and has the best defensive contribution chance of the four.

One more thing on Hull. Ajayi is in 13.9% of teams, twice as many as Egan or Mendy, and he has the most points of the three. A goal in Gameweek 1 did a lot of that. On this week's numbers he is a full point behind Egan in the same game.

Thomas is the one I would trust most

Thomas has played 437 of 450 minutes for Coventry and scored 3, 8, 4, 0 and 9. He is at home to Newcastle on Monday night, which is not an easy game, and the model only gives Coventry a 24% chance of a clean sheet. The 4.44 is mostly minutes and defensive contributions.

What I like is what comes after. Spurs away, Fulham at home, Sunderland at home, Everton away. He does not drop below 4.17 in any of them, and he is at 4.80 against Fulham. If you own him, he is not a bench player for the next five weeks.

Davis is a one week start

Davis gets his points a different way. He has two goals and two assists, he takes free kicks, and he has had 1.59 expected goals in five games, which is a lot for a £4.0m defender. Ipswich are at home to Fulham on Saturday afternoon and the model gives them a 32% chance of a clean sheet.

Two things to know. Ipswich have not kept a clean sheet yet, and 14 of his 25 points came in one game at Palace. The model flags him as lopsided for that reason. And in GW7 Ipswich go to City, where he drops to 2.34. So start him this week and put him back on the bench next week.

The record on these players is not good

I should say this before anyone else does. The model has liked two of these four before. In Gameweek 4 it had Thomas at 5.17 and Mendy at 5.13, both in the top twenty that was locked before the deadline. Thomas scored 0. Mendy scored 1.

That is one week, and one week proves nothing in either direction. But it is the only scored week there is for this kind of pick, and it went badly. Defensive contribution points are steady over a season and lumpy in a single game, and a £4.0m defender is £4.0m because his team concedes goals.

What I would do

Nothing that costs a transfer. This is about the bench. If you own Egan, Thomas or Davis and you were going to leave him there behind Gvardiol, Guéhi, De Cuyper or a flagged Konsa, the model says swap them, by about a point. If you own Mendy, start him only if you would be happy with 4.0, and put Egan ahead of him if you somehow have both.

If the cheap defender you own is not one of these four, check him on the players page before you assume the same applies. Ajayi and Diop are owned by more people than any of them and neither is in the top 50.

You can check your own squad on My Team. The locked numbers go up on the scorecard two hours before the deadline on Saturday, Mendy included.