Deasons XI

Price arrows: a solid triangle is a change the collector's price feed projects within three nights (the number is how many nights away); a faint one is drift in that direction. A rise before you buy costs a step of budget; a fall before you sell costs a step of what you get back. Start? overrides the minutes model with your own read: no prices him as a substitute (at most 1 point), yes as his if-he-starts number, and the XI is re-picked. On the board, tap a player then another to swap them; tap a selected XI player again for the captain option. Trial transfers from the Alternatives table; the card at the top shows what changes. Next = expected points this gameweek from the allocation model: your share of your club's market-anchored goals and assists, clean sheets and defensive contributions from the same fixture odds, bonus from simulated BPS, all weighted by the minutes model. 5 GW sums the next five on the same model. Start is P(start) from your last team sheets (the most recent counts most); a purple arrow is the lineup rule's shadow figure - it sees injuries, returns and slot competition, and is scored but not yet used. Run shows each opponent (shaded = home) with that week's expected points.

Everyone the brief priced, ranked by five-gameweek expected points, same columns as your squad. Score ours / mkt compares our probability the player scores with the bookmakers' anytime-scorer price (void if he does not play, so it is conditional on playing). Where we are far above the market, be suspicious of us; where the market is far above us, look at set pieces we cannot see.

Each rival's squad is the fifteen he held at the last finished gameweek - his transfers for this one are private until kick-off, and so is any chip. Every squad is picked and captained the same way yours is; Next GW and Five GW are the model's expectation for that squad if it does not change - best eleven and best captain picked fresh each week. In five GWs = points now plus the five-week expectation; Rank then is the table that implies. A gap of 18 points is about a captaincy hit and a clean-sheet weekend - not a rebuild.

Squads as held at the GW5 deadline; transfers for the coming gameweek become visible once the deadline passes. Both xP columns assume the best eleven and captain.

Every squad is picked and captained the same way yours is - best legal eleven by next-gameweek expected points, five gameweeks summed with a fresh best eleven and captain each week, no chips, no hits. Names are matched to FPL web names; a name that matches nobody is listed so you can fix it. Your row includes any trial transfers you have on.

Expected goals for each side from a Dixon-Coles rating of ten seasons (180-day half-life, promoted-team prior) anchored to the live sharp-book match odds where priced. Clean sheet is exp(-xG against). Model only means no book had priced the match at build time.

The parts of the model that are being tested rather than trusted. The lineup rule runs in shadow: its P(start) is shown, frozen with the projections and scored when the gameweek closes. The market columns are diagnostics, not inputs. The last scored gameweek is the only line here the model did not produce itself.

Last scored gameweek. GW5: 652 players, points bias -0.15, MAE 1.12 · P(start) Brier 0.077 · rule column 0.070 · your XI 59 projected, 70 scored (Haaland C)