The EV curve tells the story
The wildcard firing window is the single highest-leverage chip-spend decision in FPL. The Onside model's historical EV curve makes the right answer obvious at a glance — a steep climb from GW1 to GW4 (the peak), a gentle decay from GW4 to GW6, and a sharp decline from GW7 onward.
GW1: -3 points (using WC1 before any real data is net-negative) GW4: +47 points (post-international-break, info-rich, fixture swing visible) GW10: +25 points (more than half the chip's potential value lost) GW20: +8 points (effectively a free transfer at this point)
The right framing isn't "early or late" — it's "early, peak, or late." Peak (GW4-6) wins 70%+ of seasons. Early (GW1-3) wins only in injury/squad-decay scenarios. Late (GW10+) wins only when the alternative is doing nothing.
3 conditions that justify an early wildcard
Condition 1 — 4+ first-XI players have minutes-confirmed-cut OR injured for >3 GWs. Hard structural decay that won't self-correct via free transfers + 1-2 hits.
Condition 2 — You drafted before fixtures release and ended up with 3+ first-XI players facing top-six opposition in GW1. The fixture-mismatch is too wide to close with single transfers before the international break.
Condition 3 — The 26/27 template consolidated and you own 5 or fewer of the top-12 template names. Structural template miss = wildcard reset is cheaper than 4-5 weeks of hits trying to catch up.
If you don't hit any of these three conditions, the model strictly prefers holding for GW4-6.
Why late wildcards lose 22+ points/season
Every gameweek after GW6 you sacrifice 4-6 points of chip EV. By GW10 you've lost roughly 22 points vs the GW4 baseline. By GW15 the chip is worth ~30 points less than if you'd fired at the EV peak.
The biggest late-wildcard mistake: holding the chip "for a fixture swing" that never quite arrives. The model's view: take the GW4-6 EV peak. The chip's value compounds — every week you delay, you carry suboptimal players who underperform projection, and you forfeit the price-change window.
The exception: managers who plan to use Wildcard 2 (GW20-29 firing window) AND don't need WC1's reset. Some second-half-focused managers prefer a single wildcard concentrated in the second half. That's a valid strategy with different EV math — but it requires holding both chips, not just delaying.
The model's GW4 firing prescription
If you're reaching GW4 with no clear injury crisis and no structural template miss, the model recommends firing Wildcard 1 in the GW4 deadline window with the following skeleton:
GK: Roefs (£4.5m) + £4m rotation DEF: Gabriel + Robinson + Tarkowski + 2x rotating budget MID: Bruno + Saka + Palmer (if Alonso-bounce confirmed) + Rogers + budget bench FWD: Haaland (captain) + Watkins + Thiago
The £100m budget assumes 26/27 inflation lands at +£3-4m vs 25/26 closing prices. The structural anchors (Haaland + Bruno + Saka + Gabriel) are non-negotiable in 90%+ of fixture configurations through GW8.
GW5 wins over GW4 when there's a coordinated 3+ club fixture swing visible. GW6 wins over GW4 when premium-mid clarity (Bruno vs Saka vs Palmer) shifts in the 2-week window. Otherwise GW4 is the firing prescription.