Why the Current FPL Leader Might Not Be the Favourite

I built a Monte Carlo simulation for the top 1,000 FPL managers in the world, testing who’s most likely to win the whole thing from here. The model accounts for form trajectory, squad overlap, chip usage, scoring variance, and the fact that when Bruno hauls, 94% of the top 1,000 haul with him. It’s the most thorough attempt I can make at answering a simple question: does the current leaderboard reflect who’s actually most likely to finish first?

The honest answer is that it depends on what you think chips are worth.

Under baseline assumptions, Ted Gondwe from 4th edges it at 16.6% win probability over Erik Ibsen’s 14.6% from 1st. Halve the chip bonuses and Gondwe pulls away to 21.2%. Boost them 50% and Ibsen takes the lead at 17.2%. Head-to-head across all baseline sims, Ibsen finishes above Gondwe 54.1% of the time. Gondwe above Ibsen 45.9%. A coin toss with a slight lean, and it took me, (more than id like to admit) model variants to admit just for something that basically shrugs.

How the Model Works

Each manager’s scoring profile blends their recent form (last 6 GWs at 60% weight) with season-long averages (40%) into a single number. The simulation resamples from each manager’s score history, overweighting recent weeks because what someone scored last Tuesday tells me more than what they scored in September. Since 94% of the top 1,000 own Bruno and 93% own Gabriel, scores aren’t independent: I measure squad overlap to give each manager a correlation coefficient (rho), then generate correlated simulations. It sounds grander than it is, but the point is simple: Bruno points are not being rolled independently for 1,000 managers.

Chip bonuses get drawn randomly from a calibrated distribution, which is comfortably the weakest part of the whole model because real managers time their chips strategically, not randomly. Just like the top 100, I’m waiting to see if City have a double week. After all simulated seasons I rank every manager in every run and the win probability is just: how often did this person finish first?

The Power Rankings

Rank	Manager	Style	Pts	Avg	Win%	Top 3	Top 10	E[Rank]	Chips
4	Ted Gondwe	Rising Form	1923	74.8	16.6%	33.7%	58.7%	15.0	FH
1	Erik Ibsen	Rising Form	1926	70.9	14.6%	34.8%	66.5%	10.6	WC, FH, BB, TC
9	Jack Chapman	Fading Star	1906	73.1	8.5%	20.6%	45.3%	19.8	WC, FH, TC
2	Luka Ortulan	The Metronome	1925	67.8	5.8%	18.4%	49.8%	15.2	WC, FH, BB, TC
3	Ian Foster	Fading Star	1923	67.3	4.5%	13.8%	39.6%	21.1	WC, FH, BB
6	Bror Martin Juhasz Fossum	Rising Form	1911	69.1	4.0%	11.7%	30.2%	29.3	WC, FH, BB
5	Joe Dean	Fading Star	1913	69.8	3.9%	11.9%	34.9%	23.8	WC, FH, BB
14	GUNNER ZS	Rising Form	1900	71.3	3.0%	11.1%	36.6%	21.9	WC, FH, BB, TC
35	Kalem Boyle	Rising Form	1887	72.8	3.0%	8.2%	22.1%	36.4	WC, FH, BB
90	Ewan McNeice	—	1874	—	2.9%	6.6%	15.2%	52.6	—

Avg = form-weighted blended mean (60% last 6 GWs, 40% season). 5,000 correlated sims (rho~0.62). Chips reset at GW19.

Gondwe’s 16.6% from 4th and Ibsen’s 14.6% from 1st are close enough that the gap falls within the noise margin of any plausible model variant. But look at the Top 3 column. Ibsen at 34.8% is actually slightly more likely to finish in the top 3 than Gondwe at 33.7%, and his expected final rank of 10.6 beats Gondwe’s 15.0. The win probability gap comes entirely from Gondwe’s higher ceiling, not a better floor, and that distinction matters a lot more than the headline numbers suggest.

Jack Chapman at 9th is the name I keep coming back to. His 73.1 blended average is second only to Gondwe, and at 8.5% win probability he’s the only other contender above 5%. The model classifies him as Fading Star because his trend has dipped recently but his form-weighted average tells you his recent weeks were still elite, just coming down from an even higher peak. Fading from extraordinary to merely excellent still leaves you with the second-best scoring rate in the top 1,000, three chips in hand, and a 20.6% chance of finishing top 3.

The Template

Six players appear in more than 60% of the top 1,000 starting XIs and they’re the reason everyone’s scores move together.

Bruno Fernandes at 94.3%, Gabriel at 93.0%, Haaland at 79.4%, Semenyo at 68.6%, Igor Thiago at 64.8%, Van Dijk at 63.5%. After that it drops off: Ekitike at 49.6%, Joao Pedro at 48.1%, Rice at 43.8%, Raya at 40.0%, Timber at 36.3%. The title race isn’t about those six. It’s about the other five slots.

Squad DNA

Gondwe runs 7/11 template at rho 0.64: Raya, Van Dijk, Gabriel, Senesi, Bruno, Semenyo, Palmer, Sarr, Ouattara, Haaland, Joao Pedro. Four differentials: Senesi, Palmer, Sarr, Ouattara. He sold Rice for Ouattara in GW28 and brought in Sarr in GW26, both moves that traded template safety for attacking upside. Captaining Palmer in GW27 when almost everyone else went Joao Pedro or Haaland is genuinely bold at this level, and I respect it even though I would never have the nerve to do it myself. Haaland, please start scoring again.

The wildcard, bench boost, and triple captain are already played, leaving just the free hit. That’s a man who’s committed to the squad he has.

Ibsen runs 8/11 template at rho 0.73: Pickford, Tarkowski, Gabriel, Van Dijk, Rice, Semenyo, Bruno, Szoboszlai, Igor Thiago, Joao Pedro, Haaland. Three differentials, all relatively safe: Pickford, Tarkowski, Szoboszlai. His GW28 was a triple transfer bringing in Hill, Szoboszlai and Semenyo. All four chips unused since the GW19 reset. A wildcard, free hit, bench boost, and triple captain sitting in the bank while leading the world. I don’t know these people personally and this is pure speculation but Ibsen either has extraordinary patience or confidence in his game plan.

Foster at 8/11 template, rho 0.73: Raya, Van Dijk, Gabriel, Guehi, Rice, Bruno, Morgan Rogers, Anthony Gordon, Haaland, Igor Thiago, Joao Pedro. Three differentials: Guehi, Rogers, Gordon. Captaining Haaland four of six recent weeks while the field shifted towards Bruno. Level on points with Gondwe but playing a completely different game, one that prioritises stability over variance. Three chips in hand: wildcard, free hit, bench boost.

Gondwe and Ibsen share only 6 of 11 players: Bruno, Gabriel, Haaland, Semenyo, Van Dijk, Joao Pedro. That 5-player gap is enormous at this level and it’s where most of the simulation’s win probability difference comes from.

The Robustness Test

I don’t trust a single model configuration to tell me the truth. There are at least four modelling choices baked into the baseline that could reasonably go a different way, and I wanted to know whether the headline finding survives if you turn those dials. So I ran 13 different configurations across two field sizes: 7 plausible scenarios that represent defensible alternative assumptions, and 6 stress tests that push the model to its boundaries.

The four dials:

• Recency: how much to oversample recent form (0x to 6x)
• Fixtures: whether to adjust for fixture difficulty (0 or on)
• Chips: how valuable chip bonuses are (0x to 2x)
• Correlation: how correlated managers’ scores are (70% to 130% of measured)

#	Scenario	Favourite	Win%	#1 Win%
1	Baseline	Ted Gondwe	16.6%	14.6%
2	Mild recency	Erik Ibsen	14.7%	14.7%
3	Baseline + fixtures	Ted Gondwe	16.3%	14.4%
4	Mild recency + fixtures	Erik Ibsen	14.5%	14.5%
5	Conservative chips	Ted Gondwe	21.2%	11.8%
6	Optimistic chips	Erik Ibsen	17.2%	17.2%
7	Mild + fix + opt chips	Erik Ibsen	16.4%	16.4%

Stress tests (boundary conditions, not used for classification):

#	Scenario	Favourite	Win%
8	No recency at all	Erik Ibsen	12.7%
9	Extreme recency (6x)	Ted Gondwe	27.9%
10	No chip bonuses	Ted Gondwe	25.2%
11	Low correlation (0.7x)	Ted Gondwe	15.4%
12	High correlation (1.3x)	Ted Gondwe	19.5%
13	Bear case (no recency + fixtures + 2x chips)	Erik Ibsen	15.4%

Gondwe is the favourite in 3 of 7 plausible scenarios. Ibsen in 4 of 7. In all 7 the gap between them never exceeds 4 percentage points, and the same 5 contenders appear in the top 5 across every plausible scenario. The ordering shuffles but the names don’t.

The Chip Pivot

The single biggest sensitivity in the model is chip valuation. Strip chips out entirely (scenario 10) and Gondwe surges to 25.2% against Ibsen’s 8.9%, a gap so wide it looks like a different race. Scale chip bonuses to 150% (scenario 6) and Ibsen takes the lead at 17.2%. That entire swing, from a 16-point Gondwe advantage to an Ibsen lead, happens because Ibsen holds four chips and Gondwe holds one.

Each chip bonus gets drawn randomly from a calibrated distribution. In reality Ibsen won’t play his triple captain on a random week. He’ll wait for a double gameweek or a fixture swing that screams opportunity. The model undervalues his chip arsenal because it can’t simulate that kind of patience, which means the true picture leans closer to Ibsen than the baseline 14.6% suggests.

Conservative chip assumptions (scenario 5) push Gondwe to 21.2% because halving the bonus doesn’t just reduce the upside for Ibsen. It reduces it for every four-chip manager in the field. Gondwe, with just a free hit, barely notices. That asymmetry is the entire race condensed into a single parameter.

Head-to-Head

In the baseline simulation, Ibsen finishes above Gondwe 54.1% of the time. Gondwe above Ibsen 45.9%.

Almost exactly a coin toss. Gondwe’s higher win percentage comes from the tail: when he finishes high, he finishes higher than Ibsen’s best outcomes because his scoring variance produces more first-place finishes. Ibsen is more likely to finish above Gondwe on any given simulation run, but Gondwe is more likely to finish first overall. That’s a genuinely strange dynamic and I spent a while staring at it before it made sense to me.

Gondwe beats Ortulan 56.7% of the time, Foster 63.0%. Ibsen beats Ortulan 62.2%, Foster 68.8%. The model’s centre of gravity sits with Gondwe and Ibsen, even if Chapman at 8.5% is still close enough to matter.

Why Correlation Changes Everything

The average pairwise correlation between gameweek scores in the top 1,000 sits at 0.62, measured from 8,575 manager pairs. When 94% of the top 1,000 own Bruno and 93% own Gabriel, a haul from either lifts all boats equally. The only thing separating managers is the differential edge: the Cole Palmers the Dango Ouattaras, the captaincy calls that deviate from the herd.

I varied this correlation by +/-30% (scenarios 11 and 12). At low correlation (0.7x), Gondwe’s advantage narrows to 15.4% vs 12.8% because more independent scoring gives every contender a better shot at a wild outlier run. At high correlation (1.3x), Gondwe strengthens to 19.5% vs 17.6% because template managers move in tighter lockstep and only genuine differentials create separation. Neither direction changes who the favourite is. That’s the strongest evidence I found that Gondwe’s edge isn’t just something my correlation settings or my mind invented.

The model decays correlation at 3% per remaining gameweek because squads diverge through transfers. I did try and build a transfer model (do they buy on form/price movement etc.) but I cant understand my transfers half the time so why would I try get in the minds of the top 1,000. Managers still holding a wildcard decay faster at 6% per week, which includes Ibsen. His rho of 0.73 today could drop to around 0.42 by GW38 if and when he wildcards.

The Archetypes

I classify every manager in the top 1,000 from four signals: form trajectory, scoring variance, transfer hit frequency, and consistency. Eight profiles emerge, from Fading Stars (230 managers, declining trend) to Steady Hands (70, consistent and quiet). I’d probably classify myself as Boom or Bust, except I’d need to have had a boom at some point for that to apply.

The title race is a battle between two Rising Form managers with completely different philosophies. Gondwe’s slope is +0.30 with a cv of 0.31, the highest blended average in the entire top 1,000 at 74.8 paired with genuinely high variance. Ibsen’s slope is +0.17 with a cv of 0.26, improving but tightly controlled. Foster at 3rd is the cautionary tale: a Fading Star with slope -0.64, his 67.3 blended average compounds to roughly a 68-point deficit over 9 remaining gameweeks. That’s why the model gives him 4.5% despite being level on points.

Dark Horses

Ewan McNeice at 90th is the most improbable contender in the race. His 2.9% win chance from 90th in the world is higher than 14 managers ranked above him, including 7th-placed Janne Ojanpera at 1.5%. Only 5 of 11 template players, a rho of 0.46, and differentials that read like a parallel FPL universe: Palmer, Mbeumo, Sarr, Szoboszlai, Pickford, James Hill. Median finish 54th, but his best 5% of sims land top 7. Kalem Boyle at 35th is the other long-range threat, his 72.8 blended average ranking 3rd in the top 100 with three chips still in hand.

What This Model Can and Cannot Tell You

The fixture adjustment layer is deliberately blunt: a squad-level FDR average that adjusts weekly scores by roughly 0-1 point per GW, enough to capture the broad advantage of easy fixture runs without pretending to model individual matchups. Adding fixtures barely shifts the results (scenario 3 vs scenario 1: Gondwe 16.3% vs 16.6%), which tells you either the fixtures are genuinely balanced across contenders or my adjustment is too modest to matter.

The deeper limitation is that the model assumes managers keep their current squads for 9 weeks, modified only by correlation decay. Real managers react. Gondwe will sell a differential if it blanks twice, Ibsen will wildcard the moment he spots an opportunity. I can’t simulate reactive management and that gap matters most for Ibsen, whose four chips represent flexibility the simulation treats as random bonuses rather than strategic options.

The recency sensitivity is the single most important finding for how literally to take any of these numbers. At mild recency (1x oversampling instead of 2x), Ibsen becomes the favourite. At extreme recency (6x), Gondwe surges to 27.9%. The baseline sits at 2x, which is a judgement call on my part, not a mathematical fact. Recent form matters more than season averages with 9 weeks left but exactly how much more is genuinely unknowable.

The Verdict

Classification: Mixed. The current leader is not the simulation favourite in 3 of 7 plausible scenarios. The same 5 contenders surface across every configuration. The leaderboard overstates how secure Ibsen’s position is, but no single challenger dominates either.

The title won’t be decided by the six players everyone owns. It’ll be decided by the other five slots, when you play your chips, and whether you believe recent form tells you more than a full season of data.

Ibsen’s got first place, four chips, and the standings on his side. Gondwe’s got Palmer, Ouattara, Sarr, a free hit, and the last six weeks of form on his. The model leans Gondwe by 2 percentage points and changes its mind if you turn one dial. I did the maths. I’m still in Coal.

FPLCore Insights. 5,000 correlated simulations. 13 model variants. rho~0.62. Top 1,000 managers. After GW29. Chips reset at GW19.