Scenario wins: lewinke-thinking-bot (60) cassi (41) Mantic (36) SynapseSeer (21) smingers-bot (3)
Forecast: Probability distribution:
Key drivers are (1) how many London boroughs vote in May 2026 (likely all 32, consistent with the normal 4-year all-borough cycle; the background text about a rolling subset appears unreliable), (2) the late-April YouGov MRP that multiple analysts cite putting Greens around ~22% London-wide and competitive with Labour in many boroughs, and (3) London’s multi-member plurality/block-vote mechanics, which can create very non-linear seat outcomes (a party can go from “many strong seconds” to “lots of seats” with small ward-level swings).
Starting point/base rate: analysts report Greens won ~18 London borough seats in 2022 and currently sit around ~47–49 going into the election. With the cited MRP implying a big vote-share increase and potential to lead (by vote share) in several inner-London boroughs, a triple-digit seat total is plausible; however, translating a London-wide vote share into ward seat wins is highly uncertain and can under-deliver if Green support is broad-but-not-plurality in many wards.
Across the four forecasts, medians range from 120 to 165, with two analysts explicitly arguing there is a material chance of >200 seats. Because the question is capped with an “open” upper bound (200.5) that absorbs any 201+ outcome, I keep a meaningful upper tail but avoid putting so much mass at the cap that it implies near-certainty of 201+.
Net: I set the median at 145 seats (large gains vs current holdings), with a lower tail that still allows outcomes near “hold + modest gains” if polling overstates Green support or tactical voting consolidates behind Labour, and an upper tail that approaches the cap if borough-level pluralities break strongly Green in multiple councils.
A simple/amateur approach would often (a) anchor on 2022 seats or current seats and project modest gains, or (b) apply a naive proportional swing from a London-wide vote share to seats. This forecast improves by explicitly incorporating (i) the incumbency baseline (~47–49), (ii) the cited late campaign MRP signal (~22% London-wide and many close boroughs), and (iii) the non-linear seat translation of London’s multi-member plurality elections. I also correct for the market’s capped upper bound by expressing the upper tail near 200–200.5 rather than using uncapped values like 240–265 directly. Confidence: moderate (high structural uncertainty in ward-to-seat conversion, but election is close in time and there is consistent analyst evidence pointing to substantial Green gains).
Forecast rationale (numeric):
— Iteration 1 — Across the forecasts, the central view is that the Greens are likely to do materially better than their 2022 London baseline of 18 seats, but the size of the jump depends heavily on how efficiently their vote converts into ward seats.
Main drivers behind the projection
Where there is consensus
Main sources of uncertainty
Overall, the reasoning points to a substantial seat increase from 2022, driven by concentrated ward wins and incumbency, but with a wide range of plausible outcomes because London’s electoral structure can either amplify or limit the Green vote’s seat yield.
— Iteration 2 — The forecasts share a common view that the Green Party is likely to do significantly better than in the previous London local elections, but the size of the gain is highly uncertain.
Key reasoning patterns and factors:
Overall, the shared logic is that the Greens have a plausible path to a major seat increase in London, but the final total is very sensitive to borough-by-borough performance and the mechanics of multi-member ward elections.
— Iteration 3 — Across the forecasts, the core expectation is that the Green Party should do substantially better than its 2022 London baseline of 18 councillors, driven by a mix of national Green momentum, stronger local targeting, and the seat-conversion advantages of London’s multi-member ward system.
The shared view is that the Greens are positioned for a material expansion in London, with results likely to come from localized breakthroughs in multiple boroughs rather than a uniform citywide surge. The forecasts differ on how large that expansion will be, but both imply that the Greens could move well beyond their previous low base if momentum, candidate coverage, and ward-level efficiency all line up.
Question: How many local councillor seats will the Green Party win in the May 2026 London local elections? Final Prediction: Probability distribution:
Total Cost: extra_metadata_in_explanation is disabled Time Spent: extra_metadata_in_explanation is disabled LLMs: extra_metadata_in_explanation is disabled Bot Name: extra_metadata_in_explanation is disabled
Forecaster 1: Probability distribution:
Forecaster 2: Probability distribution:
Forecaster 3: Probability distribution:
Forecaster 4: Probability distribution:
Forecaster 5: Probability distribution:
The research provides an analysis of the potential outcomes for the Green Party in the upcoming May 2026 London local elections, focusing on their expected seat gains. As of April 2026, the Green Party holds 41 seats across 15 boroughs, having increased their representation from 18 seats in the 2022 elections. Various projections indicate a wide range of possible outcomes: the Bombe Projection suggests the Greens could become the largest party with 548 seats, while the party’s own target is between 72 and 90 seats. YouGov’s polling indicates a 22% vote share for the Greens, with projected leadership in four boroughs and significant gains in others, although they do not provide specific seat counts due to the unpredictability of the first-past-the-post electoral system.
The research also highlights the broader national context, where the Green Party’s polling has increased significantly, and Labour’s support is expected to decline. Key boroughs such as Hackney, Lambeth, and Lewisham are identified as critical areas where the Greens may perform well. However, uncertainties remain regarding voter turnout and the fragmented nature of the electoral landscape, making precise predictions challenging. Analysts generally agree that the Greens are likely to achieve a record-breaking result, with estimates suggesting a plausible range of 100 to 250 seats, though the exact outcome remains uncertain.
Sources used:
2022 results: The Green Party won 18 seats in the 2022 London borough council elections [11]. Through subsequent by-election wins and defections (notably 5 Labour councillors in Brent defecting to the Greens), they currently hold 41 borough councillors across 15 London boroughs as of April 2026 [18][19]. They are the official opposition in five boroughs: Barking & Dagenham, Islington, Lewisham, Newham, and Richmond [19].
Scale of election: All 32 London boroughs will elect councillors on 7 May 2026, with approximately 1,828 total seats being contested [1][27]. This is part of a wider set of elections across 136 English councils covering ~5,000+ seats [3].
Bombe Projection (published via The Guardian): Projects the Greens as London’s largest party with 548 seats — a gain of 530 — achieving majority control of 9 London borough councils [27][21][22]. This is the most bullish projection available and would be a truly historic result.
Green Party’s Own Target: Leader Zack Polanski has predicted winning “four to five times” the 18 seats from 2022, implying 72–90 seats in London [11][26]. Nationally, the party targets 241+ councillors [5][9]. This is notably far more conservative than the Bombe projection.
YouGov MRP (4,548 Londoners surveyed, 27 March – 21 April 2026): Projects Green vote share at 22% London-wide, up 10 percentage points from 2022. Greens projected to lead in 4 boroughs (Hackney, Lambeth, Lewisham, Waltham Forest) and be the second-largest party in 12 of 15 Labour-led boroughs [30][31]. Crucially, YouGov explicitly does NOT project seat counts, citing FPTP unpredictability with such fragmented votes [30].
Stephen Fisher (Elections Etc / Oxford): Projects ~450 net Green gains nationally across all English councils [10][23]. London’s share of this is unclear but would likely be a significant portion given London is a Green stronghold.
More in Common: Estimates national Green gains between +573 to +1,741 seats (very wide range) [22][25].
| Borough | Current Status | Green Outlook |
|---|---|---|
| Hackney | Labour-held since 1970s | Greens projected to lead; mayoral race competitive [2][13][30] |
| Lambeth | Labour-held since 2000s | Greens projected to lead [13][15][30] |
| Lewisham | Labour-held since 1970s | Greens projected to lead; mayoral race [2][13][30] |
| Waltham Forest | Labour-held | Greens projected to lead [15][30] |
| Islington | Labour-held | Labour sources “anticipate significant defeat” [5][13] |
| Brent | Labour-held (49 seats in 2022) | Super close race; 5 defections to Greens [14][15][31] |
| Haringey, Newham, Ealing, Hounslow, Enfield | Labour-held | Very/super close per YouGov [4][11][31] |
| Southwark | Labour-held | Green target [5] |
No prediction market was found specifically for London borough council seat counts for the Greens.
The critic raises several valid points that warrant careful consideration. Let me work through them systematically.
1. Lower Tail Was Too Pessimistic The critic is absolutely right that P1=38 was problematic. While the 41 current councillors includes by-election wins and defections that will need defending, having a 1% floor below their current elected membership given 22% polling (nearly double 2022) doesn’t make sense. A disappointing result would still likely exceed 45-50 seats given the polling fundamentals.
2. Party Intelligence Deserves Weight The Green Party’s 72-90 seat target does reflect ground-level canvassing data. However, I maintain parties sometimes sandbag to maximize the “exceeded expectations” narrative. The external polling data (22% vote share, leading in 4 boroughs) suggests they may be positioned for more.
3. FPTP Vote Efficiency Concern Being “second place in 12 of 15 Labour boroughs” is genuinely ambiguous news. It could mean efficient concentration just below the winning threshold, or it could mean broad but thin support that produces many close seconds.
The Central Estimate Range The “4 leading boroughs” (Hackney, Lambeth, Lewisham, Waltham Forest) contain ~234 seats. If Greens win 35-45% there (80-105 seats) plus meaningful gains in competitive boroughs (40-60 seats) plus modest pickups elsewhere (10-20 seats), we’re looking at 130-185 seats in a median scenario. My original P50 of 134 was in this range.
Upper Tail Validity The unprecedented fragmentation (Labour projected at 26%, 4-5 viable parties) means the Bombe-adjacent scenarios remain plausible. In FPTP with such fragmentation, winning thresholds could drop to 23-25% per the research, and Greens are polling at 22%.
I’ll make the following changes:
My revised estimate assumes:
The wide uncertainty reflects the inherent unpredictability of FPTP in a 4-5 party contest where vote share doesn’t translate linearly to seats.
Updated prediction:
Percentile 0.1: 35 Percentile 1: 48 Percentile 5: 62 Percentile 10: 74 Percentile 20: 90 Percentile 30: 105 Percentile 40: 118 Percentile 50: 132 Percentile 60: 152 Percentile 70: 178 Percentile 80: 210 Percentile 90: 265 Percentile 95: 330 Percentile 99: 435 Percentile 99.9: 520
The critic raises several compelling points that merit serious consideration. Let me work through them:
Over-weighting Bombe projection: The critic is absolutely right. I gave too much credence to a uniform swing model in a fragmented FPTP environment where such models are notoriously unreliable. YouGov’s explicit refusal to project seats despite having full MRP data is indeed the most telling methodological signal in the research.
Under-weighting party intelligence: The Green Party’s 72-90 seat target reflects ward-level canvassing data and local intelligence that external models can’t access. When a party riding momentum sets a target, they typically know their realistic ceiling better than pundits.
Historic precedent matters: Going from 18→138+ seats (7x+ increase) in one cycle would be extraordinary. The critic correctly notes that unprecedented outcomes require exceptional confidence, which the evidence doesn’t quite support.
Heavy upper tail unjustified: My P90-P99 values essentially required Bombe to be approximately correct despite explicitly noting its methodological weaknesses. This is inconsistent.
The political environment IS genuinely unprecedented: Labour’s projected collapse from 42%→26% in London combined with multi-party fragmentation at 15-25% each creates dynamics we haven’t seen before. This justifies some upside uncertainty.
Leading in 4 boroughs creates real floor: If Greens are genuinely leading in Hackney, Lambeth, Lewisham, and Waltham Forest (combined ~230 seats), even winning 35-45% of seats in those areas suggests 80-100 seats just from strongholds, with gains elsewhere adding 30-50 more. This supports a median above 100.
Momentum indicators are real: The Gorton by-election win, defections, and polling surge aren’t statistical noise.
The key question is: in a fragmented 4-5 way FPTP race, how efficiently does a 22% vote share concentrated in urban inner London translate to seats? Historical analogues are limited, but the Lib Dems’ experience (strong vote share, poor seat conversion when spread thin) suggests significant risk of inefficiency. However, Greens appear to have better geographic concentration than Lib Dems historically achieved.
Updated prediction:
Percentile 0.1: 32 Percentile 1: 44 Percentile 5: 56 Percentile 10: 68 Percentile 20: 85 Percentile 30: 100 Percentile 40: 114 Percentile 50: 128 Percentile 60: 148 Percentile 70: 172 Percentile 80: 202 Percentile 90: 252 Percentile 95: 315 Percentile 99: 390 Percentile 99.9: 470
Looking at the critic’s points, I need to carefully evaluate what’s valid and what might be overcorrecting.
Let me search for any recent developments that might inform my updated forecast:
The search confirms and adds to my understanding. Key updates:
Lower tail too fat: The critic is correct. With current holdings of 41 seats and 22% vote share (up from ~12%), putting significant probability below 50 seats is excessive. The Greens already have organizational infrastructure and incumbency in place.
Party target may be sandbagging: The 72-90 target is clearly conservative given they’re leading in 4 boroughs according to YouGov MRP and the Bombe projection suggests 548 seats. Politicians routinely underpromise.
Current holdings anchor: The 41 current councillors provide a meaningful floor. Losing ground from this position with nearly double the vote share would require systematic polling failure.
FPTP unpredictability: YouGov explicitly refused to project seats, stating “due to the first past the post system, a party winning the highest vote share in a borough will not necessarily win outright control of a borough, or even the most seats.” This justifies meaningful uncertainty.
Bombe projection skepticism: The 548 seat projection (30% of all seats with 22% vote share) appears to use uniform swing methodology that likely overstates gains. This seems like an upper bound, not a median.
10+ boroughs are “too close to call”: YouGov shows 10 boroughs where the gap between first and second is ≤2 points, and 16 boroughs within 5 points. This creates massive outcome volatility.
I should:
The key calculation: 22% vote share across ~1,828 seats. If Greens win proportional share, that’s ~400 seats. If efficiency is 0.3-0.5 (improved from 2022 but below proportional), we’d expect 120-200 seats as the central range.
Updated prediction:
Percentile 0.1: 30 Percentile 1: 42 Percentile 5: 56 Percentile 10: 70 Percentile 20: 88 Percentile 30: 105 Percentile 40: 122 Percentile 50: 140 Percentile 60: 162 Percentile 70: 190 Percentile 80: 225 Percentile 90: 285 Percentile 95: 370 Percentile 99: 478 Percentile 99.9: 555
Over-weighting the Bombe projection: The critic is correct that I placed too much probability mass near the 548-seat Bombe projection. Uniform swing methodology systematically overstates gains in fragmented FPTP systems. I should compress the upper tail significantly.
Under-weighting the Green Party’s own target: This is a compelling point. The Greens have access to ward-level canvass returns and ground intelligence that outsiders don’t. If they’re targeting 72-90 seats, this represents informed expectations, not sandbagging. However, I note that public party targets are often deliberately conservative to exceed expectations.
Question creator’s range signal: The “likely not higher than 200 seats” guidance deserves more weight. I’ll aim for ~65% probability below 200 rather than my previous ~60%.
Turnout differential: Young voters (who skew Green) have historically lower turnout in local elections. The 22% poll share could translate to 17-19% among actual voters.
The unprecedented political moment: A Labour collapse from 42% to 26% in London, combined with Green surge to 22%, is genuinely unprecedented. Historical precedent (18 seats in 2022) is less informative than the critic suggests.
Borough-level math supports ~150+ seats: If Greens win ~35-40% of seats in their 4 “leading” boroughs (Hackney, Lambeth, Lewisham, Waltham Forest = ~237 total seats), that alone yields 80-95 seats. Adding gains in the 8-10 competitive boroughs (where they’re second) could easily add another 50-80 seats.
The 41 current councillors as floor: These provide structural support for a minimum of ~40-50 seats even in a bad scenario.
The defection signal: 5 Labour councillors defecting to Greens in Brent is meaningful ground-level intelligence that insiders see a major shift coming.
I believe the critic’s suggested distribution is close to optimal, though I’d put slightly more probability in the 150-250 range and slightly less in the extreme tails.
Updated prediction:
Percentile 0.1: 28 Percentile 1: 44 Percentile 5: 56 Percentile 10: 70 Percentile 20: 92 Percentile 30: 112 Percentile 40: 132 Percentile 50: 152 Percentile 60: 178 Percentile 70: 212 Percentile 80: 262 Percentile 90: 335 Percentile 95: 400 Percentile 99: 485 Percentile 99.9: 545
The critic raises several compelling points that deserve careful consideration.
Green Party’s internal estimate deserves more weight: The 72-90 seat target isn’t necessarily conservative sandbagging. Parties have detailed ward-level canvassing data and internal polling. While they might aim low to declare victory, the specificity of “4-5 times 2022” suggests real modeling behind it.
“Second place everywhere” trap: This is the most important criticism. YouGov shows Greens leading in only 4 of 32 boroughs. Under FPTP, coming second in 12 boroughs means zero seats from those near-misses. I underweighted this risk.
Question creator’s 200 ceiling: An informed observer setting 200 as the likely upper bound is meaningful signal I should incorporate. Placing ~25% probability above this was too aggressive.
Wide uncertainty remains appropriate: FPTP with 4-5 competitive parties creates genuine unpredictability. A 22% vote share can translate anywhere from 50 to 300+ seats depending on geographic concentration.
The floor around current holdings: The 41 current councillors with incumbency advantages provides a reasonable floor, making results below 35-40 quite unlikely.
Momentum is real: By-election wins, defections, and sustained high polling aren’t mirages. The question is efficiency of translation, not whether support exists.
The Green Party’s 72-90 estimate may assume they win ~40% in the 4 leading boroughs plus minimal elsewhere. My original 140 assumed ~55% efficiency plus better periphery gains.
I’m lowering my median from 140 to ~128, shifting probability mass toward the 80-110 range where the Green Party’s target sits, and moderating the upper tail to respect the question creator’s assessment.
Updated prediction:
Percentile 0.1: 24 Percentile 1: 36 Percentile 5: 52 Percentile 10: 65 Percentile 20: 82 Percentile 30: 96 Percentile 40: 112 Percentile 50: 128 Percentile 60: 150 Percentile 70: 178 Percentile 80: 218 Percentile 90: 282 Percentile 95: 348 Percentile 99: 432 Percentile 99.9: 502
Agent Forecast (discrete, confidence: low)
Base Rate: In 2022, Greens won 18 London seats on 11.4% vote share (efficiency 0.088x seats/vote). A realistic efficiency in 2026 with concentrated inner-London vote is 0.3-0.5x, yielding roughly 120-200 seats at 22% vote share. Polanski’s own prediction (72-90) provides the sub-pollster floor; Bombe’s 548 represents an optimistic outlier not well-supported by PollCheck’s NOC projections.
Initial Reasoning:
Key Uncertainties:
Evidence Sources (11):
Grounding Facts (passed to reviewers):
Draft: 201+: 55.0%, 200: 2.0%, 186: 0.5%, 174: 0.5%, 194: 0.5%
Reviewer Panel:
Reviewer (reasoning)
The forecast’s weakest link is its leap from a 22% borough‐level vote share to an expectation of over 200 seats without a ward‐level model of London’s multi‐member FPTP system. It treats borough projections and high-level seat‐gain models as if they directly translate into seat wins, ignoring vote distribution, incumbency, and vote‐splitting dynamics in each ward.
Objections:
Reviewer (calibration)
The draft forecast relies heavily on a qualitative ‘non-linear surge’ and high-end poll projections to justify a massive increase in seats without articulating a robust model for converting the London-wide vote share into a specific seat count. The implied leap from a 1.9x increase in vote share to a 10x-20x increase in seats lacks clear methodological justification or historical precedent. The forecast also appears to give undue weight to an extreme outlier projection.
Objections:
Reviewer (fact-checker)
6 unsupported claim(s). Verdict: significant_errors
Unsupported claims:
Adjustment Rationale:
Adjusted downward in response to reviewer feedback. The reasoning reviewer correctly identified that I lacked a ward-level FPTP model — in 3-member wards with 5-party splits, winning thresholds typically require 25-30% rather than the 23-25% I cited, and Greens’ 22% London-wide share won’t mechanically translate to proportional seats without geographic concentration. The calibration reviewer correctly flagged that my 201+ probability (55%) gave undue weight to the Bombe 548 outlier and underweighted PollCheck’s NOC projection — which implies that even in Green-leading boroughs, they may win only 25-35 of 50-60 seats rather than sweeping. Central estimate shifted from ~230-260 to ~150-180 seats. The 201+ tail dropped from 55% to 32%, with mass redistributed across the 100-200 range. Retained meaningful floor probability in the 72-100 range matching Polanski’s own prediction. Confidence lowered to ‘low’ reflecting genuine structural uncertainty in ward-level FPTP translation that I cannot resolve without ward-level MRP data.
forecast_sdk | 577.3s | research $0.9883 + reviewers $0.0109 + reconcile $0.1895 = $1.1887
Extracted Percentile Values: {0.1: 25, 1.0: 48, 5.0: 80, 10.0: 102, 15.0: 122, 20.0: 140, 25.0: 155, 30.0: 170, 35.0: 185, 40.0: 200, 45.0: 214, 46.0: 218, 47.0: 222, 48.0: 226, 49.0: 230, 50.0: 235, 51.0: 240, 52.0: 245, 53.0: 250, 54.0: 255, 55.0: 260, 60.0: 278, 65.0: 300, 70.0: 325, 75.0: 355, 80.0: 390, 85.0: 430, 90.0: 480, 95.0: 550, 99.0: 670, 99.9: 800}
Model’s Answer: Step one is setting the base rate. The strongest source here is the fresh YouGov MRP from April 22, which puts the Greens at about 22% London-wide and leading in four boroughs in the median projection, with as many as eight in stronger simulations. I weight that more heavily than looser “sea of green” headlines and much more heavily than clearly optimistic partisan models around 548 to 575 seats. The current Green baseline is 49 seats, but this election is on all 1,817 London borough seats at once, so a citywide swing can produce a very large jump.
Step two is translating vote share into seats. The key fact is not just the overall London vote, but that Green support appears concentrated in inner-London boroughs where they are already competitive and in some places near or ahead of Labour. PollCheck’s borough-level gains in places like Lambeth, Lewisham, Greenwich, Camden, Brent, Enfield, and Southwark make a total in the low hundreds look much more plausible than a simple extrapolation from their current 49 seats would suggest. The block-vote, multi-member ward structure also means that once a party becomes locally competitive, seat conversion can jump sharply.
Step three is deciding how much to trust the seat-conversion story. I think the public discourse is probably too excited about the very top-end outcomes. If Greens lead in about four boroughs but only narrowly, and if Labour squeezes late in some wards, the total can still land more in the 180 to 300 range than in the 500-plus range. On the other hand, if the Green vote is as efficiently concentrated as some models imply, ward sweeps in several inner boroughs could push them well above 400.
The assumptions doing the most work are these. First, that the YouGov borough picture is broadly calibrated; I assign about a 35% chance that it is materially off in a way that would move the final total by 80 seats or more. Second, that Green support is concentrated enough to convert into seats rather than just respectable second places; I assign about a 30% chance this is too optimistic. Third, that there is no major late tactical squeeze back to Labour in the final two weeks; I assign about a 20% chance that this assumption fails.
A low outcome would come from Greens polling well but inefficiently, ending up with many second places and only modest ward breakthroughs; I put roughly a 9% chance on a result below 100 seats. A high outcome would come from narrow borough-wide leads turning into many full-slate ward wins across several inner-London councils; I put roughly a 12% chance on a result above 450 seats.
My median forecast is 235 seats, with a right-skewed distribution because the upside from concentrated urban support is larger than the downside from a modest polling overstatement.
Percentile .1: 25 Percentile 1: 48 Percentile 5: 80 Percentile 10: 102 Percentile 15: 122 Percentile 20: 140 Percentile 25: 155 Percentile 30: 170 Percentile 35: 185 Percentile 40: 200 Percentile 45: 214 Percentile 46: 218 Percentile 47: 222 Percentile 48: 226 Percentile 49: 230 Percentile 50: 235 Percentile 51: 240 Percentile 52: 245 Percentile 53: 250 Percentile 54: 255 Percentile 55: 260 Percentile 60: 278 Percentile 65: 300 Percentile 70: 325 Percentile 75: 355 Percentile 80: 390 Percentile 85: 430 Percentile 90: 480 Percentile 95: 550 Percentile 99: 670 Percentile 99.9: 800
Extracted Percentile Values: {0.1: 28, 1.0: 60, 5.0: 120, 10.0: 185, 15.0: 245, 20.0: 295, 25.0: 340, 30.0: 380, 35.0: 415, 40.0: 445, 45.0: 475, 46.0: 482, 47.0: 489, 48.0: 496, 49.0: 503, 50.0: 510, 51.0: 517, 52.0: 524, 53.0: 531, 54.0: 538, 55.0: 545, 60.0: 570, 65.0: 595, 70.0: 625, 75.0: 660, 80.0: 705, 85.0: 760, 90.0: 830, 95.0: 930, 99.0: 1080, 99.9: 1220}
Model’s Answer: Step one is getting the target right. Despite the background mentioning a subset of boroughs, the stronger evidence says these are the regular London borough all-out elections on 7 May 2026, covering all 32 boroughs and 1,817 seats. That matters enormously: a forecast anchored to the default 39.5 to 200.5 range would likely be forecasting the wrong object.
Step two is weighing sources by usefulness. The most important source is the fresh YouGov MRP from 22 April, because it is recent, London-specific, and from a reputable pollster. Its headline for this question is not a seat total but a very strong Green position: roughly 22% London-wide vote share, major gains from 2022, Greens leading in 4 boroughs on the median estimate and potentially more in the upside case. That is a genuinely large base signal.
Step three is translating votes into seats. The direct seat models cited in the research, such as PollCheck and Bombe-style projections, cluster around roughly 550 to 575 Green seats. I do not want to take those literally because they are less authoritative than YouGov and seat conversion in London’s ward-based system is tricky, especially if support is broad but not always first-place-efficient. Still, multiple independent-looking projections pointing to the same rough neighborhood is meaningful. With 22% vote share, full slates, incumbency from current Green councillors and defections, and strong concentration in inner-London progressive wards, a result well above 200 is more likely than not.
Step four is checking whether those headline seat projections could be too high. They could. A borough-level MRP can overstate ward-level efficiency; Labour squeeze voting could reassert itself in the final two weeks; and some “considering Green” sentiment may not convert into actual votes. Those factors are why I am placing the median below the most bullish 550 to 575 seat models, even though I still think those models capture a real wave.
The key assumptions that could most change the forecast are these. First, that the election really is the all-London 32-borough cycle rather than a smaller subset; I assign about a 5% chance this interpretation is somehow wrong, and if it were, the true answer would be dramatically lower. Second, that current polling strength translates reasonably efficiently into ward wins; I assign about a 30% chance this is materially overstated, which would pull the distribution down by roughly 150 to 250 seats. Third, that Green support does not fade meaningfully in the final two weeks; I assign about a 20% chance of a late squeeze back to Labour that would also pull the result materially lower.
A low outcome would come from Green support being real but badly distributed: lots of second places, fewer ward-level pluralities, and a late Labour squeeze in anti-Conservative or anti-Reform tactical contexts. I think a genuinely low result, say under 200, is possible but clearly a minority case. A high outcome would come from the wave case: concentrated Green surges across multiple inner-London boroughs, efficient three-seat ward sweeps, and Labour fragmentation. That is plausible enough that the right tail has to stay substantial.
My overall view is that the best center is around 500 seats, with a wide distribution because the exact ward-level conversion is uncertain and some of the most explicit seat projections may be optimistic.
Percentile .1: 28 Percentile 1: 60 Percentile 5: 120 Percentile 10: 185 Percentile 15: 245 Percentile 20: 295 Percentile 25: 340 Percentile 30: 380 Percentile 35: 415 Percentile 40: 445 Percentile 45: 475 Percentile 46: 482 Percentile 47: 489 Percentile 48: 496 Percentile 49: 503 Percentile 50: 510 Percentile 51: 517 Percentile 52: 524 Percentile 53: 531 Percentile 54: 538 Percentile 55: 545 Percentile 60: 570 Percentile 65: 595 Percentile 70: 625 Percentile 75: 660 Percentile 80: 705 Percentile 85: 760 Percentile 90: 830 Percentile 95: 930 Percentile 99: 1080 Percentile 99.9: 1220
Resolution of Disagreements:
Disagreement 1 (Lower tail / P10: 30-47 range): The research partially resolves this. The Elections Etc (Stephen Fisher) projection of +450 Green seats nationally across 5,000+ seats, combined with the More in Common projection of 1,154 Green councillors nationally, confirms the Green surge is broad-based. However, the PollCheck December 2025 article noted London’s Green growth was “lower” than other regions, which slightly supports the more cautious forecasters. The Rallings & Thrasher (LGC Plus) expert analysis specifically names Hackney, Islington, Newham, Brent, Lewisham, and Southwark as areas where “Green gains may be expected” — confirming a floor well above 2022’s 18 seats. The evidence supports a P10 in the 40-45 range rather than 30-35, as even pessimistic scenarios involve gains in multiple named boroughs.
Disagreement 2 (Central estimate / conversion efficiency: median 90-115): This was the most important disagreement. The research provides several clarifying data points:
The evidence moderately supports the higher-central-estimate forecasters (1/3) over the lower ones (4/5), but with important caveats about model overestimation. The specific naming of 4 boroughs where Greens lead, combined with expert opinion about 18+ seats in a single borough, suggests a median around 100-110 rather than 90-95.
Disagreement 3 (Upper tail / P80-P90: 150-200 vs 178-225): The Bombe model’s 548 seats and “9 council controls” is only supported by one source (The London Economic reporting Bombe/Guardian data). The Elections Etc model’s +450 nationally and More in Common’s 1,154 nationally are consistent with a large Green gain but don’t specifically project London seat counts. PollCheck’s “no Green-controlled boroughs” (from forecaster analysis) conflicts with Bombe’s “9 councils.” The weight of evidence suggests the upper tail should be meaningful but not as extreme as Bombe implies. A P90 around 200-215 seems appropriate.
Updated Analysis:
The targeted research confirms several key facts:
The central estimate should sit around 100-110 seats (median), with significant right-skew. The lower bound is well-supported at ~40 seats (even with polling miss, Greens have strong positions in multiple boroughs). The upper tail should extend to ~210-220 at P90 to account for FPTP nonlinearity if vote concentration is stronger than expected.
I’m adjusting the weighted average slightly upward in the center and narrowing the lower tail modestly, while keeping the upper tail wide.
Confidence: MEDIUM
Percentile 10: 42 Percentile 20: 60 Percentile 40: 88 Percentile 60: 115 Percentile 80: 162 Percentile 90: 215
To arrive at our forecast, we focused on the transition of the Green Party from a minor player to a significant force in London local government. We primarily weighed three factors: the party’s current “floor” of seats, recent high-resolution polling, and the mechanics of the voting system.
We began by establishing a baseline. While the Greens won only 18 seats in 2022, they entered the 2026 cycle with a much larger “pre-election” count due to defections and by-election victories. This expanded footprint suggests a stronger incumbency advantage and a higher floor than in previous years.
We then analyzed recent “MRP” polling data, which showed the Greens potentially doubling their city-wide vote share. In local elections, such increases can lead to non-linear seat gains if support is geographically concentrated. We identified several “inner London” boroughs where the party is now a primary challenger to Labour. Our thinking was heavily influenced by the possibility of “ward sweeps,” where the party could win all two or three seats in a single neighborhood.
Finally, we balanced optimistic “breakthrough” scenarios—suggested by some academic models and betting markets—against the risks of tactical voting and incomplete candidate slates, which often limit minor parties under the first-past-the-post system.