Evidence Ranking · built 2026-10-20
Famous records, on the same ruler as a wallet
Buffett, Simons, Cathie Wood, Bill Ackman, famous 13F filers, Polymarket’s French whale and Hyperliquid’s strongest wallets, each measured against its own market and then against luck. Picked blind, a record needs z ≥ 2. Picked as the best of 47,270, as famous records are, it needs 4.90.
1record clears 4.90: Renaissance Medallion, net of fees
3.15Berkshire Hathaway over 61 years against the S&P 500; 1.65 from 2001 on
87records scored; 6 famous names have no public return series to score
Length is already in z. A short record needs more return per unit of risk to reach the same z, so a 3-year vault at 3.4 and Buffett’s 61 years at 3.1 are on one scale: the vault’s returns were far steadier, so a short record could get that far. What z cannot say is whether an edge will last; a 61-year record has survived regimes a 3-year one has not met. Check the length column before comparing two rows. “Luck, picked blind” is how often a record this strong turns up with no skill behind it, for one record picked without looking at any ranking; picked as the best of 47,270, records below 4.90 are expected by chance, which is why only one row clears it.
| # | Record | Length | z after the market · bar 4.90 | Last 10 yr | Luck, picked blind |
|---|---|---|---|---|---|
| 1 | Renaissance Medallion, net of fees strongJim Simons · Yearly returns · vs S&P 500 with dividends | 31.0 yr | 4.84 | 1 in 480 billion (< 0.0001%) | |
| 2 | Fidelity Contrafund evidence if picked blindWill Danoff (since 1990) · FCNTX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | 1.15 | 1 in 17,000 (0.0059%) | |
| 3 | Vanguard Health Care evidence if picked blindEd Owens (to 2012) · VGHCX, monthly · vs S&P 500 with dividends (VFINX)Sector fund: against the S&P 500, its z includes the sector's own run. | 38.7 yr | -0.29 | 1 in 8,200 (0.012%) | |
| 4 | 0x58e1…132e evidence if picked blindHyperliquid, #7 by 30-day ROI · vs BTCno losing week | 2.8 yr | raw 4.49 | — | 1 in 4,100 (0.025%) |
| 5 | D.E. Shaw Composite evidence if picked blindD.E. Shaw · Yearly net returns as reported, 10 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 10.0 yr | — | 1 in 3,600 (0.028%) | |
| 6 | Hyperliquidity Provider (HLP) evidence if picked blindHyperliquid, #9 by all-time PnL · vs BTC | 3.5 yr | raw 3.42 | — | 1 in 2,900 (0.035%) |
| 7 | 0x0526…bb9a evidence if picked blindHyperliquid, #49 by 30-day ROI · vs BTC | 0.9 yr | raw 3.37 | — | 1 in 2,300 (0.044%) |
| 8 | Berkshire Hathaway evidence if picked blindWarren Buffett · Yearly returns · vs S&P 500 with dividends | 61.0 yr | 1.70 | 1 in 1,200 (0.082%) | |
| 9 | 0xd487…4982 evidence if picked blindHyperliquid, #32 by 30-day ROI · vs BTC | 2.3 yr | raw 3.04 | — | 1 in 930 (0.11%) |
| 10 | 0x7548…5788 evidence if picked blindHyperliquid, #10 by 30-day ROI · vs BTC | 0.4 yr | raw 2.99 | — | 1 in 920 (0.11%) |
| 11 | Millennium Management evidence if picked blindIzzy Englander · Yearly net returns as reported, 12 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 12.0 yr | — | 1 in 430 (0.23%) | |
| 12 | 0x3e7c…3909 evidence if picked blindHyperliquid, #15 by 30-day ROI · vs BTCno losing week | 0.2 yr | raw 3.06 | — | 1 in 380 (0.26%) |
| 13 | 0x67ed…623e evidence if picked blindHyperliquid, #11 by 30-day ROI · vs BTC | 0.6 yr | raw 2.72 | — | 1 in 300 (0.33%) |
| 14 | 0xff4c…463c evidence if picked blindHyperliquid, #88 by all-time PnL · vs BTC | 2.7 yr | raw 2.76 | — | 1 in 240 (0.42%) |
| 15 | Fidelity Low-Priced Stock evidence if picked blindJoel Tillinghast (from 1989) · FLPSX, monthly · vs S&P 500 with dividends (VFINX) | 36.7 yr | -0.70 | 1 in 190 (0.52%) | |
| 16 | Vanguard Wellington evidence if picked blindWellington Management · VWELX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.20 | 1 in 170 (0.59%) | |
| 17 | asparagus2012 evidence if picked blindPolymarket, #40 by all-time profit · vs price paid per share | 9 events | — | 1 in 120 (0.82%) | |
| 18 | DEGEN evidence if picked blindAI agent · Recall, Perps Trading Competition (4.0 days, 29 agents) · vs none (raw returns, days long) | — | — | 1 in 91 (1.1%) | |
| 19 | Parnassus Core Equity evidence if picked blindParnassus · PRBLX, monthly · vs S&P 500 with dividends (VFINX) | 34.0 yr | -0.37 | 1 in 80 (1.3%) | |
| 20 | Michie evidence if picked blindPolymarket, #50 by all-time profit · vs price paid per share | 8 events | — | 1 in 80 (1.3%) | |
| 21 | Jenzigo evidence if picked blindPolymarket, #32 by all-time profit · vs price paid per share | 6 events | — | 1 in 78 (1.3%) | |
| 22 | 0x5966…f804 evidence if picked blindPolymarket, #43 by all-time profit · vs price paid per share | 168 events | — | 1 in 78 (1.3%) | |
| 23 | Duquesne Family Office evidence if picked blindStanley Druckenmiller · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 11.5 yr | — | 1 in 74 (1.4%) | |
| 24 | Markel Group evidence if picked blindTom Gayner · MKL, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.44 | 1 in 62 (1.6%) | |
| 25 | RepTrump evidence if picked blindPolymarket, #12 by all-time profit · vs price paid per share | 5 events | — | 1 in 56 (1.8%) | |
| 26 | Theo4 evidence if picked blindThéo, the 2024 “French whale” (one of his four accounts) · Polymarket, #1 by all-time profit · vs price paid per share | 9 events | — | 1 in 50 (2.0%) | |
| 27 | FPA Crescent looks like luckSteven Romick (since 1993) · FPACX, monthly · vs S&P 500 with dividends (VFINX) | 33.2 yr | -0.61 | 1 in 35 (2.9%) | |
| 28 | Bridgewater Pure Alpha looks like luckRay Dalio · Yearly net returns as reported, 16 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 16.0 yr | — | 1 in 34 (2.9%) | |
| 29 | GME looks like luckAI agent · Recall, Perps Trading Competition (4.0 days, 29 agents) · vs none (raw returns, days long) | — | — | 1 in 34 (2.9%) | |
| 30 | JT looks like luckAI agent · Recall, Perps Trading Challenge (6.0 days, 11 agents) · vs none (raw returns, days long) | — | — | 1 in 33 (3.1%) | |
| 31 | Fidelity Growth Company looks like luckFidelity · FDGRX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | 1.65 | 1 in 22 (4.6%) | |
| 32 | Berkshire Hathaway, 2001 onward looks like luckWarren Buffett · Yearly returns · vs S&P 500 with dividends | 25.0 yr | — | 1 in 20 (4.9%) | |
| 33 | Vanguard PRIMECAP looks like luckPRIMECAP Management · VPMCX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | 0.82 | 1 in 20 (5.1%) | |
| 34 | Oakmark Fund looks like luckBill Nygren · OAKMX, monthly · vs S&P 500 with dividends (VFINX) | 35.0 yr | -0.49 | 1 in 18 (5.6%) | |
| 35 | H&H International looks like luckDuan Yongping (段永平) · 13F holdings, 31 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 7.5 yr | — | 1 in 18 (5.6%) | |
| 36 | Fairfax Financial looks like luckPrem Watsa · FFH.TO, monthly · vs S&P 500 with dividends (VFINX) | 22.9 yr | 0.59 | 1 in 14 (7.4%) | |
| 37 | Daily Journal Corp. looks like luckCharlie Munger (chairman to 2022) · DJCO, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | 0.23 | 1 in 13 (7.6%) | |
| 38 | Sequoia Fund looks like luckRuane, Cunniff · SEQUX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.12 | 1 in 13 (7.9%) | |
| 39 | AMG Yacktman looks like luckYacktman Asset Management · YACKX, monthly · vs S&P 500 with dividends (VFINX) | 34.1 yr | 0.37 | 1 in 13 (7.9%) | |
| 40 | Scion Asset Management looks like luckMichael Burry · 13F holdings, 32 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 4.5 yr | — | 1 in 10 (9.9%) | |
| 41 | Citadel Wellington looks like luckKen Griffin · Yearly net returns as reported, 18 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 18.0 yr | — | 1 in 9 (11%) | |
| 42 | Leader of TradeRank season 10 (openai) looks like luckAI model · TradeRank season 10, day 9 of 28 (22 models, simulated) · vs none (raw returns, days long) | — | — | 1 in 8 (12%) | |
| 43 | Mairs & Power Growth looks like luckMairs & Power · MPGFX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.45 | 1 in 8 (13%) | |
| 44 | Baron Partners looks like luckRon Baron · BPTRX, monthly · vs S&P 500 with dividends (VFINX) | 34.6 yr | 0.66 | 1 in 7 (15%) | |
| 45 | Fairholme Fund looks like luckBruce Berkowitz · FAIRX, monthly · vs S&P 500 with dividends (VFINX) | 26.7 yr | -0.26 | 1 in 6 (15%) | |
| 46 | American Funds Washington Mutual looks like luckCapital Group · AWSHX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | 0.14 | 1 in 6 (16%) | |
| 47 | Fidelity Select Technology looks like luckFidelity · FSPTX, monthly · vs S&P 500 with dividends (VFINX)Sector fund: against the S&P 500, its z includes the sector's own run. | 38.7 yr | 1.68 | 1 in 6 (18%) | |
| 48 | Dalal Street looks like luckMohnish Pabrai · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 5.0 yr | — | 1 in 5 (20%) | |
| 49 | Dodge & Cox Stock looks like luckDodge & Cox · DODGX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.44 | 1 in 5 (21%) | |
| 50 | Columbia Seligman Technology looks like luckColumbia Threadneedle · SLMCX, monthly · vs S&P 500 with dividends (VFINX)Sector fund: against the S&P 500, its z includes the sector's own run. | 38.7 yr | 1.77 | 1 in 4 (26%) | |
| 51 | Vanguard Windsor II looks like luckVanguard · VWNFX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.74 | 1 in 4 (27%) | |
| 52 | Best model in Alpha Arena season 1 looks like luckAI model (6 competed) · nof1 Alpha Arena, season 1 · vs none (raw returns) | — | — | 1 in 3 (29%) | |
| 53 | T. Rowe Price Blue Chip Growth looks like luckT. Rowe Price · TRBCX, monthly · vs S&P 500 with dividends (VFINX) | 33.2 yr | -0.08 | 1 in 3 (30%) | |
| 54 | Pershing Square Holdings looks like luckBill Ackman · PSH.L, monthly · vs S&P 500 with dividends (VFINX) | 9.4 yr | — | 1 in 3 (33%) | |
| 55 | Gabelli Asset Fund looks like luckMario Gabelli (since 1986) · GABAX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.71 | 1 in 3 (33%) | |
| 56 | Scottish Mortgage looks like luckBaillie Gifford · SMT.L, monthly · vs S&P 500 with dividends (VFINX) | 22.7 yr | 0.24 | 1 in 3 (35%) | |
| 57 | James Wynn looks like luckJames Wynn · Hyperliquid · vs BTC | 1.0 yr | raw 0.32 | — | 1 in 3 (36%) |
| 58 | Appaloosa looks like luckDavid Tepper · 13F holdings, 42 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 8.5 yr | — | 1 in 3 (37%) | |
| 59 | Royce Pennsylvania Mutual looks like luckChuck Royce · PENNX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.24 | 1 in 3 (38%) | |
| 60 | Greenlight Capital looks like luckDavid Einhorn · Yearly net returns as reported, 12 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 12.0 yr | — | 1 in 3 (39%) | |
| 61 | American Funds Growth Fund of America looks like luckCapital Group · AGTHX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.21 | 1 in 2 (40%) | |
| 62 | American Funds Investment Company of America looks like luckCapital Group · AIVSX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.09 | 1 in 2 (41%) | |
| 63 | Janus Henderson Enterprise looks like luckJanus Henderson · JAENX, monthly · vs S&P 500 with dividends (VFINX) | 34.0 yr | -1.13 | 1 in 2 (42%) | |
| 64 | Gates Foundation Trust looks like luckBill Gates · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 11.8 yr | — | 1 in 2 (42%) | |
| 65 | Jensen Quality Growth looks like luckJensen Investment Management · JENSX, monthly · vs S&P 500 with dividends (VFINX) | 34.1 yr | -1.19 | 1 in 2 (47%) | |
| 66 | NANC ETF (Democrats’ disclosed trades) looks like luckNancy Pelosi and colleagues, as disclosed · NANC, monthly · vs S&P 500 with dividends (VFINX) | 3.5 yr | — | 1 in 2 (49%) | |
| 67 | Baron Growth looks like luckRon Baron (since 1994) · BGRFX, monthly · vs S&P 500 with dividends (VFINX) | 31.0 yr | -2.08 | more likely than not (50%) | |
| 68 | Weitz Value looks like luckWally Weitz · WVALX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -2.27 | more likely than not (52%) | |
| 69 | Baupost Group looks like luckSeth Klarman · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 2.5 yr | — | more likely than not (57%) | |
| 70 | Hillhouse (HHLR Advisors) looks like luckZhang Lei (张磊) · 13F holdings, 31 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 7.0 yr | — | more likely than not (58%) | |
| 71 | Jev (TypeSafe) looks like luckAI model · TradeRank season 10, day 9 of 28 (22 models, simulated) · vs none (raw returns, days long) | — | — | more likely than not (58%) | |
| 72 | Tiger Global looks like luckChase Coleman · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 9.0 yr | — | more likely than not (59%) | |
| 73 | Akre Capital looks like luckChuck Akre · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 11.2 yr | — | more likely than not (65%) | |
| 74 | ARK Innovation ETF looks like luckCathie Wood · ARKK, monthly · vs S&P 500 with dividends (VFINX) | 11.8 yr | -0.40 | more likely than not (69%) | |
| 75 | Third Point Offshore looks like luckDan Loeb · Yearly net returns as reported, 17 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 17.0 yr | — | more likely than not (72%) | |
| 76 | Tiger Global (flagship) looks like luckChase Coleman · Yearly net returns as reported, 8 consecutive years · vs S&P 500 with dividendsyearly returns as reported by the press or the firm, one cited source per year; not audited | 8.0 yr | — | more likely than not (73%) | |
| 77 | Davis New York Venture looks like luckDavis Advisors · NYVTX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.68 | more likely than not (74%) | |
| 78 | Greenlight Capital looks like luckDavid Einhorn · 13F holdings, 37 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 3.8 yr | — | more likely than not (77%) | |
| 79 | Berkshire Hathaway, 13F stock portfolio looks like luckWarren Buffett · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 11.8 yr | — | more likely than not (79%) | |
| 80 | Himalaya Capital looks like luckLi Lu (李录) · 13F holdings, 35 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 7.2 yr | — | more likely than not (79%) | |
| 81 | Third Point looks like luckDan Loeb · 13F holdings, 47 quarters · vs S&P 500 with dividends (VFINX)estimate from 13F holdings: long US stocks only, held from one quarter end to the next | 6.2 yr | — | more likely than not (79%) | |
| 82 | Longleaf Partners looks like luckMason Hawkins (since 1987) · LLPFX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -2.48 | more likely than not (85%) | |
| 83 | Fidelity Magellan looks like luckFidelity (Peter Lynch to 1990) · FMAGX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.77 | more likely than not (86%) | |
| 84 | T. Rowe Price Growth Stock looks like luckT. Rowe Price · PRGFX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -0.59 | more likely than not (87%) | |
| 85 | Polen Growth looks like luckPolen Capital · POLRX, monthly · vs S&P 500 with dividends (VFINX) | 15.7 yr | -1.53 | more likely than not (88%) | |
| 86 | Machi Big Brother looks like luckJeffrey Huang · Hyperliquid · vs BTC | 1.4 yr | raw -1.23 | — | more likely than not (88%) |
| 87 | Legg Mason Value Trust (now ClearBridge) looks like luckBill Miller (founding manager) · LMVTX, monthly · vs S&P 500 with dividends (VFINX) | 38.7 yr | -1.29 | more likely than not (95%) |
z is the evidence after taking out the market: stocks and funds against the S&P 500 with dividends, wallets against BTC, Polymarket against the price paid for each share. Wallet rows also show the raw t that the wallet verdict uses today. “Luck alone” is the chance that, with no skill at all, a record gets this far: picked blind, and as the best of 47,270. Data: latest.json.
Every board we have rescored
Ranked by profit. On every board we have rescored, no one has strong evidence yet.
Each dot is one record’s t on a public leaderboard; the red line is the bar for being picked from that board. Dots past t = 2 are promising; none is past the red line. Why that is not the method saying no to everything.
-4-20246
Hyperliquidall-time PnL top 100, from 47,6070 / 99strong (≥ 4.90)9 promising (t ≥ 2)Hyperliquid30-day ROI top 1000 / 81strong (≥ 4.90)14 promising (t ≥ 2)Binancecopy trading, 90-day PnL top 1000 / 83strong (≥ 4.57)9 promising (t ≥ 2)OKXcopy trading, PnL top 1000 / 97strong (≥ 3.73)10 promising (t ≥ 2)Polymarketall-time profit top 50, complete records0 / 38strong (≥ 5.33)8 promising (t ≥ 2)RecallAI agent perps competitions, 10 of them0 / 149strong (≥ 2.84–3.16)1 promising (t ≥ 2)Alpha Arenaseason 1, 6 AI models0 / 6strong (≥ 2.67)0 promising (t ≥ 2)TradeRankAI models, season 10 at day 9, Jev included0 / 21strong (≥ 3.08)0 promising (t ≥ 2)Hyperliquid rows are rebuilt daily; the others are as of their audit dates. Values beyond the axis are drawn at its edge. Papertrade’s board (14,381 accounts) has no return series yet; three groups were frozen 2026-10-16 for readings on 10-23 and 10-30.
Famous, but not on the list
- Peter Lynch, Fidelity Magellan, 1977–1990: the price feed misses distributions before 1987; needs Fidelity’s yearly returns.
- John Neff, Vanguard Windsor, 1964–1995: same gap in the price feed; needs Vanguard’s yearly returns.
- George Soros, Quantum Fund: no complete yearly series published; only scattered years.
- Peter Thiel, Clarium Capital, Founders Fund: no public return series.
- Leopold Aschenbrenner, Situational Awareness LP: most of the 13F value is options, which a holdings estimate cannot price; returns reported by the press, not published.
- Ray Dalio, Bridgewater Pure Alpha: private; the public All Weather ETF is too new.
13F estimates. For investors who publish only quarterly holdings, the ranking rebuilds the long US stock book from each 13F filing and holds it to the next quarter end. It is labelled as an estimate on every row: a filing omits shorts, leverage, cash, non-US positions and trades within the quarter, and a book that is mostly options is left out.
How it is computed
- Against the market. Each record’s returns are regressed on its benchmark’s; z is the t of the alpha on the normal scale (Student’s t with n − 2 degrees of freedom), the same scale as a wallet’s t. Tested against zero instead, the S&P 500 itself reaches 4.94 over 1965–2025.
- Sources. Berkshire: the yearly table in the 2025 shareholder letter. Medallion: Cornell (2019), Table 1, from Zuckerman (2019); estimates, not audited. Other stocks and funds: Yahoo Finance adjusted closes, monthly, converted to US dollars, from 1988 on. Wallets: Hyperliquid’s public API. Polymarket: audit 008.
- Checks. Berkshire’s price feed matches the letter’s yearly figures to within 0.05 points over 45 years; the S&P 500 index fund matches the letter’s S&P column to within 0.98 points a year from 1988. Before 1987 the feed misses fund distributions, so no series starts earlier.
- Survivors. A price feed only holds funds that still exist, so the stocks-and-funds rows lean toward survivors.
- Who is named. A row is named for the person the public associates with the record; where the manager joined after launch or left before the end, the row says so.
Script: tools/build_ranking.py. Not investment advice.