Track record · we keep score in public
Every thesis, on the clock.
Anyone can sound confident. We publish the exact condition that would prove each thesis wrong, then resolve it in public — played out, or invalidated. Strictly non-monetary; just whether the falsifiable claim was falsified. Below: every live commitment and the clock on it.
Calibration · 250 scored
meaningful sampleBrier score
0.26
0 perfect · 0.25 coin-flip
Played-out rate
44%
111 of 250
Calibration error
0.014
lower = conviction tracks reality
Skill vs base rate
−0.04
Brier skill score
Read these as directional, not a verdict: with 250 resolved, a single-tier row (n=1) can show 100% off one outcome and the skill score sits near zero on a young record. The numbers are published as they stand and sharpen as more theses resolve.
What a Brier score of 0.26 means
A Brier score is the mean squared error between the probability a forecast stakes and what actually happened. Each conviction tier carries an implied probability of playing out — SUPREME around 90%, LOW around a coin-flip 50% — and every resolved thesis is graded against the odds its tier published in advance. A score of 0 would be perfect foresight; 0.25 is what you'd get by forecasting 50% on everything.
That gives the number a reference band. Expert human forecasters in Tetlock's Good Judgment work are commonly cited around 0.08, and the strongest LLM forecasters in 2024–25 evaluations around 0.10; an uninformed coin-flip sits at 0.25. Those benchmarks score different questions than equity theses, so read them as scale markers rather than a like-for-like comparison. orbyd's 0.26 is read against those marks, on a record that is still young — 250 resolved theses is enough to be directional, not enough to be a verdict. The odds each tier stakes are fixed in advance and published below, and every resolved call is listed with the tier it staked, so the score is recomputable from the public record rather than asserted.
It is not a sealed number, though, and we'd rather say that than imply otherwise. A resolved thesis is scored on the conviction tier its dossier carries, and the research pipeline can restate a tier after the call has already resolved — which moves the published score with no new outcome behind it. What we won't do is restate a call to improve the figure: once a resolution is published it stays as published, drift and all. Read this as a live measurement of a record still being built.
What a near-zero skill score tells you
The Brier skill score (−0.04) asks a sharper question than raw accuracy: does staking conviction beat simply forecasting the base rate? A positive value means the conviction tiers add information; zero means they don't yet, which is the honest read on a record this size. Publishing that plainly — rather than a self-reported headline accuracy — is the whole point.
It also draws the line between a real scoreboard and a disappearing one. Research that quietly keeps its winners and drops its losers can claim any accuracy it likes, precisely because nothing was falsifiable to begin with. A score that can sit near zero in public, attached to every resolved call, is one you can actually audit — which is what AI equity research needs to be trustworthy, and what an invalidation trigger makes possible. Whether that score reflects skill or luck is a question a sample this size can't yet settle.
The Brier score, decomposed
0.014
Reliability: the gap between the play-out rate each tier claimed and the rate it actually hit. Lower is better.
0.003
Discrimination: how far the conviction tiers pull outcomes apart from the base rate. Higher is better; near zero means the tiers don't yet separate.
0.247
The irreducible difficulty of the question itself — the variance you'd face forecasting the base rate alone, before any conviction is staked.
These three sum to the headline figure: Brier = calibration − resolution + uncertainty. The split is the honest part. A low calibration term paired with a near-zero resolution term is exactly what a skill-≈0 record looks like — confidence that tracks reality reasonably well, but conviction tiers that don't yet sort the wins from the losses. The refinement term (0.244, uncertainty minus resolution) measures the same thing from the other side: how much sharper than the base rate the forecasts have managed to be so far.
Dashed = perfect calibration. Each dot is a conviction tier; the line shows the gap between what it claimed and what happened.
Does conviction predict the outcome?
| Tier | Claimed | Actual | n |
|---|---|---|---|
| HIGH | 75% | 17% | 6 |
| MEDIUM | 60% | 50% | 60 |
| LOW | 50% | 43% | 184 |
Play-out rate by archetype
- Compounder 36% · 25
- Cyclical recovery 31% · 48
- Theme leader 53% · 19
- Special situation 42% · 69
- Earnings inflection 57% · 49
- Retail squeeze 50% · 22
- Defensive 50% · 18
Play-out rate by regime at resolution
- Risk-on 45% · 228
- Neutral 41% · 22
Resolved · 251 · 250 scored
Two ways a thesis resolves — played out or invalidated, both kept on the record:
- OKLO LOW Played out
- CCC MEDIUM Played out
- FRD LOW Played out
- SENS LOW Played out
- TNDM LOW Played out
- WEAV LOW Invalidated trigger fired
- WDC LOW Invalidated trigger fired
- ELF MEDIUM Played out
- MIRM MEDIUM Invalidated trigger fired
- RAPP MEDIUM Played out
- AKBA LOW Invalidated trigger fired
- BROS LOW Invalidated trigger fired
- CHYM LOW Played out
- FLY LOW Played out
- GGAL LOW Invalidated trigger fired
- GOLF LOW Invalidated trigger fired
- HELP LOW Played out
- LZ LOW Invalidated trigger fired
- NWPX LOW Invalidated trigger fired
- QURE LOW Played out
- ZVRA LOW Played out
- BBAR MEDIUM Invalidated trigger fired
- GHRS MEDIUM Played out
- COMP LOW Played out
- TBI LOW Played out
- UI LOW Played out
- HRI MEDIUM Played out
- NKTR MEDIUM Played out
- AEIS LOW Played out
- AOSL LOW Played out
- BRKR LOW Invalidated trigger fired
- DDOG LOW Played out
- ORIC LOW Played out
- VECO LOW Played out
- WGS LOW Played out
- AMBA MEDIUM Played out
- PGEN MEDIUM Played out
- RRGB MEDIUM Played out
- AAPL LOW Invalidated trigger fired
- AGX LOW Played out
- ATEN LOW Invalidated trigger fired
- BTE LOW Played out
- FIG LOW Played out
- KURA LOW Invalidated trigger fired
- MDB LOW Played out
- MSFT LOW Played out
- ODFL LOW Invalidated trigger fired
- RDW LOW Played out
- SABR LOW Played out
- SVV LOW Played out
- XRX LOW Played out
- VCYT MEDIUM Invalidated trigger fired
- NSIT LOW Played out
- RBLX LOW Invalidated trigger fired
- MRX HIGH Played out
- PRCH MEDIUM Played out
- KSS LOW Played out
- SNEX LOW Played out
- UNIT LOW Invalidated trigger fired
- IREN LOW Invalidated trigger fired
- NNE LOW Invalidated trigger fired
- CRDO MEDIUM Invalidated trigger fired
- CTOS MEDIUM Invalidated trigger fired
- MEI MEDIUM Invalidated trigger fired
- RMAX MEDIUM Invalidated trigger fired
- CAKE LOW Played out
- DDD LOW Invalidated trigger fired
- ENOV LOW Played out
- FOUR LOW Played out
- GTX LOW Invalidated trigger fired
- IART LOW Invalidated trigger fired
- JACK LOW Played out
- LUNR LOW Invalidated trigger fired
- ON LOW Invalidated trigger fired
- OUST LOW Invalidated trigger fired
- PAY LOW Played out
- SITM LOW Invalidated trigger fired
- RXO MEDIUM Invalidated trigger fired
- STX LOW Invalidated trigger fired
- UMC LOW Invalidated trigger fired
- MTUM LOW Invalidated trigger fired
- ACIW MEDIUM Played out
- SNDX MEDIUM Invalidated trigger fired
- ALHC LOW Invalidated trigger fired
- CDW LOW Played out
- FN LOW Invalidated trigger fired
- LILAK LOW Played out
- MTRN LOW Invalidated trigger fired
- TH LOW Invalidated trigger fired
- UCTT LOW Invalidated trigger fired
- VERU LOW Invalidated trigger fired
- EVH LOW Invalidated trigger fired
- LGIH LOW Played out
- M LOW Played out
- MYE LOW Played out
- MYGN LOW Invalidated trigger fired
- NEO LOW Invalidated trigger fired
- NOK LOW Invalidated trigger fired
- NUTX LOW Invalidated trigger fired
- REAX LOW Invalidated trigger fired
- SMCI LOW Played out
- VICR LOW Invalidated trigger fired
- AFRM MEDIUM Invalidated trigger fired
- MRCY MEDIUM Played out
- PANW HIGH Invalidated trigger fired
- BAND LOW Invalidated trigger fired
- GCMG LOW Invalidated trigger fired
- MMYT LOW Invalidated trigger fired
- XXI LOW Invalidated trigger fired
- ALGT HIGH Invalidated trigger fired
- FROG MEDIUM Invalidated trigger fired
- YETI LOW Played out
- AGYS LOW Invalidated trigger fired
- RARE LOW Invalidated trigger fired
- BULL MEDIUM Played out
- TWLO MEDIUM Invalidated trigger fired
- ZBIO LOW Played out
- BTGO LOW Played out
- CORT LOW Played out
- CPIX LOW Played out
- PLUG LOW Played out
- W LOW Invalidated trigger fired
- AMC LOW Played out
- AMKR LOW Invalidated trigger fired
- BB MEDIUM Invalidated trigger fired
- CLS MEDIUM Invalidated trigger fired
- ENPH LOW Invalidated trigger fired
- SANA LOW Invalidated trigger fired
- SPRY LOW Invalidated trigger fired
- FWRD LOW Played out
- EXTR LOW Invalidated trigger fired
- MARA MEDIUM Invalidated trigger fired
- PDFS LOW Invalidated trigger fired
- STM LOW Invalidated trigger fired
- TMC LOW Invalidated trigger fired
- KMX LOW Played out
- LEGN LOW Invalidated trigger fired
- BBY LOW Played out
- CLPT LOW Invalidated trigger fired
- NXT LOW Invalidated trigger fired
- PD LOW Played out
- JOBY LOW Invalidated trigger fired
- STUB LOW Invalidated trigger fired
- WING LOW Invalidated trigger fired
- AMD LOW Played out
- OXM LOW Played out
- PNRG LOW Played out
- NEXA LOW Played out
- PENG LOW Played out
- GCO LOW Invalidated trigger fired
- KALU LOW Invalidated trigger fired
- MTH LOW Invalidated trigger fired
- SHOO LOW Invalidated trigger fired
- FORM MEDIUM Invalidated trigger fired
- HIVE LOW Invalidated trigger fired
- KFRC LOW Played out
- KOS LOW Played out
- LGN LOW Invalidated trigger fired
- RMBS LOW Invalidated trigger fired
- VSH LOW Invalidated trigger fired
- MRVL MEDIUM Invalidated trigger fired
- IMOS MEDIUM Played out
- LPG LOW Played out
- BTDR LOW Invalidated trigger fired
- CIEN LOW Invalidated trigger fired
- FLEX LOW Invalidated trigger fired
- FPS LOW Invalidated trigger fired
- KEEL MEDIUM Invalidated trigger fired
- MTSI MEDIUM Invalidated trigger fired
- PI LOW Played out
- SAIL LOW Played out
- VIAV LOW Invalidated trigger fired
- ATI MEDIUM Invalidated trigger fired
- KYMR NONE Played out
- LRCX MEDIUM Invalidated trigger fired
- ONTO HIGH Invalidated trigger fired
- TNGX MEDIUM Invalidated trigger fired
- TSM HIGH Invalidated trigger fired
- MYRG HIGH Invalidated trigger fired
- AESI LOW Invalidated trigger fired
- AMN LOW Played out
- CCJ LOW Invalidated trigger fired
- HUT LOW Invalidated trigger fired
- ALTO LOW Played out
- SXC LOW Invalidated trigger fired
- SYRE LOW Invalidated trigger fired
- FAC LOW Played out
- AMSC LOW Invalidated trigger fired
- CLSK MEDIUM Invalidated trigger fired
- BTBT LOW Invalidated trigger fired
- SHLS LOW Played out
- PSNL MEDIUM Played out
- CEVA MEDIUM Invalidated trigger fired
- TLRY LOW Invalidated trigger fired
- RGTI LOW Invalidated trigger fired
- ADPT MEDIUM Played out
- GENI LOW Invalidated trigger fired
- NBR LOW Invalidated trigger fired
- QUIK LOW Invalidated trigger fired
- TSHA MEDIUM Played out
- NBIS LOW Invalidated trigger fired
- VRT MEDIUM Invalidated trigger fired
- OKTA MEDIUM Invalidated trigger fired
- SNOW LOW Invalidated trigger fired
- BFLY MEDIUM Played out
- DXCM MEDIUM Invalidated trigger fired
- MGTX MEDIUM Played out
- ZETA LOW Invalidated trigger fired
- CDNL LOW Played out
- BLZE MEDIUM Played out
- IPX LOW Invalidated trigger fired
- OPEN LOW Played out
- PTEN LOW Invalidated trigger fired
- HNGE LOW Played out
- RVMD MEDIUM Played out
- PRM MEDIUM Played out
- TBLA LOW Played out
- ALGM LOW Played out
- AUGO LOW Played out
- CECO MEDIUM Played out
- DIOD LOW Played out
- FIVN LOW Invalidated trigger fired
- JBLU MEDIUM Played out
- MCHP LOW Played out
- MRAM LOW Played out
- PCT LOW Invalidated trigger fired
- POWI MEDIUM Played out
- STRO LOW Played out
- TER MEDIUM Played out
- TRVI LOW Played out
- DRTS MEDIUM Played out
- TWST LOW Played out
- AZZ LOW Played out
- INDI LOW Invalidated trigger fired
- RIOT LOW Invalidated trigger fired
- CLFD LOW Invalidated trigger fired
- INOD LOW Invalidated trigger fired
- QCOM MEDIUM Invalidated trigger fired
- WEST LOW Invalidated trigger fired
- PURR LOW Invalidated trigger fired
- WTI LOW Invalidated trigger fired
- WULF MEDIUM Played out
- GLW LOW Invalidated trigger fired
- NVDA MEDIUM Played out
- RKLB MEDIUM Invalidated trigger fired
- MP MEDIUM Played out
- COHR MEDIUM Invalidated trigger fired
- GLXY LOW Invalidated trigger fired
- IONQ LOW Played out
- SGMT LOW Invalidated trigger fired
- APLD LOW Invalidated trigger fired
Open commitments · 44 · soonest to resolve first
Every thesis on the active book — held, watchlist or rotation — that ships a published, falsifiable invalidation trigger. This is not the full scored set: a name we've stopped actively tracking keeps its published trigger on its own dossier and still resolves into the record above if that trigger fires, it just doesn't sit on this board. Distinct from the names held on the live book and the forward theme-bets scored on their own board.
-
Invalidates if A daily close below $7.59 gives back the entire 2026-07-29 to 2026-08-04 advance and re-opens the $7.21 52-week low into fresh discovery; secondary: the 2026-08-05 call and the ~2026-08-14 10-Q passing with no dated, binding TVA power purchase agreement milestone.
-
Invalidates if A daily close below $239 gives back the entire 2026-08-04 advance and loses the $239.06-$241.54 post-print base. Secondary: Nvidia's 2026-08-26 print carrying hyperscaler capex-digestion language, or Q2 FY27 (~2026-10-28) royalty growth landing at or under the ~13% guide.
-
Invalidates if A weekly close below $251 forfeits the $251.33–$255.72 July base and re-opens the 2026-07-01 low of $228.63. Secondary: the 2026-08-06 call passing with no new large-load counterparty named, or an FY-2026 adjusted-EPS band moved beneath $11.00.
-
Invalidates if A weekly close below $19.47 — the 7/29 low close of the unwind, with the 7/28 $19.16 intraday just beneath — voids the eight-session base; secondarily, the 8/6 SpaceX unlock passing with SPCX taking out its $107.01 low close while PL fills the $21.62–$22.15 gap.
-
Invalidates if A daily close below $17.84 breaks the higher-low sequence built off the 2026-07-29 low at $16.18 and returns price to the pre-AT&T $16.21 shelf. Secondary: the 2026-08-06 report passing with no dollar value or RPO attribution for the AT&T and Nasdaq Verafin agreements.
-
Invalidates if A weekly close below $139 — losing the 2026-07-31 $139.11 52-week low that the $141.03 low of 2026-08-03 and the $145.60 low of 2026-08-04 held — marks the four-session base as failed; secondary, the Senate leaving 2026-08-10 with no cloture petition filed on H.R. 3633.
-
Invalidates if A weekly close below $85 forfeits the shelf defended at $86.21 on 2026-07-30 and $85.30 on 2026-08-03. Secondary breaks: the Senate reaching recess around 8/7 with no CLARITY floor vote and no closing reclaim of $97.28, or the 8/7 SDNY hearing leaving New York's Kalshi suit unrestrained.
-
Invalidates if A weekly close below $132.66 takes out the May low and the 52-week floor, leaving no defended reference level beneath the tape; secondarily, an 2026-08-07 print that only reaffirms the $6.8–7.6B FY26 EBITDA range leaves nothing company-dated before the PJM backstop window opens 2026-09-30.
-
Invalidates if A weekly close below $737 takes out the 2026-07-29 capitulation low ($739.00 close / $737.88 intraday) from which the four-session rebound to $892.67 was built, leaving no identified cycle low. Secondary: Commerce clearing CXMT DRAM into US-market products, ending the three-supplier structure the ~86% margin guide rests on.
-
Invalidates if A weekly close below $43.50, beneath the ~$43.52 undisturbed pre-deal reference implied by the disclosed 24.08% premium to $54.00, prices the merger below standalone Iridium; secondary condition is no Rocket Lab Form S-4 on the SEC file by the Q3 2026 reporting cycle.
-
Invalidates if A weekly close below $1,582 surrenders the 2026-08-03 higher low and puts the 2026-07-29 reaction low of $1,530.64 back in play; secondary confirmation would be China's revenue share printing at or below its 16% H1 level at the 2026-10-14 Q3 report, against the CFO's ~20% full-year 2026 guide.
-
Invalidates if A daily close below $110 fills the 2026-08-04 policy gap — prior close $110.21, open $128.51 — and unwinds the re-rating; secondary conditions are Q3 revenue landing below the $255M guide floor, or 2026 ending with no Federal Register entry and no effective date for the FCC transceiver import measure.
-
Invalidates if A weekly close below $62 erases the 2026-08-03/04 advance and the 20-day reclaim, leaving the 2026-08-03 reversal low of $55.74 as the next reference; secondary: the $150.0M-$200.0M FY26 range cut or withdrawn on the 2026-08-10 call.
-
Invalidates if A weekly close below $88 breaks the July shelf (lowest weekly close $91.67 on 2026-07-24) and opens the $81.81 52-week low; secondary confirmation is the ~2026-09-01 Stretch rate printing above 12.00%, or STRC closing back under the $86.52 average of the issuer's first buyback tranche.
-
Invalidates if A daily close below $8.22 forfeits the eleven-session band top reclaimed on 2026-08-03 and returns the stock to the $7.30–$8.22 range; secondarily, a 2026-08-10 Q2 print with gross profit still negative on a revenue base above Q1's $3.7M, or no order book disclosed that bridges toward the $41.24M FY2026 consensus.
-
Invalidates if A daily close below $50.24 fills the 2026-08-04 gap and negates the break-and-extend leg ($48.19 returns price inside the July range). Secondary: the ~2026-09-02 Q3 FY26 print landing revenue beneath the guided $11.5–12.1B, or 2026-08-11 Supermicro and 2026-08-26 NVIDIA passing without HPE holding the gap floor.
-
Invalidates if A daily close below $95.65 completes the 2026-08-04 gap fill that the 2026-08-05 low of $95.74 stopped nine cents short of, returning price into the late-July range whose closes ran $88.20–$93.76; secondary, the 2026-08-11 print passing with no June-quarter spectrum sale agreement disclosed and no use of the $226.7M repurchase authorisation.
-
Invalidates if A weekly close below $293 removes the June breakout zone (the $293.90 pre-breakout high, the 06-18 close of $314.44) and leaves the 200-day at $248.39 as the next reference; secondarily, the ~2026-08-26 short-interest report showing the 2,349,330-share base cut without price reclaiming the 50-day at $353.53.
-
Invalidates if A weekly close below $405 surrenders the 2026-07-31 close of $405.37 and drops price under the 50-day near $407, forfeiting the entire four-session advance off the 2026-07-29 close of $369.64. Secondary: ISG operating margin printing below Q1 FY27's 10.5% at the 2026-09-03 call.
-
Invalidates if A weekly close below $266 dates the 2026-08-04 reclaim of the June floor as a failed retest and puts the entire post-2026-07-15 advance behind the tape; secondary conditions are the 2026-08-11 Deutsche Bank appearance passing without an improved intra-quarter pricing read, or DAT dry van spot linehaul printing under the $2.25 contract rate.
-
Invalidates if A weekly close below $167 takes out the 2026-07-30 session low of $167.58 and the rising 200-day at $166.56 together, leaving no defended shelf above the $119.32 52-week low; secondary confirmation would be a third straight weekly decline in DAT dry van and flatbed from $2.32 and $2.83/mi in the ~2026-08-11 reports.
-
Invalidates if A weekly close below $45.92 fills the 2026-08-07 gap and returns price under the floor of the 2026-07-21→08-03 shelf; secondary: the ~40-unit H2 2026 Atera shipment plan cut or re-dated into 2027, or FY2026 guidance falling back below the $610M floor.
-
Invalidates if A weekly close below $68 gives back the entire 2026-08-05 to 2026-08-07 expansion leg, losing the 2026-08-03 close of $68.67 and the rising 50-day moving average near $72.41. Secondary: Lumentum (2026-08-11) or Coherent (2026-08-12) guiding datacom below consensus, or the initial $43.5M Lumentum deposit not landing by ~2026-09-04.
-
Invalidates if A weekly close below $228 returns the post-print reclaim to the broken column — daily closes on 2026-08-05 ($211.14) and 2026-08-06 ($224.70) already sat beneath it. Secondary: Lumentum (2026-08-11) or Coherent (~2026-08-12) guiding optical demand lower, or any of the seven covering firms cutting its target after 2026-08-07.
-
Invalidates if A daily close below $85.19 hands back the 2026-07-22/23 shelf that the 2026-08-03 reversal bar reclaimed; secondarily, the 2026-08-11 Q2 landing at or under the $2.525B guide midpoint with adjusted operating margin at or below Q1's 1% and the FY26 $12–13B range not reaffirmed.
-
Invalidates if A weekly close below $779 gives back the entire 2026-08-04 policy repricing and returns price beneath the 50-day reclaimed that session; secondary is the 2026-08-11 FQ4 print landing inside the $960M–$1.01B guide with no FQ1 FY27 outlook above the Street's ~$1.01B / $3.02 bar.
-
Invalidates if A weekly close below $63 undoes the 50-day zone reclaimed 2026-08-03 and ends the post-print higher-low sequence; a weekly close below $56 also loses the July shelf. Secondarily, 2026-08-12 passing with the 2027 rate position unchanged, or a Q3 print (~2026-10-27, est.) trimming the more-than-$4.80 FY guide.
-
Invalidates if A weekly close below $105 puts price under both the 2026-08-03 record low of $104.83 and the 2026-08-06 unlock-day low of $105.11, saying the absorption was pre-positioned inventory rather than demand. Secondary: the ~2026-08-12 tranche met with lower highs and no higher low.
-
Invalidates if A weekly close below $556 erases the post-print rebound off the 2026-07-30 low at $524.49 and re-exposes the $520.26 52-week low; secondary: 2026-08-18 opening statements trending toward an adverse liability finding, or the reported $10B Anthropic lease lapsing with no contracted compute revenue at the 2026-10-28 Q3 print.
-
Invalidates if A daily close below $11.80 loses the 2026-08-06 post-print reversal low and puts the 2026-07-29 close of $10.74 back in play. Secondary: the ASM share scheme failing the 2026-08-12 vote, or the 2026-08-18 court hearing adjourning implementation past 2026-08-31.
-
Invalidates if A weekly close below $100 gives up the round-number shelf and puts price into the rising 200-day zone ($91.21 as of 2026-08-04), converting the 2026-08-07 de-rating into a full round-trip. Secondary: the 35% FY2026 revenue-growth guide being trimmed at the Q3 print.
-
Invalidates if A weekly close below $196 forfeits the 2026-08-07 reclaim of the $199–$201 shelf and puts price under the 2026-08-03 low close of $197.01, leaving the rising 200-day at $182.59 as the next reference; secondary condition is July Cass (~2026-08-13) and July ATA (~2026-08-18) both printing negative year over year.
-
Invalidates if A weekly close below $26.50 confirms the August 6 gap-down as distribution — it loses the $26.54 post-print reaction close and the $26.50 median target. Secondary: the ~September 30 appropriations vehicle passing with no enhanced premium tax credit extension attached.
-
Invalidates if A daily close below $143 re-enters the unfilled 2026-08-03/04 gap ($126.97–$143.28) and points back at the $131.45 July shelf; secondary: US commercial growth printing below +120% YoY at the Q3 report against the at-least-+134% FY26 path guided 2026-08-03.
-
Invalidates if A weekly close below $297 forfeits the July and 52-week low at $297.38, which has no 2026 shelf beneath it; secondary confirmation is the ACEA July aggregate (~2026-08-20) printing an EU-wide Tesla decline consistent with Germany's -66.9%, or August ending with no disclosed Optimus unit count from the Fremont line.
-
Invalidates if A weekly close below $134.75 forfeits the post-print base between the 2026-07-29 low of $135.98 and the 2026-08-07 low, and dates the 2026-06-12 high of $176.69 as the top; secondary break if ATA July tonnage (~2026-08-18) prints a third consecutive year-over-year decline.
-
Invalidates if A weekly close below $92 fills the 2026-08-04 follow-on-order gap and gives back the 50-day; secondarily, the 2026-08-20 Needham and 2026-08-26 Jefferies appearances passing with no second distinct hyperscaler AI-ASIC order disclosed, or a discounted common-stock takedown priced off the 2026-07-27 shelf.
-
Invalidates if A weekly close below $235 breaks the floor tested at $235.11 on 2026-07-30 and 2026-08-03 and at $238.51 on 2026-08-06, reopening the unfilled gap toward the $209.68 pre-print close. Secondary: the 50-day at $286.82 left unreclaimed while the complex sells memory beats as it did 2026-08-06.
-
Invalidates if A weekly close below $85 forfeits the 2026-06-02 secondary shelf priced at $85.00 and the entire post-print re-rate; secondary condition is Q3 marketplace growth landing under the ~33% guided on 2026-08-04, or the new 33–34% FY2026 frame being cut.
-
Invalidates if A weekly close below $164 gives up the July–August base and the rising 50-day average last observed at $164.07 on 2026-08-04; secondary is the Q1 FY27 print (~2026-09-02, not company-confirmed) guiding FY27 revenue under the $7.325B low end or Q1 non-GAAP gross margin below the 69.1% floor.
-
Invalidates if A weekly close below $370 forfeits the July shelf defended at $370.83 (2026-07-17), $370.32 (2026-07-29) and $374.61 intraday (2026-08-03), and the 200-day at $366.52 with it. Nearer: a daily close under $392 unwinds the 2026-08-04 gap. Secondary: Q3 FY26 on 2026-09-02 posting AI revenue under the guided $16.0B or book-to-bill materially below ~2.8x.
-
Invalidates if A weekly close below $234 retraces the entire advance off the 2026-06-25 print and loses the $235.67 July base floor. Secondary: Dell guiding ISG materially under the ~75% growth flagged 2026-05-28, at its 2026-09-03 Q2 FY2027 print.
-
Invalidates if A weekly close below $58 takes out the $58.66 low of 2026-07-31 and the $59.47 shelf, leaving the $49.90 52-week low as the next marked level; secondarily, Arc mainnet slipping past its 2026-09-16 target with no replacement date given.
-
Invalidates if A daily close below $353 takes out the 2026-08-05 low of $353.69 and the 2026-07-29 print-day low of $355.81, making the reclaim a failed reversal; a weekly close below $327 unwinds the leg off the 2026-06-03 base close of $327.54. Secondary: the ~October 2026 CMS Star Ratings leaving 4-star-plus enrollment near the 20% of the 2026 ratings.
How we're scored
A confident model is easy. A calibrated one is the point.
When a thesis resolves, it lands in the ledger above as played out or invalidated, dated, with a flag for whether the published trigger actually fired. As outcomes accumulate, the calibration question gets a public answer: does conviction predict the outcome — does SUPREME beat LOW, does a regime call hold? Strictly non-monetary — only whether the falsifiable claim was falsified.
The odds each tier stakes · published in advance
- SUPREME 90%
- HIGH 75%
- MEDIUM 60%
- LOW 50%
Each tier claims this chance of playing out. The Brier score above measures the gap between these stated odds and reality.
Common questions
- Do AI stock predictions actually work?
- There's no honest one-word answer — only a record. orbyd publishes one: every thesis ships a dated, falsifiable kill level, then resolves in public as played out or invalidated, and the whole set is scored with a Brier score. The point isn't a headline accuracy figure but that the calls are checkable: a model's predictions only "work" if they survive the triggers they published, and that's exactly what the scoreboard above measures.
- How is orbyd scored?
- Every thesis ships with a published invalidation trigger. When it resolves it's marked played-out or invalidated — dated, with whether the trigger fired. Across resolved theses we compute a Brier score (forecasting accuracy) and break play-out rates out by conviction tier, so you can see whether SUPREME actually beats LOW. Strictly non-monetary.
- How does a thesis resolve, exactly?
- Mechanically, against the thesis's own published kill level — no discretion. The reference is the real closing price the day the thesis was last stated. The kill is the price in the published invalidation ("close below $X"); the target is a 1:1 realisation of that same published risk (as far above the entry as the kill is below it). Walking real daily closes forward, whichever is hit first decides it: a close at/below the kill is invalidated, a close at/above the target is played-out. "Weekly close" triggers are graded on weekly closes. A thesis whose stop is within 2% (too tight to grade on price) or whose kill isn't a clean price level stays open. Graded only on whether the published claim was falsified.
- Why does the score start in June 2026?
- We rebuilt our research pipeline in June 2026. The public score measures only the calls the new system makes, so it begins fresh and fills as live theses resolve. Earlier theses were written by an approach we've since replaced — scoring them would grade the wrong model. We'd rather start honest and empty than carry a retired system's record.
- What is a Brier score?
- The mean squared error between the probability a conviction tier implies (SUPREME≈0.9 … LOW≈0.5) and the binary outcome. 0 is perfect, 0.25 is a coin flip. For a rough reference band: expert human forecasters in Tetlock's Good Judgment work are commonly cited around 0.08, and the strongest LLM forecasters in 2024–25 evaluations around 0.10. Those benchmarks score different questions, so treat them as scale markers, not a like-for-like comparison.
- What does the track record measure?
- Whether each falsifiable claim was falsified — graded against its own published kill level. It scores the reasoning, not a trade.
- What happens when a thesis is wrong?
- It's marked invalidated, dated, and kept on the public record — with whether the published trigger fired first. Being wrong in public, on the record, is the point: it's what makes the score trustworthy.