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 reference, expert human Superforecasters score around 0.08 and the best LLMs around 0.10.

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.

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 resolves 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 · 8 scored

preliminary — too few resolutions to read much into yet

Brier score

0.215

0 perfect · 0.25 coin-flip

Played-out rate

63%

5 of 8

Calibration error

0.028

lower = conviction tracks reality

Skill vs base rate

+0.08

Brier skill score

Reliability diagram

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%	100%	1
MEDIUM	60%	100%	1
LOW	50%	50%	6

3 of 3 invalidations had the published trigger fire first.

Play-out rate by archetype

Play-out rate by regime at resolution

Risk-on 100% · 4
Neutral 25% · 4

Resolved · 8

TRVI LOW Played out 2026-06-12
DRTS HIGH Played out 2026-06-11
AZZ LOW Played out 2026-06-11
TWST LOW Played out 2026-06-11
INDI LOW Invalidated trigger fired 2026-06-10
APLD MEDIUM Played out 2026-06-09
CLFD LOW Invalidated trigger fired 2026-06-09
INOD LOW Invalidated trigger fired 2026-06-09

Open commitments · 37 · soonest to resolve first

HOOD Robinhood Markets, Inc. LOW catalyst due

Invalidates if revisit on a weekly close back above the 20-EMA (mid-$80s) reclaiming the ~$87-88 higher-low shelf, or a CLARITY Act floor-vote scheduling
AVGO Broadcom Inc. MEDIUM catalyst due

Invalidates if Weekly close below the 2026-06-04/05 post-Q2 reaction low AND under the 20-EMA (sell-the-news becomes a structural de-rate); OR FY27 AI revenue guide later cut below $100B; OR a confirmed Google/Meta next-gen TPU socket loss to MRVL.
META Meta Platforms, Inc. LOW catalyst due

Invalidates if Confirmed multi-tens-of-$B equity raise priced at a discount that the tape sells (validates the capex-dilution bear case); OR weekly close below the 20-EMA on rising volume; OR Q2 (~2026-07-29) ad rev <+15% YoY vs Q4 +21%; OR custom-silicon theme (AVGO/MRVL) breaks into SATURATED.
MSTR Strategy Inc LOW catalyst due

Invalidates if Avoid-stance flips to long ONLY on all three: (1) BTC weekly close back above $84K, (2) an 8-K showing net BTC accumulation with zero sell-language, (3) MSTR weekly close above its 20-EMA with mNAV premium restored above issuance-accretive level. Until then every bounce is a falling-knife trap.
PLTR Palantir Technologies Inc. MEDIUM catalyst due

Invalidates if Failure to reclaim the 200-DMA within ~2-3 weeks plus a daily close below the early-June profit-taking swing low on >1.5x avg volume; OR a DIA ruling formally barring PLTR from the contract bid; OR Q2 US commercial growth decel <45% YoY (~Aug).
AXTI AXT Inc LOW catalyst due

Invalidates if Lost the $90 base on 6/5 (-16% day, >$22M CEO insider sale); no buyable setup until a reclaim and hold above $100 on normalizing volume with a higher-low. Thesis breaks outright on Q2 (~7/29) revenue below the $34M guide, InP backlog falling QoQ from the $100M record, or denial of pending US InP export permits.
CEG Constellation Energy Corporation LOW catalyst due

Invalidates if Weekly close below $240 (200-DMA + post-Calpine gap fill); OR FERC rejects the Amazon/Susquehanna co-location (docket ER24-2172) or PJM slips Crane interconnection to 2031; OR 2026 adj-EPS guide cut below the reaffirmed $11 floor at Q2 (~2026-08-05).
CRCL Circle Internet Group, Inc. LOW catalyst due

Invalidates if Weekly close below $77 (Compass Point magnet) on above-average volume; OR failure to reclaim $85 (Mizuho PT) within two weeks of the Mastercard catalyst, confirming sold-the-news; OR Q2'26 print shows no USDC settlement-volume traction from Mastercard/Nium.
LITE Lumentum Holdings Inc. MEDIUM catalyst due

Invalidates if Daily close below ~$840 forfeits the 2026-06-01 $866.97 signal-bar reclaim; OR a weekly close under the rising 20-EMA; OR a peer optical name (COHR/MRVL/AAOI) losing its 50-day while LITE diverges lower any one flips momentum to broken.
COIN Coinbase Global, Inc. LOW catalyst due

Invalidates if Long re-arms only on BTC reclaim of ~$68K AND a COIN weekly close above $185 (back over the $180 bear anchor + 20-EMA), ideally with 3+ days of net-positive spot-ETF flows. BTC losing $60K opens $55K and a weekly close below the $142 Baird floor pushes the no-touch zone lower; do not buy oversold RSI alone.
SMR NuScale Power Corporation LOW catalyst due

Invalidates if South Korea $200B package finalizes with no named NuScale allocation, or a weekly close below the $8.85 52-wk low confirms the breakdown. Engage long only on a named SK capital figure (or binding offtake ≥$100/MWh for ≥250 MWe) PLUS a higher-low base reclaiming the 20-EMA (~$12).
VST Vistra Corp. LOW catalyst due

Invalidates if Weekly close below $132 (May-2026 swing low) on >1.3x avg volume; OR failure to reclaim the 50-DMA (~$160-165) by end-July while CEG/TLN/NRG roll over; OR a hyperscaler Q2 capex guide-cut that re-rates the power complex lower.
TSEM Tower Semiconductor Ltd. MEDIUM catalyst due

Invalidates if Weekly close back below the ~$255 May-13 earnings-gap base (failed post-beat breakout — verify exact 20-EMA before sizing), or a Q2 guide walk-back / RF-SOI inventory-reset signal from QCOM/AVGO handset commentary.
ASTS AST SpaceMobile, Inc. MEDIUM catalyst due

Invalidates if retry at the next scheduled review
CRWV CoreWeave, Inc. LOW catalyst due

Invalidates if Weekly close below the ~$104 post-Q1 low (5/20) = base failed, opening the DA Davidson $100 → Bernstein $67 air pocket. Also invalidating: the $850M data-center junk bond prices wider than the 9.750% notes (credit refusing the recovery), or a second consecutive cohort guide-down at the Aug 11 print.
OKLO Oklo Inc. LOW catalyst due

Invalidates if Avoid-thesis flips bullish on a daily reclaim of the $66–$70 shelf (above Goldman $66 PT and the insider-sale range) on expanding volume with NO fresh 424B/ATM takedown. Otherwise a weekly close below the 2026-05-18 post-Q1 low, a 424B takedown off the $400M shelf, or any new insider Form 4 within 2 weeks of June 1 confirms the breakdown.
QBTS D-Wave Quantum Inc. LOW catalyst due

Invalidates if Daily close back below the ~$20 May-20 pre-announcement base = government-bet gap filled and failed. OR a new ATM/shelf take-down 8-K beyond the $100M Commerce issuance. OR QTUM ETF rolls over while the quantum cohort (IONQ/RGTI/QUBT) prints lower lows in unison (theme→DEAD).
RGTI Rigetti Computing, Inc. LOW catalyst due

Invalidates if Daily/weekly close below ~$20 (loses the long-term average and gives back the 2026-05-21 gov-stake gap), confirmed by QBTS failing to reclaim its 20-EMA, OR any ATM-issuance 8-K filed during strength (front-runs basket peak by 5–10 sessions).
ASML ASML Holding N.V. MEDIUM catalyst due

Invalidates if Weekly close below the 20-week EMA (50% scale-out) or Q2 guide worse than −18% sequential confirming demand softening; on entry side, do not act until a 20/50-DMA pullback with RVOL>1.5 while not holding a correlated foundry name (TSM).
NNE Nano Nuclear Energy Inc. LOW catalyst due

Invalidates if Daily close below $22 forfeits the post-NRC base and opens the $18.93 52-week low; alternatively any 424B/8-K disclosing the $400M ATM selling into rallies confirms catalysts are priced and dilution is absorbing demand.
QUBT Quantum Computing Inc. LOW catalyst due

Invalidates if Flip to interest only on a weekly close back above the ~$13–14 200-DMA driven by a QUBT-specific catalyst (TFLN customer 8-K, direct federal award, or sustained options skew). Hard skip on any 424B5/S-3 ATM filing or a weekly close losing the ~$10 shelf.
PL Planet Labs PBC LOW catalyst due

Invalidates if Holding below the lost ~$42 breakout shelf keeps the structure rolled-over (no-touch); a daily close below the ~$32 post-earnings low confirms the downtrend toward DEAD-theme treatment. A long only re-validates on a weekly reclaim of $42+ on volume with a higher low.
BFLY Butterfly Network, Inc. LOW resolves in 4d

Invalidates if Daily close below the ~$4.40 breakout base (loses the rising 50-EMA structure off the $1.33 low), OR Q2 2026 revenue below the $27M guide floor / any cut to the $117–121M FY guide signals the June move was a fading retail spike, not a durable narrative leg.
TSLA Tesla, Inc. LOW resolves in 4d

Invalidates if Confirmed pass if TSLA fails to reclaim its 2026-06-02 pre-gap high within ~2 weeks, the Austin Robotaxi fleet shrinks below 20 cars, or any FSD class-cert/HW3-retrofit ruling lands. Re-engage only on a reclaim of that high plus a dated TSLA-owned robotics accelerant (Optimus milestone or Robotaxi rides inflection on an expanding fleet).
UUUU Energy Fuels Inc. LOW resolves in 8d

Invalidates if Daily close below $14 extends the failed-spike downtrend; OR ASM scheme vote fails at the 2026-06-22 Perth meeting / deal terminated; OR uranium spot holds below $78/lb for two consecutive weeks; OR any bounce fails to reclaim and hold the lost ~$16-17 mid-May shelf (no higher low).
SYRE Spyre Therapeutics, Inc. MEDIUM resolves in 16d

Invalidates if SPY002 "mid-2026" (~2026-06-30 est.) Part A induction shows weak within-patient efficacy or class-discordant TL1A safety; OR weekly close below the $62 April offering shelf on >2x 20d avg volume; OR Merck tulisokibart (ATLAS-UC) or Roche RVT-3101 prints superior anti-TL1A Phase 3 data.
CCJ Cameco Corporation LOW resolves in 17d

Invalidates if Weekly close below $100 (loses the May ~$104 low), or uranium spot back under $80/lb, or a Q2 guide cut to the 19.5–21.5M lb range on the 2026-07-31 print. A second rejection at ~$115 confirms the lower high and voids the reclaim setup.
IREN IREN Limited MEDIUM resolves in 18d

Invalidates if Weekly or decisive daily close back below the 2026-05-26 breakout gap on >1.5x avg volume (failed breakout); OR disclosure of a Microsoft contract delay/descope; OR a new ATM equity raise >$300M priced below spot (signals the $3.65B IG facility was insufficient).
AEHR Aehr Test Systems LOW resolves in 23d

Invalidates if Weekly close below the $80 breakout shelf (opens $60-65 base), or Q4/FY26 print (~2026-07-07) that walks back the non-GAAP profitability guide or shows H2 bookings momentum stalling.
HUT Hut 8 Corp. HIGH no dated catalyst

Invalidates if a daily close below the published invalidation level
SHLS Shoals Technologies Group, Inc. HIGH no dated catalyst

Invalidates if a daily close below the published invalidation level
TSM Taiwan Semiconductor Manufacturing Company Ltd. HIGH no dated catalyst

Invalidates if a daily close below the published invalidation level
OKTA Okta, Inc. MEDIUM no dated catalyst

Invalidates if Daily close below the post-earnings breakout shelf / rising 20-EMA (~$108), OR the cybersecurity/identity theme classifier flips off ACCELERATING, OR a peer (CRWD/PANW/ZS) guides demand down, OR a second sell-side downgrade follows Mizuho's 06-02 cut.
SNOW Snowflake Inc. MEDIUM no dated catalyst

Invalidates if Daily close below the published invalidation level loses the post-earnings gap shelf / rising 20-EMA, fills the gap and breaks the inflection re-rate structure.
AMSC American Superconductor Corporation LOW no dated catalyst

Invalidates if Weekly close below $40 (loses post-earnings shelf + April base pivot; opens path to the 200-DMA / $24.87 low). Also dead if the early-Aug Q1 print shows 12-mo backlog contracting from ~$280M or grid orders rolling over, or grid-power-transmission downgrades to SATURATED/DEAD.
IONQ IonQ, Inc. LOW no dated catalyst

Invalidates if Weekly close below the rising 20-EMA (refresh — no live feed); Q2 revenue sequentially below Q1 $64.67M or any FY guide cut from $260–270M; or IONQ excluded from final $2B CHIPS equity-stake awards while peers are included (unwinds the 2026-05-26 pop).
SPCX SpaceX (Space Exploration Technologies Corp.) LOW no dated catalyst

Invalidates if A read on entry timing. Constructive only on a post-lock-up higher-low base that holds above the $135 offer on expanding volume, or a first public quarter that validates Starlink's margin trajectory. More cautious if the stock loses the $135 offer on heavy volume (the day-one pop unwound and the float overhang won), if Starlink growth decelerates, or if a Starship setback pushes the optionality further out. Structural second-order effect: with the real thing now public, the 'only way to own SpaceX' premium in the proxy basket (RKLB/ASTS/RDW/LUNR and DXYZ's NAV premium) should compress.

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.

What an invalidation trigger is Why trust it

Common questions

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 reference, expert human Superforecasters score around 0.08 and the best LLMs around 0.10.
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.