Research record / 61 sleeves + the BLOCK reference

THE CORPUS

The spine is the survivor. The corpus is the record around it: live systems, validation sleeves, active research, concepts awaiting proof, and null results kept visible because failure is evidence.

New to the corpus?

Follow one sleeve from question to live spine.

BASE provides a guided path through the thesis, rejected predictive claims, surviving mechanisms, governed execution, and current role.

Start with BASE →
The corpus / by the numbers

A research program, scored in the open.

Every repository sits on one governance ladder: idea, test, validated keeper, and governed live rail. Most ideas do not make it, and the ones that fail stay published. The graveyard is the largest tier by design because the standard is visible rejection, not survivorship.

Active research / 20

LIVE / VALIDATION

BASE is the governed spine. LIVE PATH and LIVE DRIVER systems have working governed implementations; VALIDATION work has survived research but has not yet earned that bar.

001BASEEngine / governed spine / systematic risk structureWhy it survived: Predictive-return tests failed, but diversification and trend premia persisted out of sample. The spine has also completed the governed path from data through orders and reconciliation.● LIVE 002BRICSGulf-pocket EM diversifier / dollar-regime overlayWhy it survived: The broad BRICS basket was mostly emerging-market beta, but the Gulf pocket supplied genuinely lower correlation. A dollar overlay materially reduced drawdown and improved risk-adjusted return.● LIVE PATH 003BONDSBond–equity regime overlay / signal emitterWhy it survived: The stock–bond correlation regime was detectable ahead of time even though the other predictive components were null. That makes BONDS useful as a portfolio overlay and signal emitter, not a standalone trade.● LIVE PATH 004BROADManaged QQQ trend / volatility-target keeperWhy it survived: Leveraged wrappers lost their appeal once risk was matched. Unlevered QQQ with trend and volatility targeting preserved participation while controlling drawdown. The edge was exposure management, not leverage.● LIVE PATH 005BOOMMega-cap momentum / capped volatility targetWhy it survived: Twelve-to-one momentum captured persistent mega-cap leadership. Position caps and volatility targeting made the result governable, though dependence on a few winners keeps concentration risk explicit.● LIVE PATH 006BENCHMARKMarket-internals regime gate / volatility-timing emitterWhy it survived: Market internals mostly coincided with stress rather than leading it, but volatility regimes persisted long enough to inform sizing. BENCHMARK therefore emits a bounded risk input instead of claiming new alpha.● LIVE PATH 007BULLLong-growth book / validated rate-regime overlayWhy it survived: Raw growth exposure was largely beta. Conditioning it on the rate regime improved Sharpe, skew, and drawdown enough to make that beta ownable within a governed portfolio.● LIVE DRIVER 008BRIDGEWATERAll-Seasons risk parity / governed allocationWhy it survived: All Weather remained a valid low-volatility compounder, although it was highly correlated with the existing spine. The recognized allocation was retained as a governed book; a redundant overlay was not.● LIVE DRIVER 009BLACKSTONEListed private-equity factor / market-regime overlayWhy it survived: Listed private equity proved to be leveraged, procyclical equity and credit exposure, not a hedge. A market-regime overlay cut drawdown enough to turn that transparent factor into a usable governed sleeve.● LIVE DRIVER 010BRASHGoverned aggression / crypto trend / growth-vs-ruinWhy it survived validation: Aggression multiplied both growth and ruin; full Kelly and beyond deepened drawdown without improving the underlying edge. Fractional Kelly with volatility targeting preserved the return source within explicit risk limits.VALIDATION 011BOUNCERegime-gated cross-sectional reversalWhy it survived validation: Fading the index failed, but ten-day cross-sectional reversal remained modestly positive in choppy regimes and near market-neutral. The edge is relative dislocation, not a broad-market call.VALIDATION 012BLUEGrowth / thematic camp rotationWhy it survived validation: Static ownership of either thematic camp underperformed. Relative trend between the camps added value, making rotation rather than permanent allegiance the investable finding.VALIDATION 013BROWNReal-asset / conservative camp rotationWhy it survived validation: The conservative and real-asset camp was not a superior static holding. Its value emerged only when relative trend rotated exposure between Brown and Blue regimes.VALIDATION 014BITDOLLARGoverned BTC / ETH trend and volatility targetingWhy it survived validation: Trend and volatility targeting retained much of crypto's upside while sharply reducing its deepest loss. The dollar-axis thesis was rejected; disciplined exposure control was the survivor.VALIDATION 015BOTTOMDrawdown bounce / liquidity and cap-ladder disciplineWhy it survived validation: Moving down the capitalization ladder added idiosyncratic risk and trading cost rather than edge. Liquid large-cap drawdown bounces were the only version robust enough to keep.VALIDATION 016BOREBeta-hedged market-neutral momentumWhy it survived validation: Hedging removed the flattering market beta but left a smaller, real momentum residue. Used as portable alpha over an index, it improved the portfolio rather than pretending neutrality was free.VALIDATION 017BLUNTShort-horizon tactical / crude-to-refiner survivorWhy it survived validation: Five short-horizon ideas failed; a one-day crude-to-refiner lead–lag remained. The result rewards observed follow-through in a linked market rather than trading the headline itself.VALIDATION 018BLEEDRegime-spanning tail insuranceWhy it survived validation: Static long volatility carried too much annual cost. A diversified basket covering both inflationary and deflationary crises delivered negative crisis correlation with tolerable carry, so it earned a role as insurance.VALIDATION 019BEYONDReversal-to-momentum horizon transitionWhy it survived validation: Very short horizons reversed while horizons beyond roughly twenty days became momentum. Long winners survived; acceleration did not, and overlap with BOOM limits the sleeve's independent weight.VALIDATION 020APACCross-country relative strength / FX disciplineWhy it survived validation: Unhedged regional exposure was mostly US beta plus FX drag. Cross-country relative strength was the durable pulse, provided the implementation remains hedged or dollar-neutral.VALIDATION
ENTER LABS →
Meta-sleeves / 9

CAPSTONES: THE ALLOCATION LAYER.

Capstones are meta-sleeves, themselves family sleeves one layer up. They form the allocation layer that sits atop the strategy sleeves and consumes their validated keepers as ingredients. They are alternatives, not additions (stacking them is redundant at +0.92 correlation). Seven are wired to governed live rails; two are research-only controls.

021BREAKTHROUGHRisk-parity capstone / the keeper (Sharpe +1.16)Why it works: Risk parity over the validated keepers, paired with conditional insurance, converted many separate findings into one diversified and governable portfolio rather than another prediction.● GOVERNED 022BRILLIANTOptimizer capstone / min-variance can't beat 1/NWhy it works: Minimum-variance and maximum-Sharpe optimization cleared the index hurdle, but neither reliably beat risk parity. Covariance information was useful; precise return forecasts were not.● GOVERNED 023BOSSYConviction / leverage capstone (return, not edge)Why it works: Concentrating conviction worsened risk-adjusted results. Modest leverage on diversified risk parity raised return without inventing a new edge, while making the deeper drawdown explicit.● GOVERNED 024BELIEVERFaith benchmark / buy-and-hold the keepersWhy it works: Buying and holding the keepers was already strong, but drift recreated concentration. Rebalancing by risk budget improved the portfolio, showing that allocation discipline matters after selection.● GOVERNED 025BEMUSEDRandom-allocation null / the edge is the keepersWhat it established: Random allocations of validated keepers were already strong, while BREAKTHROUGH ranked near the top of the distribution. Most value came from ingredient quality; allocation skill was real but smaller.RESEARCH 026BRIGADEExpanded capstone / curated void-fillers (breadth)Why it works: Adding carefully selected, low-correlation void-fillers improved risk-adjusted return. Breadth helped only when each addition filled a portfolio need rather than repeating an existing bet.● GOVERNED 027BOUNDLESSKitchen-sink capstone / more is not betterWhat it established: Adding every near-keeper diluted the portfolio because correlated ingredients could not manufacture diversification. The control proves that curation, not maximal breadth, is the allocation edge.RESEARCH 028BASTIONInsurance capstone / a null rehabilitated by −corrWhy it works: A weak standalone bear sleeve became valuable at a small portfolio weight because its negative correlation reduced drawdown. Portfolio contribution, rather than standalone return, earned its place.● GOVERNED 029BALSAMICCore + managed-futures trend satellite (Sharpe +1.26)Why it works: A managed-futures trend satellite complemented the risk-parity core across different regimes. The combined book improved resilience and passed the governed dry-run without duplicating the core.● GOVERNED
ENTER CAPSTONE →
Research / 7

WORK STILL BEING RESOLVED.

Research repositories remain open while their evidence, implementation, or production case is still being developed.

030BEARRegime-gated short / costly incomplete tail insuranceWhy it remains research: The gated short still paid persistent carry and covered only part of major crashes. A coincident regime signal cannot reliably time the short, and BLEED already provides broader protection.RESEARCH 031BALANCEDValue factor / rate sensitivity / margin-of-safety testWhy it remains research: Value behaved largely as a rate-sensitive factor: volatility fell, but crashes remained and growth led over the test period. It may be a conditional ingredient, not a durable standalone edge.RESEARCH 032BOGLELow-cost index hurdle / factor and fund comparisonWhat it established: None of the tested Vanguard tilts reliably beat the low-cost index on a risk-adjusted basis. Apparent winners were mostly beta, making BOGLE the hurdle every active idea must clear.RESEARCH 033BUFFETTQuality / minimum-volatility / value / moat blendWhy it remains research: The factor decomposition was coherent, but neither the public factors nor Berkshire reliably beat the index in the recent sample. Selection and cheap float appear central, and the proxies do not fully capture them.RESEARCH 034BASKETExchange funds / tax-deferral and estate vehicleWhy it remains research: The vehicle creates no alpha; its value is tax deferral under narrow conditions: large embedded gains, low fees, and a long holding period. Suitability depends more on the owner than the market.RESEARCH 035BLOCKDerivatives reference / catalog / pricers / analyticsWhat the research found: Cross-asset lead–lag relationships were too weak to trade reliably. The durable output is the governed four-block allocation, supported by an open derivatives reference rather than a single predictive sleeve.RESEARCH 036BOMBS AWAYFragility monitor / tail-hedge sizing / not a crash timerWhy it remains research: The fragility tells describe building stress but do not predict forward drawdowns or sudden shocks. The monitor survives only as a bounded tail-hedge sizing input, not a market-timing switch.RESEARCH
Concept / 6

THE NEXT TESTS.

Concept repositories make the hypothesis visible before evidence earns a research, validation, or null designation.

037BELT-TIGHTENERSInflation domino / sector and global correlation tighteningWhy it is still a concept: The proposed inflation domino is plausible, but the test still needs breakeven and CPI data plus enough global history to separate a durable cascade from one episode.CONCEPT 038BATTEREDDistressed survivors / fallen angels / left-tail costWhy it is still a concept: The hypothesis depends on surviving distress through payout, but liquid ETFs are poor proxies and survivorship is severe. Tradable data has not yet isolated the intended carry.CONCEPT 039BUREAUCRATSCongressional trading / disclosure and factor testWhy it is still a concept: Public congressional-trading funds have too little history to distinguish information from a technology tilt. A credible test needs transaction-level disclosure data and lag-aware execution.CONCEPT 040BETSGambling complex / house premium or consumer betaWhy it is still a concept: The true house edge lives in private prediction and sportsbook data, while listed companies mix that economics with consumer beta. The accessible proxies have not yet isolated the premium.CONCEPT 041BIGBROTHERGovernment equity stakes / support versus dragWhy it is still a concept: State ownership could create a backstop premium or an operating drag, but investable events are rare and heterogeneous. The test requires a hand-curated event history rather than a broad industry proxy.CONCEPT 042BUYOUTSSponsor-backed exits / post-lockup driftWhy it is still a concept: Private-company performance is not directly observable. The testable residue is post-lockup sponsor behavior, but it still needs a clean event set and realistic borrow assumptions.CONCEPT
Null / graveyard / 20

THE NULL IS PART OF THE WORK.

These ideas did not produce a durable, executable edge. They remain public so the failed thesis, constraint, or risk finding can inform the next test.

043LATAMLatin America regional sleeveWhy it failed: While US and commodity beta dominated, the sleeve carried the worst tail and unhedgeable FX, and rotation reduced rather than improved the result. Geography within equities proved to be more beta with a fatter tail.NULL 044BENEFACTORSForeign Treasury creditors / causality runs backwardWhy it failed: Foreign Treasury flows were weak and coincident; markets and economic conditions often moved before the creditor data. The series is a macro read, not a tradable leading signal.NULL 045BINDLiquid-alt hedge-fund replication / diluted betaWhy it failed: Liquid hedge-fund replicators delivered watered-down market beta while the true pod-level edge remained private. The useful defensive residues were already represented elsewhere in the family.NULL 046EMEAUS beta / FX drag / no country-rotation edgeWhy it failed: Eleven country ETFs collapsed into fewer than two effective bets, dominated by US beta and currency drag. Rotation was negative, so regional labels did not produce real diversification.NULL 047BULKDefense sector / systematic stock-picking nullWhy it failed: Defense stocks behaved as equity beta with a compelling narrative, while geopolitical shocks were priced too quickly for the tested rules. Trend filters added whipsaw rather than protection.NULL 048BRITTLENear-expiry far-OTM options / short-vol trapWhy it failed: The volatility risk premium was real, but naked short volatility eventually surrendered the carry in the tail. Defined-risk versions then underperformed the index after paying for protection.NULL 049BIOBiotech dispersion / no systematic price signalWhy it failed: FDA and clinical outcomes created genuine dispersion but were not predictable from price alone. Without domain-specific fundamental data, the sleeve remained a risk diagnostic rather than an edge.NULL 050BASELUS banks / regulated macro factor / no selection alphaWhy it failed: Seventeen banks reduced to roughly one and a half effective bets, driven by the economy and the short end of the curve. Stock selection and timing added no durable alpha.NULL 051BELTWAYParty-aligned baskets / macro dominates narrativeWhy it failed: Party-aligned baskets often ran opposite the political story as rates and macro conditions dominated. Recent winners also reverted, leaving no stable partisan signal.NULL 052BURRYConcentration / equity beta / anti-systematic styleWhy it failed: The public style was mostly equity beta; hated-name reversal failed, froth shorts produced ruinous drawdown, and concentration worsened risk-adjusted return. Discretion did not survive systematization.NULL 053BLANKSPAC decay / borrow and right-tail constraintWhy it failed: Post-merger decay was visible but difficult to short because borrow and the occasional explosive winner dominated execution. The safer trust-carry trade required deal data the public proxy did not provide.NULL 054BRUTE FORCEPropped-up-by-flows sleeve / fade rejectedWhy it failed: Price and volume alone could not distinguish artificial support from ordinary momentum. Fading the pattern lost money; a credible test would require positioning and flow data.NULL 055BLURREDDeliberate low correlation / floor rather than edgeWhy it failed: Equity correlations shared a common beta and rose precisely when diversification was needed. Low historical correlation was unstable, while pair spreads lacked durable cointegration.NULL 056BACKSLIDERSFalling-knife shorts / long bounce survivesWhy it failed: Recent losers tended to bounce, so systematic shorting lost. True terminal declines were obscured by delistings and missing data, implying a fundamental specialist edge rather than a price rule.NULL 057BUBBLEAI complex / crowded-factor risk diagnosticWhy it failed: AI names formed one crowded growth factor, but price alone could not time its reversal. The practical response is to size and insure the exposure, not mechanically short the story.NULL 058BRACEVol-risk-premium harvest / short-vol is the insurer’s sideWhy it failed: Listed short-volatility products lagged the index and carried severe negative skew. Regime gates merely exchanged return for drawdown because coincident signals could not defuse the tail.NULL 059BOUNTYCross-asset carry / crash-risk premium, not alphaWhy it failed: Liquid carry proxies offered weak excess return, equity correlation, and negative crisis skew. The apparent yield was compensation for crash exposure, not portable alpha.NULL 060BITEMerger-arb / event-driven / watered-down excessWhy it failed: The public merger-arbitrage proxy earned little over cash while retaining negative skew and crisis correlation. The real edge requires deal selection, leverage, and access the ETF cannot reproduce.NULL 061BRIDGEStat-arb / pairs / ideal vehicle, no alpha to carryWhy it failed: The pairs framework achieved neutrality but earned approximately nothing after costs. The vehicle was sound; the historical mean-reversion signal had decayed.NULL 062BLOATEDNVDA capstone bolt-on / higher Sharpe, worse concentrationWhy it failed: Adding NVIDIA raised backtested Sharpe but repeated growth exposure already embedded in the capstones and deepened drawdown. The apparent improvement was concentration and sample-period luck, not diversification.NULL
VIEW ALL REPOSITORIES ↗
The website interprets the research program; GitHub remains the source for current status, methodology, code, validation, and repository history.