Multiple Testing & the Deflated Sharpe Ratio
Best-of-160 BTC rule: IS Sharpe 1.14 is only 1.24× the pure-noise expectation : DSR = 0.70 (fail).
Best-of-160 in-sample Sharpe1.14
RESEARCH · WAREHOUSE · OPEN9 STRATEGIES · FOUR FACTORS · 12 SOURCES
CHRISTIAN.T.MACIONUTC+89 STRATEGIES12 DATA SOURCESOWNER-VERIFIED
9 strategies. Four factors. 12 sources.
A market-data view of the research book. No narrative. the table is the narrative. Every number on this page traces back to a public dataset and a reproducible notebook.
Statistics that survive discipline are statistics that didn't need to lie.
quant strategies
9
all open-source on GitHub
eval gates
31
G1 to G31 mechanical harness
cert credentials
102
STA Tier-1 + 4 cert groups
shipping projects
15
6 AI · 9 quant
Asset class × regime coverage
Where each strategy works, and where it doesn't.
Four asset classes × four regime states. Each cell lists the strategy (or two) that has a defensible claim on that intersection. drawn from the 9 quant projects on this page. Empty cells are intentional: a honest book has regimes it doesn't claim.
Bulldirectional up, low vol
Beardirectional down, rising vol
Rangemean-reverting, low vol
Crashtail event, vol expands
Equities
Crypto
All 9 strategies
The book, card by card.
One card per quant project. Headline metric, one-line summary, and a canonical link to the long-form note on/projects/{slug}.
01Sat Sep 20 
02Wed Oct 15 Cross-Sectional Momentum (18 coins)
IS Sharpe 0.91 → OOS −0.03 : an honest decay; bootstrap CI straddles zero.
In-sample Sharpe0.91
03Sat Dec 20 Time-Series Momentum + Vol Targeting
Vol-targeting lifts Sharpe 0.27 → 0.39 and halves max DD (−62% → −30%).
Sharpe (raw)0.27
04Sun Jan 25 The Variance Risk Premium (VIX vs Realized)
Implied > realized 85% of 36 yrs; predicts returns, Newey-West t = +6.5.
Years VRP > 085%of 36 yrs
05Wed Feb 18 Pairs Trading via Cointegration (BTC / ETH)
ADF −2.11 (not cointegrated), half-life 208 d : the fade loses, as the test predicts.
ADF test statistic−2.11fail to reject unit root at 5%
06Sun Apr 05 Crypto Funding-Carry
Funding +11.9% annualized premium; fade IS 1.11 → OOS −0.05 (decayed post-2023).
Annualized funding+11.9%BTC perp, ~30-day centered mean
07Sun May 10 Macro / Volatility-Regime Overlay
Vol-managed exposure: Sharpe 0.64 → 0.67, max DD −58% → −42%.
Sharpe (base)0.64
08Thu May 28 Backtest Engine + Cost Model
High-turnover signal wins gross (0.20 > 0.13) but loses net of cost : break-even 20 bps.
Sharpe (gross)0.20
09Thu Jun 18 Look-Ahead Bias Audit (the shift test)
A leak inflates Sharpe 0.59 → 5.07; the shift test exposes it as 88% phantom.
Clean Sharpe0.59
Strategy × Factor decomposition
The four numbers that decide whether a strategy survives review.
Each strategy is scored against the four most-borrowed multiple-testing diagnostics: raw Sharpe, Probabilistic Sharpe Ratio (PSR), Deflated Sharpe Ratio (DSR), and Probability of Backtest Overfitting (PBO). Empty cells report n/an/a the metric was not computed in the source notebook, not omitted to hide a bad number.
| Strategy | Sharpe | PSR | DSR | PBO | Notes |
|---|---|---|---|---|---|
| 01Multiple Testing & the Deflated… | 1.14 | n/a | 0.70fails the multiple-testing adjustment | n/a | 2/4 |
| 02Cross-Sectional Momentum (18 coins) | 0.91 | n/a | n/a | n/a | 1/4 |
| 03Time-Series Momentum + Vol Targeting | 0.27 | n/a | n/a | n/a | 1/4 |
| 04The Variance Risk Premium (VIX… | n/a | n/a | n/a | n/a | 0/4 |
| 05Pairs Trading via Cointegration… | ≈ 0as the test predicted | n/a | n/a | n/a | 1/4 |
| 06Crypto Funding-Carry | 1.11 | n/a | n/a | n/a | 1/4 |
| 07Macro / Volatility-Regime Overlay | 0.64 | n/a | n/a | n/a | 1/4 |
| 08Backtest Engine + Cost Model | 0.20 | n/a | n/a | n/a | 1/4 |
| 09Look-Ahead Bias Audit (the shift… | 0.59 | n/a | n/a | n/a | 1/4 |
Data source registry
Every dataset behind the 9 strategies and live terminal.
16 sources. 4 live feeds probe the upstream in real time; the rest are static research baselines. No proprietary feeds, no employer scrapes, no platform API keys. Each line carries its URL pattern and licence so a hiring manager can audit the data lineage with curl.
Registry last verified · live feeds re-probed every 60s.
| Source | URL | Cadence | Last sync | Status | Licence |
|---|---|---|---|---|---|
| Yahoo Finance (delayed quotes, free tier)4 equity + 1 ETF quotes via Cloudflare Worker proxy. | https://query1.finance.yahoo.com/v7/finance/quote?symbols=SPY,QQQ,DIA,BTC-USD | 60s | pending | Yahoo terms (delayed >=15 min, non-commercial) | |
| CoinGecko free tier (no API key)4 spot quotes via Cloudflare Worker proxy. | https://api.coingecko.com/api/v3/simple/price?ids=bitcoin,ethereum,solana,binancecoin | 60s | pending | Free tier, attribution required | |
| ECB SDMX (reference exchange rates)Daily reference rates; edge-cached for 1h. | https://data-api.ecb.europa.eu/service/data/EXR/D.USD+GBP+JPY.EUR+USD | 1h | pending | Public (European Central Bank, free re-use) | |
| GDELT 2.0 (global events)15-min event publish cadence; edge-cached for 15 min. | https://data.gdeltproject.org/gdeltv2/lastupdate.txt | 15min | pending | Public (GDELT Project, free re-use) | |
| FRED-equivalent (public macro time-series)US macro series, daily/quarterly frequency. | https://fred.stlouisfed.org/ to public CSV download per series | static | macro | Public domain (US Federal Reserve release) | |
| CBOE VIX Historical (free)Used by project 04 and 07. variance risk premium + regime classifier. | https://cdn.cboe.com/api/global/us_indices/daily_prices/VIX_History.csv | static | equity-deriv | Free for non-commercial redistribution | |
| Yahoo Finance (delayed quotes, free tier)Daily OHLCV equity / ETF data for cross-sectional and pairs work. | https://query1.finance.yahoo.com/v7/finance/download/{ticker} | static | equity-deriv | Yahoo terms (delayed ≥15 min, non-commercial) | |
| CFTC Commitment of Traders (COT)Positioning baseline for any carry/liquidity signal. | https://www.cftc.gov/dea/newdeacom.html. annual + weekly disaggregated reports | static | cftc | Public domain (US Government publication) | |
| CoinGecko free tier (no API key)Daily and intraday crypto price history across the 18-coin universe. | https://api.coingecko.com/api/v3/coins/{id}/market_chart | static | crypto-onchain | Free tier, attribution required | |
| Public-domain academic datasets (Bailey, López de Prado, Chan)Replicated baselines for DSR / multiple-testing papers. | https://www.quantresearch.org/ to published companion data | static | macro | Academic use, attribution required | |
| Treasury par-yield curve (US Treasury)Term-structure ground truth for any rates / curve work. | https://home.treasury.gov/resource-center/data-chart-center/interest-rates/daily-treasury-rates.csv | static | rates | Public domain (US Treasury publication) | |
| EIA petroleum spot prices (weekly)Commodity baseline for cross-asset regime overlays. | https://www.eia.gov/dnav/pet/pet_pri_spt_s1_d.htm | static | commodities | Public domain (US EIA) | |
| Binance public klines (delayed REST)Daily OHLCV on the BTC/ETH/SOL universe for project 02/03/05. | https://api.binance.com/api/v3/klines?symbol=BTCUSDT&interval=1d | static | crypto-onchain | Public REST endpoint, no key required | |
| Binance public funding-rate history (read)Source for project 06 funding-carry study. | https://fapi.binance.com/fapi/v1/fundingRate?symbol=BTCUSDT | static | funding | Public REST endpoint, no key required | |
| FRED-equivalent: World Bank open data (macro cross-country)Cross-country macro for sanity-checking single-country signals. | https://data.worldbank.org/indicator. open data API | static | macro | Open licence (CC-BY 4.0) | |
| FRED-equivalent: BIS real-effective-exchange-rateFX basket reference for any carry / FX-overlay discussion. | https://data.bis.org/topics/RPP | static | fx | Public (Bank for International Settlements, re-use permitted) |
Public data only. NDA-clean by construction.
Every dataset on this page is enumerated in the registry above. No employer-internal feeds. No scraped proprietary sources. No client data. A hiring manager at a regulated shop can audit the data lineage with curl and arrive at exactly the same numbers. The repository contains the notebooks that produced every figure shown here.
Want the methodology behind the table?
The 31-gate evaluation harness (G1 to G31) on /methodology explains how each strategy got its gate verdicts, and why most of the cells in the factor table are filled with an em-dash.