AlphaGlyph trades live market prices with simulated capital — driven by classical quantitative strategies and a multi-modal ML transformer — then validates every result with the same statistical tests real quant funds use to separate genuine skill from luck.
Quant strategies · ML transformer · Statistical validation · Live paper trading
Every indicator and statistical method is implemented from first principles — no black-box libraries.
ADX, Bollinger Band Width, and realized volatility classify the market into four regimes and automatically select the optimal strategy.
A trailing stop locks in gains as price climbs; the Kelly Criterion sizes each position by your historical edge, not a fixed percentage.
Bootstrap-resamples the return sequence 1,000 times and ranks the real result against the distribution of random paths.
Corrects the Sharpe ratio for multiple-testing bias and fat tails (Lopez de Prado, 2014) — tells you if the result is real or just luck.
Regresses returns against market, size, and value factors to separate genuine alpha from passive factor exposure any ETF could replicate.
Computes the efficient frontier via quadratic programming and returns the max-Sharpe and minimum-variance portfolios with a correlation heatmap.
Any strategy can look good on historical data by chance. AlphaGlyph runs three independent statistical tests and synthesizes them into a single verdict.
1,000 bootstrap paths build a distribution — your result either stands out or it doesn't.
The Deflated Sharpe Ratio penalizes testing many strategies, the way good science corrects for multiple comparisons.
Factor decomposition strips out market, size, and value exposure to reveal whether any genuine alpha remains.
# Probability the Sharpe is real after # testing N strategies on noisy data def deflated_sharpe_ratio(returns, n_strategies): sr = annualized_sharpe(returns) skew = scipy.stats.skew(returns) kurt = scipy.stats.kurtosis(returns) # expected max Sharpe from pure luck sr_star = expected_max_sharpe(n_strategies) return psr(sr, sr_star, skew, kurt) # → 0.964 ✓ significant
Backtest a strategy across any date range and watch it paper-trade in seconds, check the statistical verdict, then see the bot running live — no account required.