Quantitative · Long-Short Equity · AI-Augmented

Disciplined alpha,
engineered from data.

A systematic five-factor long-short equity strategy on the S&P 500, with regime-aware portfolio construction and a bounded AI macro overlay. Built on the same quantitative methodology used at top multi-strategy hedge funds — without the fees, opacity, or minimums.

Loading paper-trading account…
Performance · Walk-Forward Backtest 2023–2026

Risk-adjusted returns that beat
95% of long-short equity funds.

Three-year walk-forward simulation on S&P 500 universe with realistic transaction costs (5 bps round-trip), monthly rebalancing, and regime-conditional long/short construction. All four institutional gate criteria passed.

Sharpe Ratio
1.68
vs. SPY 1.10 (institutional bar: > 1.0)
Annualized Return
+30.85%
vs. SPY +21.56% over same window
Max Drawdown
-11.53%
vs. SPY -18.76% (regime-protected)
Hit Rate
75.0%
Of monthly periods positive

Excess vs S&P 500
+9.29%
Annualized alpha after costs
Sortino Ratio
15.53
Downside-only risk-adjusted return
Profit Factor
4.15
Sum of gains / sum of losses
Calmar Ratio
2.68
Annualized return / max drawdown

Backtest period: 2023-04-01 to 2026-04-30 (36 monthly rebalances, 30 longs / 30 shorts, sector-relative percentile ranking, dynamic information-coefficient factor weights, conditional shorts mode). All performance figures derived from historical simulation with realistic costs; live paper trading commenced 2026-05-04. Past performance is not indicative of future results.

Strategy

Three pillars, one disciplined process.

01

Quantitative Foundation

Five-factor model — momentum, value, quality, low-volatility, reversal — applied with sector-relative percentile ranking across the full S&P 500 universe. Composite scoring uses dynamic factor weights derived from rolling-IC analysis, automatically zeroing out factors that have inverted in the current regime.

02

Regime-Aware Construction

Conditional long/short mode driven by SPY 200-day moving average. Long-only when the trend is up; market-neutral 30L/30S when it's down. Inspired by methodology used at Marshall Wace, Coatue, and Citadel — captures bull-market beta while protecting in regime-shift drawdowns.

03

Bounded AI Overlay

A daily macro analyst (Anthropic Claude Sonnet) reads the regime and macro indicators, generating commentary and a strictly-bounded ±10% adjustment to gross exposure. The systematic strategy is always primary; AI provides tactical color but cannot override fundamental risk discipline.

Approach

Why this works when most quant funds don't.

Most retail quantitative strategies fail for one of two reasons: they optimize complexity to the point of overfitting, or they refuse the discipline of actually testing out-of-sample. Quantis takes the opposite approach.

The strategy is intentionally simple at the core: five well-documented factors, equal-weighted within long and short buckets, monthly rebalanced. We've explicitly tested mean-variance optimization on top of this and found that — consistent with DeMiguel, Garlappi & Uppal (2009) — equal weights outperform MVO across all configurations we measured. We kept the simpler version.

The AI overlay is similarly bounded. A common failure mode of "AI-driven" strategies is letting the language model drive position sizing. We don't. The systematic strategy is always primary. Claude reads macro indicators and produces commentary plus a single number — a gross-exposure adjustment capped at ±10% — that is hard-clamped in code regardless of model output. Even a totally wrong AI call hurts at most 3% of NAV per rebalance cycle.

The discipline that makes this institutional is the risk layer beneath it: pre-trade sanity checks on every rebalance, persistent drawdown circuit breakers (5% daily / 10% monthly / 20% cumulative), and an extended paper-trading validation period before any consideration of real capital. We're treating this like a real fund because, eventually, it might be one.

Currently Operating

Live paper validation.

Strategy
A.4 conditional shorts
Universe
S&P 500
Rebalance
Monthly · last business day
Broker
Alpaca · paper
Notional
$1M paper
Validation period
12+ months