What is this page? We compared all 19 of our live-monitor trading algorithms against
published academic research and proven quant strategies. Every algorithm was evaluated against peer-reviewed
studies (Jegadeesh & Titman 1993, Brock et al. 1992, Moskowitz et al. 2012, Connors 2009, and more).
This page documents what’s working, what’s not, and why — with citations.
Study date: February 10, 2026 |
Data source: 58 tracked consensus picks, 24 algorithms graded, 19 live-monitor algos analyzed |
Performance window: Genesis (Feb 10, 2026) to present
System Performance Overview Live Data
| Metric | Our System | Industry Benchmark | Verdict |
|---|
| Average Return |
+0.58% |
Good retail: +0.5–1.5%/trade |
On Track |
| Win Rate |
60% (21W / 14L) |
Balanced system: 50–60% |
Solid |
| Position Sizing |
5% fixed |
Quarter Kelly (6.25%) |
Conservative |
| Open Positions |
58 |
10–20 max for active mgmt |
Too Many |
| Holding Periods |
3–24 hours |
Days to weeks (academic) |
Too Short |
| Regime Gating |
1 of 19 algos |
All algos should gate |
Expand |
Algorithm Report Card vs. Academia Research
A Top Performers — Keep & Enhance
▼
| Algorithm | Our Return | Picks | Win Rate | Academic Benchmark |
|---|
| Sector Momentum | +5.17% | 4 | 100% | Factor rotation: 4x S&P Sharpe (Fama-French) |
| Sector Rotation | +2.88% | 6 | 66.7% | Sector rotation: 10–15% annualized |
These strategies align with decades of factor investing research. Sector momentum captures
capital flows between industries — currently Energy/Staples/Industrials winning while Tech bleeds.
Small sample size but thesis is sound.
B Solid Performers — Minor Tweaks
▼
| Algorithm | Return | Picks | Win Rate | Academic Benchmark |
|---|
| Blue Chip Growth | +1.31% | 20 | 77.8% | Value/growth factor: 2–5% annual |
| Alpha Factor Momentum | +1.25% | 11 | 75% | Jegadeesh & Titman: ~1%/month |
| Alpha Factor Safe Bets | +0.78% | 16 | 83.3% | Low-vol anomaly: 0.5–1% premium |
| Alpha Factor Composite | +0.64% | 19 | 62.5% | Multi-factor: varies by implementation |
| Cursor Genius | +0.90% | 7 | 57.1% | Mixed factor: varies |
Alpha Factor Safe Bets at 83.3% WR is exceptional. Blue Chip Growth’s 78% WR across 20 picks
is statistically meaningful. These are outperforming their academic benchmarks.
C Mediocre — Need Attention
▼
| Algorithm | Return | Picks | Win Rate | Problem |
|---|
| CDR Zero-Fee Play | +0.18% | 16 | 55.6% | Returns barely positive after fees |
| Alpha Factor Value | 0.00% | 10 | 33.3% | Value factor dead since 2020 |
| Composite Rating | -0.25% | 12 | 50% | 50% WR + negative return = fees eating gains |
| Alpha Factor Growth | -0.36% | 16 | 28.6% | Growth stocks crushed in sector rotation |
Alpha Factor Growth at 29% WR is a disaster — worse than random. Alpha Factor Value at 33% reflects
the well-documented death of the traditional value premium post-2020. Both need regime filters to pause
during unfavorable rotations.
D-F Underperformers — Fix or Remove
▼
| Algorithm | Return | Picks | Win Rate | Root Cause |
|---|
| Momentum Continuation | -0.92% | 8 | 0% | Holding period too short for momentum edge |
| Technical Momentum | -1.49% | 12 | 20% | Hourly momentum ≠ academic momentum |
| Mean Reversion Sniper | -1.82% | 2 | 0% | Wrong regime (trending, not range-bound) |
| ETF Masters | -1.00% | 3 | 0% | Small sample, timing issue |
Critical: Momentum Continuation at 0% WR across 8 picks is a clear regime mismatch.
Academic momentum (Jegadeesh & Titman 1993) works over 3–12 months, not hours.
Mean Reversion Sniper works in range-bound markets but fails in trending environments.
Live Monitor: 19 Algorithms Deep Dive Research
Holding Period Mismatch — Our #1 Issue
▼
| Algorithm | Our Hold | Academic Optimal | Mismatch Factor |
|---|
| Momentum Burst | 4h | Days to weeks | 10–50x too short |
| RSI Reversal | 6h | 1–5 days | 4–20x too short |
| MACD Crossover | 6h | 2–10 days | 8–40x too short |
| DCA Dip | 24h | Weeks to months | 7–30x too short |
| Trend Sniper | 4h | Days to weeks | 6–24x too short |
| Ichimoku Cloud | 8h | Days (designed for daily) | 3–6x too short |
| RSI(2) Scalp | 3h | 1–3 hours | Aligned |
Academic evidence overwhelmingly shows momentum persists over weeks–months, trend following works over
days–weeks, and mean reversion needs 1–5 day windows. Our hourly holds are capturing noise, not signal.
Only RSI(2) Scalp is correctly matched to its academic source.
Algorithm-by-Algorithm Academic Comparison (all 19)
▼
Detects >2% hourly candle moves. The 2% threshold is reasonable for crypto but too loose for stocks
(where 2% hourly is rare and significant). The 4h hold is far too short for momentum strategies —
academic evidence shows momentum persists over weeks to months.
Solid textbook implementation. RSI(14) at 30/70 thresholds is the industry standard.
Degrades in strong trends where RSI stays overbought/oversold for extended periods.
More selective thresholds (20/80) trade frequency for accuracy.
Volume confirmation at 1.5x is best practice. However, breakout strategies have inherently low
win rates (20–40%). Our 3:2 TP/SL ratio won’t compensate — need 4:1+ to profit at <40% WR.
A 5% drop in 24 hours almost never happens for individual stocks — this algo effectively only fires
for crypto. Vanguard research shows lump sum beats DCA 68–75% of the time, but during corrections
dip-buying outperforms. Needs stock-specific threshold of -2%.
Academic backtests show BB squeeze “does not do particularly well for most assets tested.”
Our 20th percentile bandwidth threshold is a smart quantitative approach. Works better as a
secondary confirmation rather than standalone entry signal.
Evidence
Limited standalone
“MACD has limited use as a standalone signal” — academic consensus is clear.
Standard 12-26-9 parameters are correct, but win rate is below 50% without additional filters.
Our Trend Sniper (#9) uses MACD correctly as one of 6 weighted factors.
Strongest approach in our system. Multi-indicator confluence is consistently validated
by academic research. When 2–3 independent systems agree, false signal rates drop dramatically.
This is the core thesis behind our Consolidated Picks page.
Standard Donchian breakout + ATR approach used by professional CTA (Commodity Trading Advisor) funds.
ATR ratio of 1.5x is a reasonable threshold. Hold time of 8h may be too short for CTA-style strategies
which typically hold days to weeks.
Reference
Brock 1992, Moskowitz 2012
Most sophisticated algorithm. 6-factor weighted composite (RSI 20%, MACD 25%,
EMA Stack 25%, BB %B 15%, ATR 10%, Volume 5%) with a 4+/6 agreement threshold.
The regime gate that suppresses crypto BUY signals in bear markets is the gold standard approach.
Only algorithm with academic-grade risk management built in.
Multi-lookback approach (2, 3, and 4 candle dips) tests multiple patterns rather than relying on a
single formation. Asset-specific thresholds (2% for crypto/stocks, 0.5% for forex) are appropriate.
Volume bonus for institutional confirmation is a smart addition.
Z-Score of 2.0 is the correct statistical threshold for “unusual” activity. The directional
filter (0.3% minimum candle move) prevents triggering on volume spikes with no price movement.
Clean, statistically sound implementation.
Sophisticated and rare in retail systems. The Martin Ratio (momentum / Ulcer Index)
penalizes volatile pumps and rewards smooth trends. Based on Moskowitz, Ooi & Pedersen 2012
Time Series Momentum paper — one of the most cited in momentum research.
Conditions
Triple-confirmed
TP Target
BB Middle (dynamic)
Triple confirmation (%B <0.15 + RSI <35 + MACD histogram turning up) is one of the strongest
mean-reversion setups. Dynamic TP targeting the Bollinger middle band is data-driven rather than arbitrary.
Needs a regime gate: suppress in trending markets, activate in range-bound.
ADX >25 is the industry standard for daily/weekly charts, but research suggests >20 for intraday
data where trends are naturally weaker. The DI spread requirement of >5 points adds valuable
directional confirmation.
Standard StochRSI parameters with zone-filtered crossovers. Requiring the crossover to happen
specifically in oversold/overbought zones (not just any crossover) is a best practice that filters noise.
Backtests report 52–73% win rates.
Evidence
Not viable standalone
Source
QuantifiedStrategies
Better Use
Confirmation only
Weakest algorithm by academic evidence. Extensive backtesting by QuantifiedStrategies.com
found the AO is “not a viable trading strategy” on multiple asset classes when used standalone.
The zero-line cross generates too many false signals. Should be removed as a standalone algo or
restricted to confirmation role (as it’s used in Alpha Predator #19).
The 75% historical win rate is impressive. SMA(20) trend filter is the correct enhancement per academic
literature. The 3h hold is correctly aligned with ultra-short-term RSI(2) signals. However, the strategy
has “lost most of its power since Connors published in 2008” due to widespread adoption.
vs Buy&Hold
Loses 90% of time
20yr Backtest
5.2% vs 6.9% S&P
Academic evidence is unfavorable. A 20-year backtest on US equities shows Ichimoku underperforms
buy-and-hold 90% of the time (5.2% CAGR vs 6.9% for S&P 500). Its main value is reducing
maximum drawdown by ~50%. Originally designed for daily charts; hourly adaptation compounds the issue.
Should be demoted to secondary confirmation only.
High-conviction signal requiring all 4 factors aligned (ADX >25, DI direction, RSI 40–70,
AO >0, Volume >1.2x avg). The RSI “healthy trend zone” (not overbought/oversold)
is sophisticated — identifies trends with room to run. Generates very few signals but each
has multi-factor confirmation. Weakest link is the AO component.
Sector Rotation Analysis Live Data
Current Sector Performance (58 picks)
▼
| Sector | Picks | Avg Return | W/L | Status |
|---|
| ETF-Energy | 1 | +6.29% | 1/0 | Strong |
| ETF-Staples | 1 | +4.99% | 1/0 | Strong |
| ETF-Industrial | 1 | +4.82% | 1/0 | Strong |
| Consumer | 9 | +3.03% | 9/0 | Strong |
| Energy | 3 | +0.78% | 1/0 | Neutral |
| Industrial | 6 | +0.64% | 2/0 | Neutral |
| Healthcare | 7 | +0.31% | 2/1 | Neutral |
| Finance | 6 | +0.30% | 1/1 | Neutral |
| Tech | 13 | -1.16% | 1/6 | Weak |
| ETF-Tech | 2 | -1.34% | 0/2 | Weak |
| Real Estate | 2 | -1.38% | 0/1 | Weak |
| Auto | 2 | -1.71% | 0/2 | Weak |
Sector Rotation Detected: Capital flowing from Tech/Auto/Real Estate into Energy/Staples/Industrials.
Consumer sector is a perfect 9-for-9. Our sector-aware algorithms (Sector Momentum, Sector Rotation)
correctly identified this shift and are our top-performing strategies.
Missing Strategies We Should Consider Research
Proven Strategies Not In Our System
▼
| Strategy | Typical Returns | Sharpe | Why Add It? |
|---|
| Pairs Trading |
16–27% CAGR |
1.0–3.0 |
Market-neutral, works in all regimes. We have correlated stocks (GOOGL/GOOG, JPM/BAC/WFC) |
| VWAP Mean Reversion |
0.5–1.5%/trade |
0.8–1.5 |
Institutional standard. Better than BB-based for intraday. Widely used by prop desks |
| Factor Rotation |
4x S&P Sharpe |
1.5–2.0 |
Our sector rotation is strong — formalize with Fama-French factors |
| Overnight Gap |
55–65% WR |
0.7–1.0 |
Stocks tend to gap up overnight (documented anomaly in academic literature) |
Key Takeaways Actionable
What We’re Doing Right
1
Consensus/multi-indicator approach — Academically proven to outperform any single indicator. Brock et al. 1992 validates this across decades of data.
2
Trend Sniper’s regime gate — Suppressing crypto buys in bear markets is aligned with CTA research. Only algo with academic-grade risk management.
3
VAM / Martin Ratio — Sophisticated risk-adjusted metric (momentum / Ulcer Index) rarely seen in retail systems. Based on Moskowitz 2012.
4
Self-learning system — Adaptive TP/SL/hold via walk-forward optimization. 392 parameter combinations tested per algo per asset class.
5
Corrective scoring — Penalizing losers (momentum lag, falling knife) and boosting winners provides a real-time feedback loop on pick quality.
What Needs Fixing (Priority Order)
1
Extend holding periods — Our #1 issue. Academic momentum/trend works over days–weeks, not hours. At minimum, double all hold times.
2
Remove Awesome Oscillator standalone — “Not viable as a standalone strategy” per extensive backtesting. Demote to confirmation-only role.
3
Reduce Ichimoku weight — Underperforms buy-and-hold on US equities 90% of the time in 20-year backtests. Demote to secondary confirmation.
4
Expand regime gating to all algorithms — Currently only Trend Sniper (#9) uses it. All algos should gate: suppress momentum buys in bear markets, suppress mean-reversion shorts in bull markets.
5
Fix Breakout R:R ratio — 3:2 TP/SL won’t compensate for 20–40% win rates. Need 4:1 or 5:1 risk-reward for breakout strategies.
6
Add position correlation check — 5 tech stocks losing simultaneously = correlated drawdown. Cap sector concentration at max 3 same-sector positions.
7
Lower ADX threshold to 20 for hourly data — 25 is designed for daily/weekly charts. Intraday trends are weaker.
8
DCA Dip stock-specific threshold — -5% in 24h is crypto-appropriate but almost never triggers for stocks. Use -2% for stocks.
Academic Sources & References
Key Papers & Data Sources
▼
- Jegadeesh & Titman 1993 — “Returns to Buying Winners and Selling Losers” — foundational momentum paper. 12-month lookback, ~1%/month premium.
- Brock, Lakonishok & LeBaron 1992 — “Simple Technical Trading Rules and the Stochastic Properties of Stock Returns” — validates technical analysis on DJIA.
- Moskowitz, Ooi & Pedersen 2012 — “Time Series Momentum” — TSM across 58 liquid instruments. Basis for our VAM algorithm.
- Lo, Mamaysky & Wang 2000 — “Foundations of Technical Analysis” — pattern recognition in prices. Basis for our Dip Recovery algorithm.
- Fama & French 2015 — Five-Factor Asset Pricing Model. Profitability and investment factors explain most value premium.
- Connors & Alvarez 2009 — “Short-Term Trading Strategies That Work” — RSI(2) strategy with 75% historical WR.
- Vanguard DCA Study — Lump sum beats DCA 68–75% of the time across US, UK, and Australian markets.
- NBER 2024 — Trading Volume Alpha — volume anomalies as predictive signals.
Backtest Data Sources
Disclaimer: This analysis is for educational and research purposes only. Past academic results do not
guarantee future performance. All trading involves risk. Our algorithms are paper-traded and have no audited
live-money track record. Academic benchmarks reflect historical averages and may not apply to current market
conditions. Always do your own research before making investment decisions.