SPY RSI(3) Pullback
Backtest report · Friction-aware mean reversion still clears holdout and benchmark-relative checks.
The full research path, start to finish.
Fixed example data. Real decision surfaces — backtest evidence, comparison selection, and forward paper tracking in one pass.
01
Backtest
Cost-aware result with out-of-sample validation and benchmark-relative context. Caveats are part of the output, not a footnote.
Sharpe
1.28
after costs
OOS Sharpe
0.82
holdout
Max DD
−12.4%
stress point
Active return
+4.2%
vs benchmark
Win rate
58%
in-sample
→
Result is credible. Do not promote yet — put it through comparison first.
Next: compare
02
Comparison
All candidates evaluated on the same ticker, window, costs, benchmark, and validation lens. Selection cannot be driven by one flattering chart.
| Candidate | Validation | Sharpe | Active return | Decision |
|---|---|---|---|---|
| Connors RSI | Credible | 1.28 | +3.6% | Promote to paper |
| MACD | Mixed | 0.61 | +0.7% | Keep as reference |
| TSMOM | Fragile | −0.14 | −1.2% | Reject for now |
→
Connors RSI is the only candidate to survive comparison.
Promoted to paper
03
Paper tracking
Live days
104
since promotion
Forward return
+5.8%
live
Worst DD
−3.1%
forward
Status
Monitor
no action
Proof
Open the real artifacts.
The demo is compressed. These public surfaces show the full research depth Alphrex exposes when a result is shared.
QQQ Trend Filter Review
Compare review · Donchian survives after MACD and TSMOM are forced through the same benchmark-aware window.
Balanced Tactical Portfolio
Portfolio review · Forward sleeve mix judged against a factor-matched 60/40 reference with concentration and alert context intact.
Next
Run the same path on your own idea.
Create a free workspace, pick a strategy, and use the first result as a decision surface — not a claim.