Скачать или смотреть Connors RSI(2) Mean Reversion on SPY — Backtest Video (1996-2024)

This episode backtests the standard Connors RSI(2) mean-reversion system against Buy & Hold SPY under explicit assumptions, over 1996–2024.

We treat it as a rules-based way to buy sharp pullbacks inside a longer-term uptrend, then exit on a quick bounce. We then walk the equity curve through trends, chop, breaks, and snapbacks so you can evaluate it as an exposure decision.

Can a short-term RSI(2) dip-buying rule materially reshape SPY’s drawdown profile without paying too high a price in whipsaw and missed upside?

- -

WHAT THE VIDEO DELIVERS:

1) The full strategy spec. Every input, condition, and decision rule is stated explicitly.

2) A complete timing model. Signal timing and execution timing are specified so the backtest is reproducible.

3) Implementation-grade clarity. Definitions, rules, and edge cases are spelled out to prevent ambiguity.

4) A transparent benchmark. The benchmark is run under the same assumptions for a clean comparison.

5) On-screen backtest assumptions. Timeframe, costs, cash treatment, and exclusions are stated in the video.

6) A head-to-head scorecard with a fixed order. Exposure first, risk second, payoff last.

7) An equity-curve walkthrough that matches the scorecard. Narration ties the numbers to what actually happened across the backtest window.

8) Regime-aware interpretation. The episode pressure-tests strategy behavior through distinct market conditions.

9) Frictions treated as real. Turnover and implementation implications are discussed as constraints, not footnotes.

10) A structured conclusion. The takeaway is framed as a Benefit, Cost, Role portfolio decision, not a headline.

- -

TUNED VARIANT:
Get the tuned parameter pack at backtestedstrategies dot com slash RSI2

- -

LEGAL DISCLAIMER:
Not investment advice. Hypothetical backtested performance based on historical data and specific assumptions. Does not reflect taxes, advisory or platform fees, or all market frictions. Actual results may differ materially. Past performance, whether actual or simulated, is not indicative of future results.