TITAN-MID
A leveraged Nasdaq-100 mean-reversion strategy with multi-indicator vote-weighted entries. Backtested from 1994 (32 years), Sharpe 1.33, ROR 49% with MaxDD -47%. This page is rendered from real backtest data — every number, equity point, and stress-test return is computed from the actual .rts output.
Results are hypothetical and based on a historical backtest. Past performance does not guarantee future results. See risk disclosure for full details.
The backtest, in full.
Year-by-year, month-by-month.
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Total | MaxDD |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1994 | -20.5 | 5.8 | -2.8 | 10.2 | 2.5 | 6.5 | 12.9 | 18.4 | 23.6 | 4.1 | 69.4% | -26.7% | ||
| 1995 | -6.9 | 4.1 | 0.8 | 13.5 | 0.9 | 17.1 | 14.8 | 9.6 | 8.1 | 18.8 | 2.1 | 9.2 | 136.1% | -15.0% |
| 1996 | 3.9 | -4.6 | 12.8 | 5.9 | 20.6 | 7.7 | 0.7 | 0.5 | 9.5 | -12.5 | 3.8 | 21.4 | 88.0% | -24.7% |
| 1997 | 4.2 | 5.7 | 0.6 | 9.3 | 0.0 | 7.2 | 7.4 | 3.9 | 6.4 | -39.2 | 12.6 | 0.8 | 9.2% | -46.7% |
| 1998 | 3.3 | 0.0 | 15.6 | 23.0 | -1.6 | 10.2 | -2.5 | -13.1 | 16.6 | -5.1 | 2.3 | 19.9 | 83.3% | -16.4% |
| 1999 | 12.9 | -6.9 | 18.8 | 1.6 | -1.6 | 15.4 | 1.4 | 5.3 | 3.4 | 8.0 | -0.7 | 25.1 | 113.5% | -14.8% |
| 2000 | 5.7 | 17.9 | 1.4 | -3.5 | 7.1 | 2.9 | -18.0 | 7.9 | 19.2 | 6.5 | -10.1 | 26.0 | 70.9% | -29.6% |
| 2001 | 2.2 | -14.5 | 7.4 | 1.2 | -1.7 | 13.3 | 9.4 | -4.8 | -9.9 | 6.2 | 9.9 | -3.6 | 11.6% | -21.5% |
| 2002 | 4.1 | 3.7 | 10.1 | 2.1 | 1.3 | -0.7 | -3.6 | 1.4 | -1.6 | 4.5 | 5.3 | -2.7 | 26.7% | -13.7% |
| 2003 | 8.3 | -0.7 | 6.5 | 5.0 | 5.2 | 4.4 | 7.5 | -1.0 | -13.3 | 25.6 | 9.0 | 0.0 | 58.3% | -19.0% |
| 2004 | -3.2 | 0.0 | 3.8 | -6.3 | 14.1 | 2.1 | -4.2 | 1.3 | 1.0 | 10.4 | 7.0 | 17.1 | 48.6% | -15.2% |
| 2005 | -2.2 | -0.3 | 10.2 | 6.1 | 5.4 | 3.1 | 1.2 | 1.8 | 6.0 | 9.3 | 6.2 | 8.2 | 69.7% | -13.1% |
| 2006 | 23.5 | -11.1 | 14.8 | -3.3 | -21.1 | 9.4 | -0.2 | 4.8 | 8.6 | 9.9 | 10.7 | 3.1 | 36.8% | -26.4% |
| 2007 | -2.9 | -1.5 | -0.6 | 2.1 | 6.0 | 16.1 | 10.4 | 12.5 | 2.5 | -11.6 | 2.8 | -2.0 | 34.2% | -19.0% |
| 2008 | -16.7 | 1.5 | 2.4 | 7.1 | 15.8 | -1.4 | -2.6 | 1.9 | -0.5 | -5.5 | 1.3 | 1.5 | 3.1% | -22.9% |
| 2009 | -0.5 | 3.1 | 3.9 | 6.5 | 7.2 | -0.6 | -0.6 | 9.5 | 14.2 | 9.2 | 10.7 | 8.2 | 96.7% | -7.5% |
| 2010 | -17.4 | 7.3 | 0.0 | 5.2 | 2.3 | -5.5 | 5.5 | 2.8 | 5.5 | 0.0 | 0.1 | 2.8 | 4.6% | -20.8% |
| 2011 | 2.4 | -0.7 | 2.6 | 6.1 | 18.9 | -7.8 | 3.1 | -13.1 | 9.6 | 2.0 | -0.1 | 1.6 | 23.2% | -28.0% |
| 2012 | -0.5 | 1.7 | 2.7 | 1.0 | -6.5 | 3.4 | 8.0 | 2.0 | -3.7 | 8.4 | 5.7 | 6.5 | 30.7% | -15.5% |
| 2013 | 6.4 | 4.3 | 6.7 | -1.2 | -1.9 | 18.5 | 1.5 | 10.9 | 3.6 | 9.1 | 12.5 | 19.0 | 118.2% | -12.6% |
| 2014 | 35.6 | 3.9 | -3.2 | 7.1 | 3.6 | 3.1 | 4.7 | 4.6 | 3.6 | -8.0 | 0.0 | 2.5 | 66.9% | -31.1% |
| 2015 | -3.3 | -0.8 | 4.8 | -1.1 | 3.6 | -3.4 | 9.7 | -9.1 | -1.9 | 5.1 | 10.2 | 4.5 | 18.3% | -24.9% |
| 2016 | -10.2 | 4.4 | 6.3 | 1.4 | -6.3 | 4.7 | 0.0 | 12.3 | 6.9 | -0.5 | -2.7 | 8.1 | 24.9% | -14.0% |
| 2017 | 4.2 | 2.7 | 0.6 | 4.1 | 6.2 | -5.6 | -4.0 | 2.4 | 7.0 | 1.8 | 1.9 | 0.0 | 20.0% | -18.9% |
| 2018 | -0.4 | -9.2 | -5.0 | -0.8 | 1.3 | -3.9 | 6.2 | 3.0 | 5.2 | -7.9 | 1.5 | -3.3 | -13.7% | -29.8% |
| 2019 | 0.8 | 4.4 | 5.4 | 0.0 | 4.0 | 6.5 | 2.9 | -5.5 | -7.1 | 7.9 | 1.2 | 7.5 | 30.2% | -18.0% |
| 2020 | -5.3 | 9.9 | 15.6 | 2.9 | 8.6 | 6.0 | 6.8 | 4.3 | 11.1 | -9.5 | 13.2 | 13.1 | 104.5% | -12.0% |
| 2021 | -9.1 | -4.9 | 6.4 | 10.7 | 2.7 | 3.1 | 12.5 | 11.1 | 10.7 | 0.6 | 12.5 | 10.0 | 85.8% | -29.7% |
| 2022 | -7.2 | 1.3 | 2.0 | -1.5 | -6.3 | -4.1 | 7.6 | -0.2 | -5.2 | 1.9 | 8.6 | -3.1 | -7.2% | -18.9% |
| 2023 | 6.1 | 0.3 | 4.0 | 5.5 | 3.4 | 4.2 | 12.5 | -9.9 | -3.7 | -6.2 | 3.0 | 11.7 | 31.6% | -29.4% |
| 2024 | 8.3 | 22.7 | 10.5 | -2.8 | 9.2 | 7.2 | 2.2 | 3.4 | 2.0 | 4.0 | 18.1 | 9.5 | 141.6% | -15.3% |
| 2025 | 19.1 | -11.1 | -10.1 | 0.8 | 8.4 | 4.2 | 0.0 | 11.5 | 8.5 | 9.8 | 12.6 | 7.2 | 76.3% | -23.6% |
| 2026 | 8.7 | 7.0 | -9.6 | 2.4 | 7.2 | 20.0% | -18.8% | |||||||
| Mean Yr | 2.3 | 1.2 | 3.9 | 3.5 | 3.4 | 4.6 | 3.2 | 2.5 | 4.2 | 2.2 | 6.1 | 7.3 | 52.5% | -21.0% |
Year rows: Total = compound of the row's monthly returns. Mean Yr row: per-month cells are column means; Total is the arithmetic mean of annual totals (not compound-of-cells, so the two won't agree — this is by design).
| Metric | Strategy | SPY |
|---|---|---|
| ROR (CAGR) | 49.30% | 9.4% |
| Sortino Ratio | 1.57 | — |
| MAR Ratio | 1.06 | — |
| Net Profit | $38.90B USD | $1.72M USD |
| Trades | 2,506 | 1 |
| Win Rate | 72.2% | — |
| Expectancy | 1.48% | — |
| Avg Exposure | 57% | 100% |
| Best Year | 141.6% (2024) | — |
| Worst Year | -13.7% (2018) | — |
Multi-indicator vote-weighted mean reversion
TITAN-MID enters Nasdaq-100 stocks when multiple short-RSI windows confirm an oversold extreme simultaneously. The vote-weighted entry filters out single-indicator false signals and concentrates capital on setups where multiple lookback periods agree.
Leveraged sizing for higher capital efficiency
Position sizing uses a 1.34x leverage multiplier, targeting ~57% average exposure across the lifetime backtest. This requires a portfolio-margin (PM) account; peak gross exposure has touched ~268% during clustered entry windows.
32 years of validation
Backtested from March 1994 through May 2026 against point-in-time Nasdaq-100 membership using Norgate Premium survivorship-bias-free data. The TITAN family of strategies has been the operator's primary production system. Each leverage variant was validated independently for OOS robustness and joint correlation with the rest of the HoF cohort.
| Crisis | Strategy | SPY |
|---|---|---|
2008 Financial Crisis 2007-10-09 → 2009-03-09 | -7.2% | -47.3% |
COVID-19 Crash 2020-02-19 → 2020-03-23 | +4.9% | -28.1% |
2022 Bear Market 2022-01-03 → 2022-10-12 | -13.3% | -20.9% |
2025 Tariffs Crash 2025-02-19 → 2025-04-08 | -20.8% | -16.1% |
- titan-mid.rtsRealTest strategy file (plain-text source)
- README.txtStrategy summary, install steps, Norgate data requirements, support contact
- LICENSE.txtSingle-user license, watermarked at delivery with your purchase fingerprint
Full plaintext .rts source delivered after purchase.
- Style
- Mean Reversion
- Universe
- Nasdaq-100 Current & Past Members
- Timeframe
- Daily
- Side
- Long
- Entry
- Limit Order
- Exit
- Limit & Market
- Commission
- $0.005 USD / Share (IBKR Pro tier)
- Slippage
- $0.01 USD / Share
- Limit Buffer
- 0.05% past limit
- Risk-Free Rate
- 4.0%
SPY Buy & Hold benchmark does not include transaction fees, slippage, or management costs. The strategy figures include both.
Backtests use daily-resolution OHLC data (Norgate does not provide intraday). Stops and limits are checked against the day's High/Low; same-day stop-and-target collisions are resolved pessimistically. See Risk Disclosure § 11 for full detail.
- Data source
- Norgate Premium Data — point-in-time Nasdaq-100 membership
- Backtest period
- 1994-03-01 to 2026-05-20
- Initial capital
- $100,000.00 USD
- Position sizing
- LevMult=1.34 — leveraged sizing requiring portfolio margin account
- Compounding
- Compounded
- Survivorship bias
- None — Norgate's point-in-time membership and delisted symbols
- Look-ahead bias
- None — signals computed at prior-day close
- Benchmark
- S&P 500 ETF (SPY) buy & hold
- Delivery
- RealTest v3.x .rts file + research include
| Item | Cost |
|---|---|
TITAN-MID strategy What you're buying here | $999.00 USD one-time |
RealTest license Required. realtest.net | $249.00 USD/yr or $599.00 USD lifetime |
Norgate Premium Required for point-in-time data | $36.00 USD/mo (Platinum) |
Broker commissions IBKR Pro Fixed | ~$0.005 USD/share |
Portfolio Margin account Required for the leveraged sizing | $125,000.00 USD IBKR minimum |
OrderClerk (optional) Only if you want automated execution | ~$50.00 USD/mo |
- 01Download the .zipClick your download link in the receipt email
- 02Open it in RealTestFile → Open → select titan-mid.rts
- 03Import symbolsClick Import (one-time, ~30 seconds with Norgate)
- 04Run the backtestClick Test — your numbers will match the published numbers
Portfolio margin account required
TITAN-MID uses 1.34x leverage. Reg-T accounts cap at 2x intraday / 1x overnight; you need a Portfolio Margin account (typically $125k USD minimum at IBKR) to run this strategy at the published sizing. You can run it at lower leverage in a Reg-T account by editing the LevMult parameter.
RealTest is Windows-only
Mac or Linux users will need a Windows VM or dual-boot. RealTest does not run natively on macOS or Linux.
Norgate Premium is required
Survivorship-bias-free Nasdaq-100 history requires Norgate Premium (~$36 USD/mo). Free or basic data sources will produce inflated backtest results because delisted symbols are missing.
Drawdowns happen, and -47% is real
Maximum historical drawdown of -46.7% occurred during the dot-com unwind. That's nearly half the account. Sized correctly, the strategy recovers; sized too aggressively, you'll get margin-called out of recovery. The published equity curve assumes you stay in the trade through full drawdown — be honest with yourself about whether you would.
AvgExposure 57% means cash drag is intentional
The strategy is roughly half-invested on average. You'll see cash sitting idle during regimes where the entry filters reject most setups. That's by design — over-investing during low-quality regimes is exactly what the multi-indicator vote was built to prevent.
Inspectable proof, not promises.
Every claim in the Validation panel above is backed by a data artifact you can scrutinize. We publish the data behind the badges so a skeptical buyer can verify the strategy is genuinely robust, not a curve-fit that happens to clear our gates.
More than 100 historical trades per parameter. Even very generous tuning room can't manufacture statistical significance out of a sample this size — the result is structural.
Industry rule of thumb: a strategy needs at least 30 trades per parameter to be considered reliably backtested. Below 10 is widely considered unreliable.
We split the 32-year history at 2019-12-08 and replay the strategy on each side independently. If the curve to the right of the line slopes like the curve to the left, the strategy is generalizing — not memorizing the past.
Out-of-sample Sharpe retained at least 70% of the in-sample value. The strategy did not depend on the specific dynamics of the training period.
Cutoff date inferred as the 80/20 chronological split. This is a regime-shift stress test, not a strict held-out validation — the development data included both sides of the line. A strategy that works equally well on both halves is still meaningful evidence of robustness.
Each cell is a separate backtest. We varied RsiEntry and LevMult across the grid and re-ran the full history. A strategy that depends on a lucky parameter choice would show one bright cell in a sea of dark ones — this one doesn't.
- 100% on each axis is the shipped (production) value of that parameter — the starred cell at the center.
- 75% / 125% means we re-ran the strategy with that parameter scaled down or up by a quarter; other ticks are intermediate steps.
- Cell color follows the selected metric scale (legend below the grid). Greener is better, redder is worse — irrespective of where the star sits.
★ marks the production parameter combination (ringed in accent green). Colors map to an absolute MAR scale — a cell can be greener than the star and still represent a fundamentally similar result.
Most cells in the grid are within 25% of the production value. The parameters sit on a broad performance plateau — not on a needle peak that would only work for one specific combination.
We took the strategy's actual daily returns and re-arranged them in 500 random orders, rebuilding the equity curve each time. The end-of-history value is identical for every order (compounding is commutative) — but the path isn't. The shaded band shows the range of equity curves you might have seen with the same trades in a different sequence.
The production trades landed in a slightly favorable order. Plan for somewhat deeper drawdowns than the backtest curve shows when trading live.
A strategy whose entire edge comes from a handful of jackpot winners is a different beast from one whose edge is broadly distributed. We removed the top and bottom of the trade list by P&L and recomputed the headline stats. If most of the edge survives, the strategy is durable.
| Sample | Trades | CAGR* | Win % | Profit Factor | Retained P&L |
|---|---|---|---|---|---|
| All trades | 2,502 | 48.8% | 72.2% | 2.28 | — |
| Trim plus/minus 5% (extremes) | 2,252 | 44.5% | 74.6% | 9.76 | 39% |
| Trim plus/minus 1% (worst outliers) | 2,452 | 48.2% | 72.6% | 4.05 | 87% |
* CAGR is computed against a fixed starting balance and the strategy's actual trade duration. The figure is intended for apples-to-apples comparison across the rows of this table, not as a re-statement of the production headline.
Removing the best and worst 5% of trades retains at least 70% of the CAGR. The edge is distributed across many trades — not the result of a handful of jackpot winners.
Common questions.
Is this real backtest data?+
Wait, isn't this the operator's own production strategy?+
How was this validated?+
Why the 1.34x leverage?+
TITAN-MID -- $999.00 USD
One-time purchase. Instant download. Add 4+ strategies to your cart for a volume discount at checkout.
By purchasing you accept the Terms of Service, Privacy Policy, and Risk Disclosure.