PrimeQuantum Almanac
- Experts
-
Masahiro Tanigawa
PrimeQuantum — building explainable, event-flow trading systems. Currently running a public forward test of PrimeQuantum Veritas: no grid, no martingale, a hard stop on every position, and a Parity Certificate so you can reproduce the reference backtest yourself. Verify, don't trust. - Version: 1.1
- Activations: 10
PrimeQuantum Almanac - Seven-Engine Portfolio EA
No martingale. No grid. No averaging down. Seven engines, 31 instruments.
Instead of depending on one trading idea, PrimeQuantum Almanac spreads risk across seven mechanically different engines. It is a multi-strategy portfolio EA for MetaTrader 5, running rule-based engines across 31 FX, metals, index, energy and crypto instruments - trend following, month-end flows, index opening effects and volatility-regime hedging. The portfolio is therefore less dependent on any single market behaviour.
You choose one of three strategy presets: Low Risk, Standard or High Risk. All trading logic is sealed.
BACKTEST REFERENCE
- IC Markets real data, May 2018 - June 2026
- 1-minute OHLC model, real spread and real swap charged
- Standard preset: Profit Factor 2.45, max equity drawdown 30.3%
- All seven complete calendar years (2019-2025) positive in the historical simulation; both partial years (2018 from May, 2026 H1) also positive
- Historical simulations are not live results. Slippage is not modeled. For conservative evaluation, the developer uses a stress assumption of 0.7-0.8x backtest Sharpe and 1.3-1.5x backtest drawdown; these are not forecasts of live performance
Before buying, test it yourself. The free demo reproduces the published Strategy Tester configuration - the exact recipe is below.
Forward monitoring: a demo run of the Standard preset has been live on a VPS since August 2026; its results are posted monthly in the Comments section of this page. A real-money MQL5 signal is planned and will be linked here once it starts.
After purchase, send me a private message to receive the reconciliation guide and calendar-update notifications. I do not offer anything in exchange for reviews.
THE ANNUAL RE-BASE
Every January 1st the EA re-bases its position sizing to the same fixed stake: the account balance at first activation. Profits accumulate in the account, but position sizes do not grow with them. Three things follow:
- Position sizes do not compound with account growth, helping avoid orders expanding toward broker size limits.
- Every calendar year is measured against the same yardstick, like pages of an almanac. The yearly figures below are exactly what the EA reproduces.
- You can withdraw profits at any time without changing the EA's behavior.
The current year, the fixed stake and the year-to-date result are always shown on the chart panel (YEAR / STAKE / P/L).
THE THREE RISK LEVELS - ALL NUMBERS ARE BACKTEST FIGURES
Test conditions: MetaTrader 5, IC Markets real data, 2018.05.01 - 2026.06.30, 100,000 USD stake, leverage 1:100.
| Preset | Profit Factor | Max equity drawdown | Worst full year | Note |
|---|---|---|---|---|
| LOW RISK | 2.50 | 25.7% | +41.7% | Factory default for new users |
| STANDARD | 2.45 | 30.3% | +48.4% | Reference preset - the published monitoring runs this |
| HIGH RISK | 2.39 | 37.9% | +74.5% | Risk capital only - intra-year ledger drawdown reached -41% |
All three presets use the same engines, same entries, same exits. Only position sizing changes.
Annual backtest results, Standard preset, measured against the fixed stake:
| Year | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|---|---|
| Return vs stake | +198% | +3,173% | +186% | +48% | +113% | +403% | +305% |
Two clarifications the headline numbers need:
- 2020 dominates the sample. Excluding 2020 entirely, the pooled profit factor is still 2.06.
- The report profit factor is a whole-run figure. Year-by-year profit factors range from 1.26 to 3.05; the choppy year 2022 printed 1.38.
SEVEN ENGINES. SEVEN DIFFERENT JOBS.
| # | Engine | What it does | Share of net profit |
|---|---|---|---|
| 1 | Trend following | Gold, silver, crude, Bitcoin, Ethereum - multi-week momentum | 35.6% |
| 2 | Index overnight-overreaction fade | US indices, 30-minute window at the US open; the crisis-year engine (most of 2020) | 28.1% |
| 3 | Month-end FX fixing flow | Three pairs, calm months only - London fixing flow | 20.3% |
| 4 | FX month-end rebalancing reversal | Five majors | 7.4% |
| 5 | Market-neutral FX pair arbitrage | FX crosses, long/short pairs | 6.6% |
| 6 | VIX-backwardation crisis hedge | Long US indices when the volatility curve inverts | 3.5% |
| 7 | Pre-FOMC drift | US indices ahead of FOMC meetings - kept deliberately | -1.5% |
One engine is intentionally negative over the full sample. Removing it improved the headline profit factor slightly but collapsed the portfolio's weakest year. The frozen removal test is documented - details in the FAQ.
RISK CONTROL
- Position sizes never react to losses - they are set by the fixed yearly stake and step back to it every January
- Every position is time-boxed: each engine has a scheduled, rule-based exit window
- Engines are independent: if your broker lacks a symbol, only that engine disables itself and says so in the log; the rest keep trading
- Two engines use event calendars (FOMC, month-end classification) shipped through end-2027; the EA warns on screen 45 days before a table expires, and updates drop into MQL5\Files without reinstalling
RISK AND LIMITATIONS
- 27-28% of trading days have no trades at all; the longest quiet stretch in the backtest was 13 consecutive trading days. If silence makes you switch an EA off, this product is not suitable for you.
- The worst single backtest day was about -15 to -18% of the yearly stake, driven by the intraday index engine.
- A bootstrap over yearly resamples puts the probability of a sub-2.0 live profit factor at roughly 37%. Below 2.0 does not mean unprofitable - it means the backtest profit factor should not be treated as a live expectation. No certainty is claimed.
- Positions are carried across January 1st, so part of a strong December can land on the next year's page.
- Roughly a third of gross backtest profit went to costs, mostly financing (swap). At 1.5x historical swap the Standard profit factor still prints at or above 2.0.
- There is no fixed stop-loss on every trade; exits are scheduled and rule-based, and risk is controlled at the sizing layer. The worst-day and drawdown figures above are the honest measure of that choice.
SETUP AND REQUIREMENTS
- Attach to any one chart - symbol and timeframe do not matter; the EA trades its own 31-instrument universe
- Inputs: risk level, symbol names (pre-filled, auto-resolving prefixes/suffixes), magic number - nothing else
- Broker: one carrying the full instrument set (FX majors and crosses, gold, silver, crude, BTC, ETH, three US index CFDs). Published numbers were produced on IC Markets data and not tuned to any other broker.
- Hedging account recommended. Leverage 1:100 or higher (1:30 is not sufficient for High Risk)
- Minimum balance: 1,000 USD (verified). 10,000 USD or more is representative - below that, minimum-lot rounding thins out the smaller legs
- VPS recommended. No DLLs. One optional WebRequest (public CBOE endpoint) refreshes the VIX reading; the EA runs without it
WHO THIS EA IS FOR
- Traders leaving martingale, grid or averaging EAs - here, sizes never react to losses, so there is nothing to recover by doubling
- Buyers who care about drawdown, yearly consistency and transparent risk control rather than headline gains
- Patient traders - this is a selective system with silent stretches, not an overtrading scalper
Not for you if you need daily action, cannot tolerate double-digit drawdowns, or expect backtest performance to repeat unchanged in live trading.
ABOUT THE DEVELOPER
Developed by a systematic-trading research practice. Every published figure was replicated exactly before release, removal tests were run with gates frozen before results were seen, and the documentation - including the tests that failed - ships with the product.
FREQUENTLY ASKED QUESTIONS
Is this martingale or grid?
No. Position sizes are set by the fixed yearly stake and never react to losses. You can verify this in the strategy tester in under a minute.
Why is there a losing engine inside?
Because removing it materially weakens the portfolio's worst-year result, and we measured that. In a frozen A/B test, removing the pre-FOMC engine collapsed the worst full year from +48.4% to +12.2% on Standard and failed every risk tier's gates. Its -1.5% full-sample backtest contribution is an insurance premium: its gross profits concentrate in exactly the years when everything else is quiet.
Is there a stop loss on every trade?
Not in the classic fixed-pips sense, and we would rather tell you than let you find out. Every position is time-boxed, and risk is controlled at the portfolio sizing layer - where testing showed the main risk constraint to be - pinned to the fixed yearly stake. The measured worst day (-15 to -18% of stake) and worst drawdowns are published above.
Why 1-minute OHLC instead of real ticks?
The portfolio spans 31 symbols over 8 years; the engines enter on scheduled windows measured in minutes to weeks, not tick microstructure. Real spread and real swap are charged in every published test, and we tell you to derate the results for live trading.
What happened in 2020 - is the whole result just that one year?
2020 dominates the sample and we say so. Excluding 2020 entirely, the pooled profit factor is still 2.06, and every complete calendar year is profitable on all three presets, including 2022. The account-equity curve also looks flatter after 2020 for an accounting reason, not a performance reason: sizing stays on the fixed yearly stake, so a large account grows by a small percentage per year. The per-year figures are the honest yardstick - 2024 printed +403% and 2025 +305% of stake, versus +198% in 2019.
Can I withdraw profits without breaking it?
Yes - this is the point of the annual re-base. Sizing follows the fixed yearly stake, not the account balance, so withdrawals and deposits do not change the EA's behavior.
How often does it trade?
About 72% of trading days have at least one deal; the median active day has around 3. The longest silent stretch in the backtest was 13 trading days. Month-end weeks are the most active.
Does it work on other symbols or timeframes?
The chart it is attached to is irrelevant - the EA trades its own 31-instrument universe internally.
Do the presets really contain the whole configuration?
Yes. Before release, the plain preset selections were run against fully pinned research configurations and reproduced the official equity curves to 0.0000% on all 2,029 daily equity rows, for all three risk levels, with identical trade counts. No user-adjustable strategy parameter changes the certified trading logic.
Will there be updates?
Yes - lifetime updates, including refreshed event-calendar files (currently shipped through end-2027; the EA warns 45 days before a table expires).
PRICING
The launch price reflects a short public track record. It rises on published evidence milestones, not on a marketing countdown, and it will not be discounted.
| Stage | Evidence milestone | Price |
|---|---|---|
| Launch | - | 399 USD |
| Stage 2 | 3 published forward months | 499 USD |
| Stage 3 | 6 months, drawdown inside the published range | 599 USD |
WHAT'S NEW - VERSION 1.00
- Initial release: seven engines, 31 instruments, annual re-base
- Low / Standard / High Risk presets, all other inputs sealed
- Live panel with per-engine state and YEAR / STAKE / P&L row
- Event calendars through end-2027 with 45-day expiry warnings
Risk Disclaimer: Backtest results are historical simulations and do not guarantee future performance. Trading involves risk, and losses are possible. Only trade with capital you can afford to lose.
