AbacuQuant Portfolio: 11-Asset Multi-Strategy EA with 11,333 Real-Tick Trades and 6.4% Max Drawdown

Anyone who has spent a few months with automated trading has seen the same pattern: a backtest with a smooth, steep equity curve, followed by a live account that behaves nothing like it.

In most cases, the cause is one of three things:

• Loss recovery. Martingale and grid systems add size after losses, so a single extended trend can erase months of small gains.
• Curve-fitting. Parameters tuned to the noise of one historical period stop working as soon as the market behaves differently.
• Single-market dependence. A system that trades one symbol spends long stretches flat or in drawdown whenever that market goes quiet or changes character.

AbacuQuant Portfolio was built around one question:

Can an automated portfolio keep every position behind a hard stop, never add size after a loss, spread its decisions across unrelated markets, and admit only markets whose edge survives a strict statistical test?

This post explains how that question was answered, which tests every market had to pass, and how the portfolio behaves when it runs as one account.

Key figures from the Strategy Tester (real ticks, 0.5 percent risk per trade, January 2020 to July 2026):

Three Gates Every Market Has to Pass

No market enters the Expert Advisor until it clears three separate gates. The thresholds are the same for every market, and they are not adjusted to let a candidate in.

Gate 1: does the past predict the future at all?

For each market, a genetic optimization explores roughly 100,000 to 340,000 parameter sets on H1. January 2020 to May 2024 is the selection period, and May 2024 to August 2026 is a second period. Instead of picking the best pass, the research first asks whether ranking configurations by the first period predicts their ranking in the second.

Two numbers decide it: a selection ratio of at least 1.00 and a Spearman rank correlation of at least +0.10. A market that fails either one is rejected before any preset is built, whatever its best result looked like.

Gate 2: a stand-alone test on real ticks

Every surviving preset is then tested alone, with "Every tick based on real ticks", so the result carries the broker's real spreads. It has to meet three conditions:

The break-even level comes from each configuration's own stop and target, and the margin above it is tested in the usual way:

break_even = SL / (SL + TP)

margin = win_rate - break_even

t = margin / sqrt( win_rate * (1 - win_rate) / n )

Gate 3: the combined portfolio

Only presets that clear the first two gates are switched on together in the master Expert Advisor, under one account-level risk engine. A preset with too little evidence stays off. GBPUSD is the current example. It passed its real-tick test with a wide margin, but on only 18 losses, so it ships disabled until a rebuilt preset meets the 40-loss minimum.

Of the 39 markets researched, 22 were rejected along the way, including the S&P 500, the DAX, the CAC 40, crude oil, and ten individual US stocks.

The Eleven Validated Markets

The portfolio follows the trading day around the clock across three asset classes. When one market is quiet or in a losing period, others are often active, and that is what smooths the combined result:

The Combined Test

All eleven presets ran together on one account, with the default 0.5 percent risk per trade, in the Strategy Tester with real ticks:

The win rate moved by about one percentage point between the two periods, and the deepest decline stayed in the same range. That consistency is encouraging, but it should be read for what it is. The second period was also used as a consistency requirement when configurations were chosen, so it is not a blind test.

The stronger checks are elsewhere. The same research process rejected 22 of 39 markets, and the engine has been running on the developer's own real-money account since August 2026.

These are Strategy Tester results on historical data, not live trading results.

You can check the same configuration yourself: download the free AbacuQuant Portfolio demo and run it in the MetaTrader 5 Strategy Tester with your own broker and balance.

Why It Matters for Drawdown-Limited Accounts

Traders working under daily and maximum drawdown limits, such as prop-firm challenges, are hurt less by a lack of gains than by a single bad stretch. Three design choices address that directly:

• Risk per trade starts at 0.5 percent of balance. Position size is calculated from each trade's own stop distance. In the second period of the combined test, the longest losing streak was four trades.
• Each market has its own magic numbers. They are derived from the instrument's name, so configurations never collide, and the EA refuses to trade if it detects a conflict.
• The stop loss travels with the order. It is part of the opening instruction, so a VPS disconnect or a power failure does not leave a position unprotected.

Every firm's rules are different. Before using the EA on a funded account, run it in the Strategy Tester with your own balance and check the result against your firm's daily and total limits.

Frequently Asked Questions

What balance does it need to run all eleven markets?

At the default 0.5 percent, sizing works properly from about 3,000 for the three currency majors only, about 8,000 when gold, EURJPY, and the indices except the Nikkei are added, and about 20,000 for all eleven markets. Below these levels, the broker's minimum lot can push real risk above your setting. The EA prints the balance each configuration needs at startup, and cent accounts, where available, allow finer sizing on small balances.

Will it work with my broker?

It needs MetaTrader 5 and a hedging account. The EA detects broker prefixes and suffixes such as EURUSD.r, and it recognizes GOLD as gold. Indices that a broker names differently, such as US30 instead of WS30, need one entry in the SymbolMap input, for example WS30=US30.cash. A low-spread account is recommended.

Does it fit prop-firm rules?

It uses no martingale, no grid, and no averaging, and every position has a stop loss from the moment it opens. Whether it fits a specific firm's daily and total drawdown limits depends on your balance, your risk setting, and that firm's exact rules, so test it in the Strategy Tester before trading a funded account.

Trading currencies, indices, and metals on margin carries a high risk of loss. Strategy Tester results describe the past and do not guarantee future results.