Timeless strategies.
Tested, not guessed.

Software for NinjaTrader 8 that tunes documented futures strategies to your market, then tests them on data they never saw.

Start free trial How it works
Download 3 algos for free

Built by our engine for MES, MNQ and MYM. Free. Backtest them tonight.

17 strategiesdocumented for decades, never invented by the engine
A few settingstiming, entry, stop, target and one or two filters
Out of sampleevery result is checked on data it never saw
01 How it works

One strategy. A few settings. Nothing invented.

Most strategy builders mix hundreds of indicators until a curve goes up. Theory of Edge starts from a strategy with a known logic and only adapts it to your market and your session. Fewer moving parts, far less room to fool yourself.

STEP 1

Choose one strategy

Export your 1-minute history from NinjaTrader 8 with the indicator we provide. Pick one of the 17 documented strategies, an opening range breakout for example, and the session you trade: US, European or Asian hours. Not sure which one? Leave it on Automatic and the engine tests them all.

STEP 2

Let it tune, not invent

The engine adjusts the entry window, the entry trigger, the stop and the target, and adds at most one or two filters from 150+ alphas, such as relative volume at the open or distance to VWAP. It keeps only the settings that also hold on data it never saw.

STEP 3

Trade it in NinjaTrader

Each strategy is a short line of text. Copy it from the Studio, paste it into the trading algo we provide, and NinjaTrader runs it: first in the Strategy Analyzer to confirm the numbers, then on a simulation or live account.

02 The approach

Most strategy builders find noise. We start from structure.

A genetic algorithm will always find a curve that goes up. The real question is whether there was anything there to find.

See a hedge fund manager use similar methods
Data-mining buildersTheory of Edge
Starting pointRandom indicators and thresholdsA documented market behaviour, published on arXiv or SSRN
What the engine searchesAny combination that fits the pastA few settings of that one strategy: timing, entry, stop, target, one or two filters
Why it should workBecause it did, in the backtestA structural reason, observed for decades
What you usually getA strong backtest and a weak live accountFewer strategies, each with a reason to exist
Examples
Opening Range BreakoutEvery trader knows it. Few know how to apply it properly: which range, which session, which stop.
Early MomentumTrickier, especially by hand: the first minutes of the session decide, and timing is everything.
17documented strategies, none invented by the engine
3sessions, US, Europe, Asia, each its own market
±10%every setting nudged on its own after each run: a lucky peak breaks, an edge holds
Trade ideas with a reason to work.Overfitting can be reduced, never removed. Every strategy comes with its out-of-sample record and its Monte Carlo range, before you risk a dollar.
Get access
03 In NinjaTrader

What a result looks like, in your own platform.

Strategies found by the engine, replayed in the NinjaTrader 8 Strategy Analyzer on MNQ with one contract. The Studio and NinjaTrader run the same code bar by bar, so what you test in the Studio is what you see here.

NinjaTrader 8 Strategy Analyzer: cumulative net profit of a strategy on MNQ
NinjaTrader 8 Strategy Analyzer: cumulative net profit of a second strategy on MNQ
NinjaTrader 8 Strategy Analyzer: cumulative net profit of a third strategy on MNQ
NinjaTrader 8 Strategy Analyzer: cumulative net profit of a fourth strategy on MNQ
A strategy running on an MNQ 1-minute chart in NinjaTrader 8, entry and exit marked
The same strategy on another day, two trades marked
See it on your own market.Export your data once, run the engine tonight.
Get access

Backtests on historical data, hypothetical by nature. They include commission and one tick of slippage per side. They say what a strategy did, not what it will do.

04 The Studio

One window, from data to portfolio.

A Windows application. Everything below is a real capture of the current build.

Portfolio Builder: strategies with contracts, combined equity, correlation, prop-firm simulation
Monte Carlo of the whole portfolio
Training: equity curve and the strategy databank
Monte Carlo: the fan of 1,000 resampled equity paths around the original
05 Access

One license. Everything included.

Launch price, limited number of licenses
$99.99$49.9950% offper month after a 24-hour free trial, cancel anytime
  • The Studio for Windows: Training, Monte Carlo, Portfolio Builder
  • Two NinjaTrader 8 algos: one exports your 1-minute data, one trades any strategy, backtest and live
  • US, European and Asian sessions, on any futures market your NinjaTrader has data for
  • A step-by-step guide: export your data, run the engine, read the results, trade a strategy
  • Updates for as long as you subscribe
Start your free trial
24 hours free, card required: cancel before the end of the trial and you pay nothing. Then $49.99 per month, the launch price, kept for as long as you stay subscribed. Payment and taxes handled by Stripe at checkout.

Yes, if you are ready to learn the basics. The step-by-step guide takes you from your first data export to a backtest in NinjaTrader, and explains what each number means: drawdown, profit factor, out-of-sample, Monte Carlo. Experienced traders will find every detail of the validation on the Method page. Whatever your level, start on a simulation account: the software does the heavy work, the discipline stays yours.

Three things: the Studio (a Windows application), a NinjaTrader 8 export algo that writes your 1-minute data into the file the Studio reads, and a NinjaTrader 8 trading algo that runs any strategy you tuned in the Studio. Plus a step-by-step guide.

Windows 10 or 11, NinjaTrader 8 with historical 1-minute data (a free NinjaTrader license is enough to backtest). Nothing else to install: the Studio ships with everything it needs. No GPU, no external data feed. There is no Mac version: NinjaTrader 8 itself runs on Windows.

Yes: 24 hours free, with every feature of the Studio. A card is required to start; cancel before the 24 hours are over and you are not charged. Otherwise your subscription starts at $49.99 per month, and you can cancel it at any time.

Cancel at any time from your account: the next monthly charge stops and the current month stays active until its end. If the Studio cannot be activated or run on a supported setup, that payment is refunded. Access starts as soon as your license is activated, so customers in the EU and the EEA confirm immediate access, and the end of their 14-day withdrawal right, on the activation screen of the Studio; a confirmation is sent by e-mail.

We prefer a small community of traders who know what they are doing. This is not a product for mass distribution, and we keep it that way. When the cap is reached, new sign-ups go on a waiting list.

Yes, open to everyone: 30% of every payment of the customers you bring, for as long as they stay subscribed. Create your account in the partner space. Three rules: no fake testimonials, no promise of gains, nothing illegal.

The trading algo runs inside your own NinjaTrader, on whatever account you connect. Whether automated strategies are allowed on a given account is between you and your broker or prop firm: read their rules first. The Portfolio Builder can simulate a challenge with the rules you enter, so you see the pass rate before you pay a fee.

You can never be certain, and anyone who says otherwise is selling something. What the Studio does: it never invents a strategy, it only tunes a handful of settings of a documented one, which leaves far less room to fit noise; every result must be profitable on data it never saw during the search; near-duplicates are dropped; after every run, each setting is nudged on its own (entry and exit a few minutes either way, each indicator, stop and target up to 10%) to show whether the result holds or was a lucky peak, and which setting is the weakest; the Monte Carlo shows you the range of outcomes; the portfolio shows how strategies behave together. The decision stays yours, with the numbers in front of you.

06 What you end up with

One strategy is a bet. Ten uncorrelated ones are a business.

Every strategy has losing months. What keeps an account alive is several strategies that do not lose at the same time. That is what you build here, in an evening.

BUILD

A databank of strategies

Each run gives you strategies that held up on data they never saw, across different ideas and sessions.

COMBINE

A portfolio that smooths itself

You see the combined curve, its drawdown next to the sum of the parts, and whether your strategies really diversify.

KNOW

Your odds before you pay

The drawdown to expect at 95%, and how often your prop-firm challenge would have passed, started on hundreds of past dates.

Build your portfolio tonight.24 hours free, then $49.99 per month at launch. Cancel anytime. Limited number of licenses.
Get access
07 Who is behind it

Two traders, two worlds, one tool.

One of us spent years in market making and arbitrage: inventory models, tiny edges, recalibration every few days, a millisecond between profit and loss. The other built genetic strategy engines: breeding, selection, and the long fight against overfitting.

Theory of Edge is where the two met. The engineering discipline of the first, so the engine processes 1-minute data as fast as most platforms process daily bars. The search power of the second, aimed only at ideas with a structural reason to work.

We built it for our own trading and still run it on our own accounts. You get the same tool we use: the engine, the numbers, and the decisions, which stay yours.