You install NinjaTrader 8, import the strategy, open a chart, and 9:30 AM ET comes and goes without a single order. Nothing is broken. The chart is being fed delayed prices, the strategy is looking at a market that already happened, and reinstalling it forty times will not change that.
The data feed is the piece beginners skip, because the platform is free and the charts look fine. They are fine. They are just not showing you the present.
People say "I have NinjaTrader" as if it were one thing. It is three, and they can come from three different places.
A live strategy needs all three. A backtest needs only the first two, and only needs the data historically. That distinction is where most of the confusion lives.
A backtest can run on data you downloaded last night. A live automated strategy cannot. If your feed is delayed by ten minutes, the system is deciding at 9:40 what it should have decided at 9:30, on prices that no longer exist. No partial credit: real-time or nothing.
They get lumped together because they all put candles on a screen.
Delayed data is the exchange's real feed, held back. For CME index futures the usual delay is about ten minutes. It is free because it is worthless for trading, which is exactly why it is given away. Charts built from it look completely normal, and that is the trap: nothing on screen tells you the last bar is stale.
Simulated or historical data is what a demo connection hands you. Fine for learning the interface, fine for loading a chart, fine for the Strategy Analyzer. Not a live market.
Real-time data is the streaming feed: every tick as it prints, with no deliberate delay. This is what an automated strategy requires, and it is the one you pay for.
The exchange owns it. For US index futures that is CME Group, which produces the price stream for the Micro E-mini contracts described in futures and micro contracts. Distributors carry that stream to your platform: NinjaTrader's own data service, or routing technologies such as Rithmic, CQG, Continuum and Tradovate.
Separate charges can land on you:
| Charge | Who levies it | What it is for |
|---|---|---|
| Exchange market data fee | CME Group | The right to receive real-time prices, billed monthly per exchange group |
| Distributor or platform fee | Your data provider or broker | Delivering that stream to your machine, sometimes bundled or waived |
| Depth of market surcharge | Exchange or distributor | Level 2 order book on top of level 1 quotes |
| Often nothing extra | The prop firm | Many firms bundle data into the evaluation or activation fee |
Exchange fees are much lower for non-professional subscribers than for professional ones, and the classification is not yours to pick casually. Your provider makes you attest to it at sign-up, and being reclassified later comes with a bill.
On a prop-firm account, check before you subscribe to anything: firms usually include the data, and paying twice for the same CME stream is a classic first-month mistake. The wider cost picture is in what it costs to run an automated trading system.
Level 1 gives you the last trade, the best bid, the best ask and volume. That is everything a bar-based strategy needs, because a 1-minute candle is assembled from trades, not from the order book.
Level 2 adds market depth, the resting orders at each price level. It is an extra subscription, useful if you read order flow by eye, and completely unread by a bar-based system. Buying depth so a strategy can "see more" is paying for a stream nothing consumes.
One instrument, one timeframe, one window a day. The setup page lists exactly what has to be connected before anything is enabled.
A live feed sends you what happens from now on. It does not retroactively fill in the last several years. For backtesting you download history through the Historical Data Manager, and the depth of that download decides what your test is worth.
The published Algomyth engine backtest covers January 2022 to September 2026, day by day. Reproducing it means downloading the same span, on the same instrument, at the same bar size, with the same session template. Skip any of those and you get a different report and a wasted afternoon. The step by step is in how to backtest in NinjaTrader 8.
Hypothetical performance. The backtest referenced here is a simulation over historical data, not a live account. Simulated results are prepared with hindsight, carry no financial risk, and cannot fully reflect real execution, slippage or liquidity. Past performance, real or simulated, does not guarantee future results.
Not much, but all of it.
That last one matters more than it looks. The Algomyth engine takes one trade a day at 9:30 AM ET, with the stop and the target placed in the market at the same instant as the entry. If the feed drops at 9:29 and returns at 9:34, there was no trade. If it drops while a position is open, the stop and target are already resting at the exchange, which is exactly why they are put there instead of held in the software. Contract months are covered in contract rollover explained.
Feed problems rarely announce themselves. They show up as a strategy that "does nothing."
More of these, with the fixes, are in common NinjaTrader 8 errors. If your home internet is the weak link, a hosted Windows instance solves the uptime half of the problem, and the honest case for and against one is in VPS for automated trading.
You can run it in simulation on the free connection, and you should, for as long as it takes to trust it. You cannot run it live on delayed data, because a ten-minute delay means every decision is made on prices that no longer exist. Once real orders are involved you need a real-time feed, from your broker, a data provider or the prop firm that issued the account.
No. NinjaTrader 8 keeps market data and order routing as separate connections, so you can chart from one source and route orders to another. In practice most people use a single connection that carries both, because that is what brokers and prop firms issue. Splitting them is worth the trouble only if you have a specific reason, such as a data provider you already pay for.
It depends on the distributor and on whether you are classified as a non-professional or a professional subscriber, so check the current schedule with your provider rather than a number in a blog post. The exchange fee for a non-professional on one exchange group is a small monthly amount, and providers sometimes bundle or waive their own charge on top. On a funded account the firm often includes it, so ask before subscribing separately.
Not for a bar-based system. The Algomyth engines decide from bars built out of level 1 trade and quote data, so the order book is not part of the input. Depth is worth paying for if you trade off the tape yourself and want to see resting size, but adding it will not change what an automated bar strategy does.
At least 15 trading days for live operation, so the strategy has context before the window opens. For backtesting you need the whole period you intend to test, downloaded in advance at the same bar size you will trade. A test run on a partial download produces a report that looks real and means nothing.
Everything we write comes back to one idea: verify, then decide. The live terminal exists for exactly that.
Unlock Algomyth Live