Free backtesting software: what really exists and where free stops

Free backtesting software exists, and there is more than one: MetaTrader's Strategy Tester, cTrader's cBot tester, NinjaTrader in simulation, thinkorswim's OnDemand and thinkBack, TradingView's Bar Replay, QuantConnect's free plan and open-source Python libraries such as Backtrader. What they give away is the right to backtest; what they limit is the data: how fine it is, how far back it goes and how realistically fills are simulated. This guide reviews each tool from its own documentation and says where free stops.

• Free covers the test itself: an engine, a bar-by-bar replay, years of daily data.
• It stops at the data: real ticks, intraday depth and option histories are almost always paid, or reserved for the higher plans.
• There is no best tool, only the one that fits your strategy: manual or coded, forex, futures, stocks or options.
• A free test that is too coarse produces results that are too good: the simulation mode matters as much as the price.

This guide recommends no tool, no broker and no paid plan: every fact comes from the tool's documentation or pricing page, read at the time of writing and cited with a link, and that page is what counts on the day you choose. The method itself is in how to backtest a trading strategy honestly.

What 'free' means for backtesting software

Trading backtesting software does three things: it supplies past data, it applies your rules to that data, by hand or through a program, and it simulates the execution of your orders. Free almost always covers the second one; the limits show up on the first and the third, because fine data is expensive and realistic fills need more than four prices per bar. Hence four criteria for comparing backtesting tools, all of which can be read in the documentation:

• The data: how fine it is (real ticks, one-minute bars, open price only), how deep it goes (how many years, on which timeframe) and where it comes from (the broker's server, the site, your own file).
• Simulated costs: are spread, commission and slippage applied, and with which default values? What a round trip costs is explained in the lesson on trading basics.
• Markets covered: forex and CFDs, futures, stocks, crypto, options. A broker's tester only knows that broker's instruments.
• Manual or automated: a bar-by-bar replay for a discretionary strategy, an engine that reads coded rules for a mechanical one. The difference is detailed in manual vs automated backtesting, and the vocabulary in the glossary under backtesting and replay.

The testers built into broker platforms

The first family is the platforms your broker gives you: the tester comes with them and the platform is a free download, so you pay nothing beyond your account, which can be a demo account. A trading backtester that ships with the platform is the most common case. Backtesting free of charge is then simply part of having the platform.

MetaTrader 4 and 5: the Strategy Tester

The MetaTrader 5 website offers the platform as a free download, with the Strategy Tester built in. According to the MetaTrader 5 Strategy Tester documentation, it tests Expert Advisors, that is, coded strategies, on several instruments at once, offers a Forward option that splits the period in two so that parameters are checked outside the zone where they were fitted, and downloads missing history from the broker's server: MetaTrader backtesting therefore depends on what your broker serves.

The simulation mode matters as much as the price. The same documentation describes four modes: 'Every tick', the most accurate but the slowest, where all ticks are simulated; 'Every tick based on real ticks', which uses the ticks accumulated by the broker, with no simulation; '1 minute OHLC', where only four prices per one-minute bar are emulated; and 'Open prices only'. The rest of the platform is described in the guide What is MT5?.

MetaTrader 4 is also a free download, according to the MetaTrader 4 website, and MT4 backtesting has three modelling methods, described in the MetaTrader 4 Strategy Tester help: 'Every tick', the most accurate; 'Control points', twelve interpolated points per bar, whose results the page says should be considered estimates; and 'Open prices only', the fastest. The same page advises against the 'Every tick' mode without the lower timeframes that fully cover the period, otherwise the results will not be accurate.

cTrader: backtesting cBots

cTrader, another platform offered by forex and CFD brokers, tests robots called cBots. According to cTrader's help page on backtesting a cBot, the data source is chosen among tick data from the server, which the page calls the most accurate historical data source, one-minute bars from the server, one-minute bars from a CSV file, or one-hour bars. The spread can be fixed at a single value or drawn at random between two bounds, and the commission is set per million units traded.

NinjaTrader: Strategy Analyzer and simulation

For futures, NinjaTrader's pricing page states that the platform is always free to use for advanced charting, backtesting and trade simulation, with commissions applying only to live trading. According to the NinjaTrader 8 Strategy Analyzer guide, the tool runs historical analysis on strategies coded in NinjaScript: backtest, optimization, walk-forward optimization, Monte Carlo simulation and basket testing across instruments.

thinkorswim: OnDemand and thinkBack

According to thinkorswim's FAQ on analysis, OnDemand replays historical market data from a trading day you select, and thinkBack, designed for option back-testing, stores nearly a decade of historical option trade data and lets you enter hypothetical trades in it. The same page notes that the results are hypothetical and that there is no guarantee the same strategy implemented today would produce similar results. For what an options test really requires, read option backtesting.

Online backtesting sites: TradingView and QuantConnect

The second family installs nothing: a browser is enough, and free takes the form of an entry plan. A free backtesting website gives you its engine; its data depth is what the plan sets. A free backtester in the browser is enough to discard an idea on a year of daily data; it becomes limiting when you want ticks or years of intraday bars.

TradingView: Bar Replay and strategy tester

According to TradingView's pricing page, the Basic plan costs €0. Bar Replay steps the chart forward bar by bar for a manual test, and the strategy tester runs a strategy written in Pine Script over the chart's history. Replay depth depends on the plan: TradingView's help page on Bar Replay data describes intraday depth from the Essential plan upwards, six months of one-minute bars; one year on Plus; all time-based data from Premium upwards; and tick data, seven days back, on Ultimate only. For free backtesting online, this is the natural first stop, and those plan limits deserve a close read.

On the coded side, the help page on Deep Backtesting, the mode that calculates a strategy on all available history, reserves it for Premium and higher plans, with a limit of two million bars and one million trades. And whatever the plan, simulated fills rest on assumptions: according to the Pine Script documentation on strategies, the broker emulator assumes that price moved through each bar in the order open, high, low, close when the open is closer to the high, in the reverse order otherwise, and that no gaps exist inside a bar.

QuantConnect: a free plan for code

For those who code, QuantConnect's pricing page describes a free plan with unlimited backtesting on one backtesting node, equity, index, forex, crypto, futures and options data at minute, hour and daily resolution, one research node and paper trading; tick and second resolutions, local coding and live trading sit in the paid plans. The engine underneath, LEAN, is published under the Apache 2.0 licence on the LEAN GitHub repository, in C# and Python.

Open-source libraries and the spreadsheet

The third family costs nothing and limits nothing, because it supplies no data: you bring your own. The Backtrader GitHub repository presents a Python backtesting library for trading strategies under the GPL-3.0 licence, with data feeds from CSV files or pandas, indicators, analyzers and commission schemes. The price is elsewhere: learning Python, finding clean data and checking your own code.

The spreadsheet, finally, is the oldest free testing tool there is, and it remains the right one for a manual test: one row per trade, the risk, the result in R, and columns of statistics. It replays nothing and simulates nothing, but it forces you to write every trade down, which no engine does for you. The Excel trading journal template serves that purpose too.

The comparison table

The backtesting platforms below are grouped by family, not ranked by merit: the 'where it stops' column sums up what each tool's documentation says about its free version.

Where free stops: ticks, intraday, options, fills

Tick data. This is the sharpest limit: in MetaTrader 5, only the ticks accumulated by the broker avoid simulation; on TradingView, replay ticks are reserved for the most expensive plan; at QuantConnect, tick and second resolutions sit outside the free plan. For a strategy that lives at tick scale, such as scalping, a free backtest runs on simulated ticks or one-minute bars, and its result is an estimate.

Intraday depth. Intraday, depth is counted in months, six on TradingView's first paid plan, or depends on what the broker's server keeps on MetaTrader and cTrader. Six months of intraday data is a single market regime.

Option histories. An option is not tested on the price of its underlying: its price also depends on time and implied volatility. Historical chains, strike by strike, are rare in free access: thinkBack keeps nearly a decade of them according to thinkorswim, QuantConnect's free plan includes options data at minute resolution and above, and the rest is sold by data vendors.

Simulated fills. The last limit is not about price: it is how a tester decides that an order was filled. With four prices per bar, it has to assume the path of price between them, and each tool documents its own, from the Pine Script emulator to MetaTrader 4's control points. A stop and a target touched within the same bar are settled on that assumption, which is why a backtest, free or not, is confirmed in a forward test before a cent is at risk.

Which tool for which profile

There is no single best backtesting software, only the one that fits your strategy and what you know how to do.

• You trade by hand, from a read of the chart: a replay tool, Bar Replay or OnDemand, and a spreadsheet to record each trade. The best backtesting platform for you tests your reading, not a piece of code.
• You have mechanical rules on forex or CFDs: the MetaTrader Strategy Tester or cTrader's cBot tester, on the demo account of the broker whose data you will actually use, preferring real ticks when it provides them.
• You trade futures: NinjaTrader in simulation, Strategy Analyzer for coded rules, market replay for manual work.
• You can code in Python: QuantConnect online to start with supplied data, Backtrader or LEAN locally with your own files.
• You trade options: thinkBack if you have access to the platform, otherwise QuantConnect's free plan, and in every case the option backtesting guide before concluding.

In every case, start with what is free, measure what you are missing, and pay only for that precise gap: ticks, years of intraday data, option chains.

Where Tradoshi fits

Tradoshi has its own backtesting module, in beta: it replays history candle by candle without showing what comes next, simulates your orders with spread, slippage and commissions, and scores each session with the same statistics as your journal. It is included from the Pro plan, so it is not a free tool and this guide does not count it among them. For everyone, Tradoshi takes over after the test: your real trades produce the statistics your backtest promised. The module is presented on the backtesting page.

Frequently asked questions