# Free crypto backtesting tools compared

> TradingView, QuantConnect, Freqtrade, backtrader, 3Commas, Gainium and Janorium compared: code or no code, history, fees, funding, liquidation, price.

Guide · Janorium · Published 2026-11-14  
Web page: https://janorium.com/guides/crypto-backtesting-tools-compared

Seven tools people use to backtest crypto strategies — TradingView, QuantConnect, Freqtrade, backtrader, the backtesters inside the bot platforms 3Commas and Gainium, and Janorium — compared on the same criteria: code or no code, markets, history, which costs are modeled, robustness checks, price and exports. Janorium is ours, so its minuses are listed as plainly as everyone else’s. Facts about the other tools come from their own documentation and pricing pages, as of 14 November 2026.

## What was compared

Every tool was judged on the same questions, using only what its own site says. Where a tool’s public pages do not answer a question, the table says **not stated**: the information is not published, which does not mean the feature is missing.

- **Code or no code** — do you write a program, or fill in rules?
- **Markets** — crypto spot and perpetual futures, and from which exchanges.
- **History** — how far back the data goes, and who has to supply it.
- **Costs** — fees, slippage, funding on perpetuals, leverage and liquidation.
- **Robustness checks** — optimization, walk-forward, out-of-sample and Monte Carlo.
- **Free tier** — what you get without paying.
- **Exports** — whether you can take the trades and the equity curve elsewhere.
- **Trading** — whether the same tool can also run the strategy on an exchange.

## Side by side (November 2026)

|  | TradingView | QuantConnect | Freqtrade | backtrader | 3Commas | Gainium | Janorium |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Code or no code | Pine Script; built-in and community strategies run without code | Python or C# | Python | Python | No code: bot settings | No code: bot settings | No code: visual editor or JSON config |
| Crypto markets | Crypto from many exchanges, plus stocks, forex and more | Binance spot and perpetuals data, plus stocks, options, futures and forex | Spot on many exchanges; perpetuals on Binance, Bitget, Bybit, Gate, Hyperliquid, Kraken, OKX | Any market, with data you supply | Not stated | Pairs on the exchanges it connects to, spot and futures | Binance only: spot and USDT-M perpetuals, delisted perpetuals included |
| History | 5,000–40,000 intraday bars by plan; Deep Backtesting up to 2 million bars on Premium and up | Binance spot from 2017, perpetuals from 2020; tick and second data on paid tiers | What the exchange’s API returns; you download it | What you load | Up to the full history, depending on the plan | As much as the exchange provides for the pair | From 2020; 1h and 4h for every tracked coin, down to 5m for major pairs |
| Fees | Percent or fixed; 0 unless set | Binance fee models built in; the default model charges none on crypto | The exchange’s fee by default, or your own | Percent or fixed; you set it | Not stated | Exchange fees included | Maker and taker; the editor starts at Binance’s taker fee |
| Slippage | In ticks; 0 unless set | Constant, volume-share or market-impact model; 0 by default | None: orders fill at the requested price inside the candle | Percent or fixed; 0 by default | Not stated | Not simulated in bot backtests | Fixed, ATR-based or volume impact; the editor’s default depends on the coin |
| Funding on perpetuals | Not mentioned in the strategy docs | Funding models for Binance, Bybit and dYdX | From downloaded funding-rate data | No built-in support | Not stated | Not stated | Every real Binance settlement |
| Leverage and liquidation | Leverage; a generic margin-call model | Leverage; a generic margin-call model | Liquidation price computed; no liquidation fee | Leverage and margin; no liquidation | Not stated | Liquidation emulated from leverage and margin | 1–125×, Binance formula with a flat 0.4% maintenance margin |
| Robustness checks | No optimizer, Monte Carlo or walk-forward in the docs | Parameter optimization (paid), walk-forward by re-optimizing in code, holdout periods | Hyperopt; look-ahead and recursive-indicator analysis | Parameter sweeps on several cores; Sharpe, drawdown and SQN analyzers | Grid bots: backtests plus AI suggestions for the grid step | Comparison with random entries | Optimizer with walk-forward (rolling or anchored windows) and robustness scores; Monte Carlo, regime breakdowns, Deflated Sharpe and a random-entry baseline in the report |
| Free tier | The free Basic plan includes backtesting; deeper history on paid plans | Unlimited backtests on one node with minute data; the LEAN engine is open source | Free, open source (GPL-3.0) | Free, open source (GPL-3.0) | Starter, Pro and Expert plans; backtests per month and history depend on the plan | In-app backtests free and unlimited | Free; no paid tiers |
| Exports | Trades as CSV, report as XLSX, on paid plans | Orders and trades as CSV, results as JSON, a PDF report | Result files with every trade; analysis tables as CSV | Analyzer results in Python; charts via matplotlib | Not stated | Trade table from manual backtesting | ZIP with bar-by-bar and trades CSV; a printable tear sheet |
| Trades on an exchange | Through supported brokers | Yes, on paid plans | Yes, plus dry-run with virtual money | Interactive Brokers, Oanda, VisualChart | Yes | Yes, plus paper trading | No: backtesting only |

“Not stated” means the tool’s public pages did not say on 14 November 2026. Defaults matter as much as features: where a cost is 0 until you set it, a backtest run on the defaults is optimistic in any tool — Janorium included, since a config sent to its API without costs runs without them.

## When to pick which

- **TradingView** — you already chart there, trade more than crypto, or want the largest library of ready-made strategies, and Pine Script does not put you off.
- **QuantConnect** — you write Python or C#, want detailed cost models (fees per exchange, market-impact slippage, funding on perpetuals) and the same code to trade later, and the free tier’s limits suit you or you will pay.
- **Freqtrade** — you want to run your own crypto bot in Python, on your own machine and one of many exchanges, and test it with the same code that will trade.
- **backtrader** — you want a free Python library with full control and already have your data; nothing crypto-specific is built in.
- **3Commas or Gainium** — you trade DCA, grid or combo bots on that platform: its backtester tests exactly the bot you will run. Gainium’s in-app backtests are free.
- **Janorium** — you want no code, test on Binance spot or USDT-M perpetuals, want funding, fees, slippage and liquidation in the result, walk-forward and robustness checks, for free — and you accept that it only backtests.

## TradingView Strategy Tester

The strengths are reach and community: the charts most traders already use, about 3.5 million instruments, and a very large library of built-in and published strategies that run without code. Pine Script can express logic that rule builders cannot, and on Premium plans and up the “Bar detalization” setting fills orders from lower-timeframe data inside each bar. Since July 2026 the Strategy Tester is called the strategy report.

The caveats for crypto: commission and slippage are 0 until the script or you set them, the strategy documentation does not mention funding on perpetuals, margin calls follow a generic model rather than an exchange’s liquidation formula, and intraday history is capped by plan unless you use Deep Backtesting on Premium and up. The detailed comparison is in [TradingView backtesting vs Janorium](https://janorium.com/guides/tradingview-backtesting-vs-janorium).

## QuantConnect

QuantConnect is the most complete engine here for people who code. Strategies are Python or C# on LEAN, an open-source engine (Apache 2.0) you can also run yourself. Its cost models are documented one by one: Binance spot and futures fees; constant, volume-share and market-impact slippage; funding on Binance, Bybit and dYdX perpetuals; margin calls. Walk-forward is done in code by re-optimizing on a trailing window, and an organization can lock a holdout period. Binance data starts in 2017 for spot and in 2020 for perpetuals, results download as CSV, JSON and a PDF report, and the same algorithm can trade on an exchange.

The caveats: you need to program, and the defaults are optimistic — the default brokerage model charges no fees on crypto and the default slippage is zero, so the right models have to be chosen. The free tier runs unlimited backtests on one node with minute-to-daily data; tick and second data, parameter optimization and trading on an exchange are on paid tiers.

## Freqtrade and backtrader (Python)

**Freqtrade** is a free, open-source (GPL-3.0) crypto trading bot with a backtester. Its strengths are breadth and care about bias: spot on many exchanges, perpetual futures on seven, funding applied from downloaded funding-rate data, hyperparameter search, and two tools most backtesters lack — a look-ahead analysis and a recursive-indicator analysis that catch strategies peeking into the future. The same strategy runs in dry-run mode and with real money.

Its documented assumptions are worth knowing: orders fill at the requested price with no slippage as long as that price is inside the candle, stop-losses fill exactly at the stop even if the low was lower (the loss is raised by twice the fees), and liquidation fees are not calculated. You download the history yourself, so its depth is whatever the exchange’s API gives.

**backtrader** is a free, open-source (GPL-3.0) Python library rather than a platform: a flexible event-driven engine with market, limit, stop, OCO and bracket orders, commission schemes, percentage or fixed slippage, and a large set of indicators and analyzers. It is market-agnostic, so nothing crypto-specific is built in: you bring the data (crypto exchanges come through third-party add-ons), and there is no funding or liquidation model. The latest release on PyPI, 1.9.78.123, dates from April 2023.

## Backtesters inside bot platforms: 3Commas and Gainium

Bot platforms test their own bots: you set up a DCA, grid or combo bot the way you would run it, then replay it on history. The strength is that what you test is exactly what will trade. The limit is the same thing: you test the bot types the platform offers, with the conditions they support, not an arbitrary strategy.

**Gainium** says its in-app backtests are free and unlimited on every plan, run in your browser, include exchange fees and emulate liquidation from your leverage and margin, and it can compare a bot with random entries over the same period. Its own FAQ says slippage is not simulated in bot backtests, and its pages do not mention funding.

**3Commas** advertises backtesting on 1-minute candles with indicators in the bot logic over a coin’s full history, and its plan table limits backtests per month and the history they use. Its public pages do not say how backtests treat fees, slippage, funding or liquidation, and several of its bot pages now redirect to QuantPilot, a separate product by 3Commas. Check what your plan includes before relying on it.

## Janorium’s limits

Janorium is a free, no-code backtester for crypto strategies on Binance. It does fewer things than most tools above, and these limits matter most:

- **Binance only.** Spot and USDT-M perpetual futures. No other exchanges, no coin-margined futures, no stocks or forex.
- **History from 2020.** A backtest cannot start earlier, while QuantConnect’s Binance spot data goes back to 2017. Delisted perpetuals are kept; delisted spot coins are not.
- **No order-book fills.** An order fills at the candle’s close (or the next open), moved by modeled slippage. Order-book depth, market impact beyond the volume-based slippage and the bid-ask spread are not modeled.
- **No partial fills.** An entry that does not fit the free equity is skipped, never filled in part. Latency, lot size, tick size and minimum order value are not modeled either.
- **Simplified liquidation.** A flat 0.4% maintenance margin, no tiered margin, no liquidation fee and no auto-deleveraging — real liquidations can come earlier and cost more.
- **No custom code.** A strategy is a declarative config. Logic the config cannot express needs Pine Script, Python or C#.
- **No trading.** Janorium never places orders, on any exchange. To run a strategy, you need another tool.

Every assumption is listed in [how Janorium backtests](https://janorium.com/guides/how-janorium-backtests). What it does offer, for free: historical funding on every perpetual position, maker and taker fees, three slippage models, an optimizer (grid or random search over up to 12 parameters, walk-forward with rolling or anchored windows, robustness scores and sensitivity heatmaps), Monte Carlo resampling, drawdown episodes, regime breakdowns, a Deflated Sharpe note and a random-entry baseline in every report, a bar-by-bar CSV export, and [an AI assistant that writes the config](https://janorium.com/guides/build-a-strategy-with-ai) from your description.

## Comparing results across tools

The same strategy can show very different results in two tools, and the cause is rarely the logic. Before comparing, align what differs most:

- **Costs.** TradingView, QuantConnect and backtrader start with zero slippage, and TradingView and QuantConnect’s default model charge no fees on crypto either. Set the same fees and slippage everywhere.
- **Funding.** On perpetuals, funding can outweigh fees over months of holding. A tool that leaves it out flatters longs whenever the rate is positive.
- **Fills.** Filling at the signal candle’s close or at the next open, and whether the stop or the target counts first inside a bar, move a result more than most parameters.
- **History.** Test over the same period. A strategy tested only on 2021, or only after 2022, has seen one market.

Whichever tool you use, read [what backtesting is and how it misleads](https://janorium.com/guides/what-is-backtesting) before trusting a result.

TradingView, QuantConnect, Freqtrade, backtrader, 3Commas and Gainium are names or trademarks of their owners. Janorium is an independent project, not affiliated with or endorsed by any of them. Plans and features change — check each tool’s own site for the current details.

## Questions

### Is there a free crypto backtesting tool without coding?

Yes. Janorium is free and needs no code for Binance spot and USDT-M perpetuals. TradingView’s free Basic plan runs built-in and community strategies without code, though your own logic means Pine Script. Gainium’s in-app backtests of DCA, grid and combo bots are free too.

### Which backtesting tools include funding fees on perpetual futures?

Of the tools compared here: QuantConnect (funding models for Binance, Bybit and dYdX), Freqtrade (from downloaded funding-rate data) and Janorium (every Binance settlement). TradingView’s strategy documentation does not mention funding, backtrader has no built-in support, and 3Commas and Gainium do not say.

### What is a good TradingView Strategy Tester alternative for crypto?

If you code, QuantConnect or Freqtrade: both model crypto fees and funding and can also trade on an exchange. If you do not code and test on Binance, Janorium. If you run DCA or grid bots, the backtester of the bot platform you use.

### Is QuantConnect free?

There is a free tier with unlimited backtests on one backtest node and minute-to-daily data, and its LEAN engine is open source under Apache 2.0. Tick and second data, more nodes, parameter optimization and trading on an exchange are on paid tiers.

### Does Freqtrade model slippage in backtests?

No. Its documentation says orders fill at the requested price, with no slippage, as long as that price is inside the candle’s range, and stop-losses fill exactly at the stop price. Setting a fee above the exchange’s leaves some margin for it.

### Can Janorium run a strategy on an exchange?

No. Janorium only backtests and never places orders on any exchange. Freqtrade, QuantConnect and the bot platforms can trade, but the strategy has to be rebuilt there.

## Sources

- [TradingView — Pine Script docs: Strategies](https://www.tradingview.com/pine-script-docs/concepts/strategies/)
- [TradingView — Pine Script release notes (July 2026)](https://www.tradingview.com/pine-script-docs/release-notes/)
- [TradingView — plans and pricing](https://www.tradingview.com/pricing/)
- [TradingView Help Center — Deep Backtesting](https://www.tradingview.com/support/solutions/43000666265/)
- [TradingView Help Center — historical bars per plan](https://www.tradingview.com/support/solutions/43000480679/)
- [TradingView Help Center — bar detalization](https://www.tradingview.com/support/solutions/43000786180/)
- [TradingView — data coverage](https://www.tradingview.com/data-coverage/)
- [QuantConnect — pricing](https://www.quantconnect.com/pricing)
- [QuantConnect — features by tier](https://www.quantconnect.com/docs/v2/cloud-platform/organizations/tier-features)
- [QuantConnect — transaction fees: key concepts](https://www.quantconnect.com/docs/v2/writing-algorithms/reality-modeling/transaction-fees/key-concepts)
- [QuantConnect — transaction fees: supported models](https://www.quantconnect.com/docs/v2/writing-algorithms/reality-modeling/transaction-fees/supported-models)
- [QuantConnect — slippage](https://www.quantconnect.com/docs/v2/writing-algorithms/reality-modeling/slippage/key-concepts)
- [QuantConnect — margin interest (funding) models](https://www.quantconnect.com/docs/v2/writing-algorithms/reality-modeling/margin-interest-rate/supported-models)
- [QuantConnect — margin calls](https://www.quantconnect.com/docs/v2/writing-algorithms/reality-modeling/margin-calls)
- [QuantConnect — Binance crypto price data](https://www.quantconnect.com/docs/v2/writing-algorithms/datasets/quantconnect/binance-crypto-price-data)
- [QuantConnect — Binance crypto future price data](https://www.quantconnect.com/docs/v2/writing-algorithms/datasets/quantconnect/binance-crypto-future-price-data)
- [QuantConnect — walk-forward optimization](https://www.quantconnect.com/docs/v2/writing-algorithms/optimization/walk-forward-optimization)
- [QuantConnect — backtest results and downloads](https://www.quantconnect.com/docs/v2/cloud-platform/backtesting/results)
- [QuantConnect — LEAN engine on GitHub](https://github.com/QuantConnect/Lean)
- [Freqtrade — documentation and supported exchanges](https://www.freqtrade.io/en/stable/)
- [Freqtrade — backtesting and its assumptions](https://www.freqtrade.io/en/stable/backtesting/)
- [Freqtrade — leverage, funding and liquidation](https://www.freqtrade.io/en/stable/leverage/)
- [Freqtrade — downloading data](https://www.freqtrade.io/en/stable/data-download/)
- [Freqtrade — hyperopt](https://www.freqtrade.io/en/stable/hyperopt/)
- [Freqtrade — look-ahead analysis](https://www.freqtrade.io/en/stable/lookahead-analysis/)
- [Freqtrade on GitHub (license)](https://github.com/freqtrade/freqtrade)
- [backtrader — documentation](https://www.backtrader.com/docu/)
- [backtrader — commission schemes](https://www.backtrader.com/docu/commission-schemes/commission-schemes/)
- [backtrader — slippage](https://www.backtrader.com/docu/slippage/slippage/)
- [backtrader on PyPI (latest release)](https://pypi.org/project/backtrader/)
- [3Commas — home page](https://3commas.io/)
- [3Commas — pricing](https://3commas.io/pricing)
- [Gainium — crypto backtesting](https://gainium.io/crypto-backtesting)
- [Gainium — pricing](https://gainium.io/pricing)
- [Gainium Help — manual backtesting](https://gainium.io/help/manual-backtesting)
