Features

With Interactive Brokers, Oanda v1, VisualChart and also with external 3^rd party brokers (alpaca, Oanda v2, ccxt, ...)

Use 0 in arrays for the present moment to address the look-ahead bias when accessing values in arrays
Use -1, -2 (i.e.: negative values) for the last moments, to keep in sync with Python's definition
Any positive index means the future (test your code in event-only mode and it will break)

The trading logic and the broker are always run on an event by event basis
The calculation for indicators is vectorized if possible (source data can be preloaded)

Everything can be run in event-only mode with no data preloaded, just like if things were live.

Built-in support for several sources: CSV, Database-Sources, YahooFinance, Interactive Brokers, Oanda v1, ...
Any number of simultaneous data feeds (memory constrained, obviously) can be run simultaneously

Warning

Beware of survivorshipt bias!
Multiple timeframes can be mixed and run
Integrated Resampling and Replaying

Order Types: Market, Limit, Stop, StopLimit, StopTrail, StopTrailLimit, OCO, Bracket, MarketOnClose
Long Short selling
Continuous Cash Adjustment for Future-like instruments
User defined commission schemes and credit interest
Fund Mode
Volume Filling Strategies
Custom Slippage

Automatic warm-up period calculation before operating
Multiple Strategies (against same broker) can run in parallel
Multiple order generation methods (buy/sell, order_target_xxx, automated signals)
Event notification for: Incoming Data, Data Feed provider, orders, trades, timers.

Over 122 indicators with the usual suspects on board

Many moving averages (SMA, EMA, ...), classic: (MACD, Stochastic, RSI, ...) and others
ta-lib integration

Several built-in performance analyzers (TimeReturns, TradeAnalyzer, SharpeRatio, VWR, SQN, ...)

Automated (customizable) plotting with a single command.

Note

For this to work, matplotlib has to be installed

Define and plug smart automated staking policies

Agents which will be plotted and can observe everything in the system (usually will be used to plot statistics)

Work with one of the most powerful, yet easy to use, programming languages. No external libraries required.

Uses OO to easily let the pieces of the puzzle be fitted to each other.
Operators are overloaded, where possible, to provide natural language constructs such as:
```
av_diff = bt.ind.SMA(period=30) - bt.ind.SMA(period=15)
```
where av_diff will contain the difference of the simple moving averages of 30 and 15 periods.
For language constructs, cannot be overrideen, like and, or, if, equivalent functions are provided to ensure no functionality is missing, such as
```
av_and = bt.And(av_diff > 0, self.data.close < bt.ind.SMA(period=30))
```