freqtrade_origin/docs/backtesting.md

257 lines
15 KiB
Markdown
Raw Normal View History

# Backtesting
2018-06-13 04:52:17 +00:00
2019-09-21 08:09:14 +00:00
This page explains how to validate your strategy performance by using Backtesting.
2019-09-21 08:09:14 +00:00
Backtesting requires historic data to be available.
2019-09-24 05:05:30 +00:00
To learn how to get data for the pairs and exchange you're interested in, head over to the [Data Downloading](data-download.md) section of the documentation.
## Test your strategy with Backtesting
2018-06-13 04:52:17 +00:00
Now you have good Buy and Sell strategies and some historic data, you want to test it against
2018-12-30 16:14:56 +00:00
real data. This is what we call
[backtesting](https://en.wikipedia.org/wiki/Backtesting).
Backtesting will use the crypto-currencies (pairs) from your config file and load historical candle (OHCLV) data from `user_data/data/<exchange>` by default.
2020-06-02 08:06:26 +00:00
If no data is available for the exchange / pair / timeframe combination, backtesting will ask you to download them first using `freqtrade download-data`.
2019-09-24 05:05:30 +00:00
For details on downloading, please refer to the [Data Downloading](data-download.md) section in the documentation.
2019-09-21 08:09:14 +00:00
The result of backtesting will confirm if your bot has better odds of making a profit than a loss.
2020-04-25 13:38:20 +00:00
!!! Warning "Using dynamic pairlists for backtesting"
Using dynamic pairlists is possible, however it relies on the current market conditions - which will not reflect the historic status of the pairlist.
Also, when using pairlists other than StaticPairlist, reproducability of backtesting-results cannot be guaranteed.
Please read the [pairlists documentation](configuration.md#pairlists) for more information.
To achieve reproducible results, best generate a pairlist via the [`test-pairlist`](utils.md#test-pairlist) command and use that as static pairlist.
### Run a backtesting against the currencies listed in your config file
2019-09-21 08:09:14 +00:00
#### With 5 min candle (OHLCV) data (per default)
2018-06-13 04:52:17 +00:00
```bash
freqtrade backtesting
```
#### With 1 min candle (OHLCV) data
2018-06-13 04:52:17 +00:00
```bash
2020-06-02 08:06:26 +00:00
freqtrade backtesting --timeframe 1m
```
#### Using a different on-disk historical candle (OHLCV) data source
2018-06-13 04:52:17 +00:00
2019-08-20 05:05:17 +00:00
Assume you downloaded the history data from the Bittrex exchange and kept it in the `user_data/data/bittrex-20180101` directory.
You can then use this data for backtesting as follows:
2018-01-07 09:15:26 +00:00
```bash
2019-10-05 07:56:01 +00:00
freqtrade --datadir user_data/data/bittrex-20180101 backtesting
2018-01-07 09:15:26 +00:00
```
2018-06-13 04:52:17 +00:00
#### With a (custom) strategy file
```bash
freqtrade backtesting -s SampleStrategy
```
2018-06-13 04:52:17 +00:00
2019-08-27 04:41:07 +00:00
Where `-s SampleStrategy` refers to the class name within the strategy file `sample_strategy.py` found in the `freqtrade/user_data/strategies` directory.
2019-08-06 04:27:38 +00:00
#### Comparing multiple Strategies
```bash
2020-06-02 08:06:26 +00:00
freqtrade backtesting --strategy-list SampleStrategy1 AwesomeStrategy --timeframe 5m
2019-08-06 04:27:38 +00:00
```
2019-08-27 04:41:07 +00:00
Where `SampleStrategy1` and `AwesomeStrategy` refer to class names of strategies.
2018-06-13 04:52:17 +00:00
#### Exporting trades to file
2018-01-11 14:45:39 +00:00
```bash
freqtrade backtesting --export trades --config config.json --strategy SampleStrategy
2018-01-11 14:45:39 +00:00
```
The exported trades can be used for [further analysis](#further-backtest-result-analysis), or can be used by the plotting script `plot_dataframe.py` in the scripts directory.
2018-07-12 19:37:54 +00:00
2018-06-13 04:52:17 +00:00
#### Exporting trades to file specifying a custom filename
```bash
2019-08-27 04:41:07 +00:00
freqtrade backtesting --export trades --export-filename=backtest_samplestrategy.json
```
Please also read about the [strategy startup period](strategy-customization.md#strategy-startup-period).
2019-10-05 13:33:44 +00:00
#### Supplying custom fee value
2019-10-07 05:02:43 +00:00
Sometimes your account has certain fee rebates (fee reductions starting with a certain account size or monthly volume), which are not visible to ccxt.
To account for this in backtesting, you can use the `--fee` command line option to supply this value to backtesting.
2020-01-03 10:20:08 +00:00
This fee must be a ratio, and will be applied twice (once for trade entry, and once for trade exit).
2019-10-05 13:33:44 +00:00
For example, if the buying and selling commission fee is 0.1% (i.e., 0.001 written as ratio), then you would run backtesting as the following:
2019-10-05 13:33:44 +00:00
```bash
freqtrade backtesting --fee 0.001
```
2020-01-03 10:20:08 +00:00
!!! Note
2020-01-03 19:11:58 +00:00
Only supply this option (or the corresponding configuration parameter) if you want to experiment with different fee values. By default, Backtesting fetches the default fee from the exchange pair/market info.
2019-10-05 13:33:44 +00:00
2019-09-26 09:00:26 +00:00
#### Running backtest with smaller testset by using timerange
2019-09-21 08:09:14 +00:00
Use the `--timerange` argument to change how much of the testset you want to use.
2018-01-10 23:14:36 +00:00
2019-09-26 09:00:26 +00:00
For example, running backtesting with the `--timerange=20190501-` option will use all available data starting with May 1st, 2019 from your inputdata.
2018-06-13 04:52:17 +00:00
2018-01-10 23:14:36 +00:00
```bash
2019-09-21 08:09:14 +00:00
freqtrade backtesting --timerange=20190501-
2018-01-10 23:14:36 +00:00
```
2019-09-24 05:05:30 +00:00
You can also specify particular dates or a range span indexed by start and stop.
The full timerange specification:
2018-06-13 04:52:17 +00:00
- Use tickframes till 2018/01/31: `--timerange=-20180131`
- Use tickframes since 2018/01/31: `--timerange=20180131-`
- Use tickframes since 2018/01/31 till 2018/03/01 : `--timerange=20180131-20180301`
- Use tickframes between POSIX timestamps 1527595200 1527618600:
`--timerange=1527595200-1527618600`
2018-01-15 21:25:02 +00:00
## Understand the backtesting result
2018-06-13 04:52:17 +00:00
The most important in the backtesting is to understand the result.
A backtesting result will look like that:
2018-06-13 04:52:17 +00:00
```
2019-01-24 14:24:38 +00:00
========================================================= BACKTESTING REPORT ========================================================
| Pair | Buys | Avg Profit % | Cum Profit % | Tot Profit BTC | Tot Profit % | Avg Duration | Wins | Draws | Losses |
|:---------|-------:|---------------:|---------------:|-----------------:|---------------:|:---------------|------:|-------:|--------:|
| ADA/BTC | 35 | -0.11 | -3.88 | -0.00019428 | -1.94 | 4:35:00 | 14 | 0 | 21 |
| ARK/BTC | 11 | -0.41 | -4.52 | -0.00022647 | -2.26 | 2:03:00 | 3 | 0 | 8 |
| BTS/BTC | 32 | 0.31 | 9.78 | 0.00048938 | 4.89 | 5:05:00 | 18 | 0 | 14 |
| DASH/BTC | 13 | -0.08 | -1.07 | -0.00005343 | -0.53 | 4:39:00 | 6 | 0 | 7 |
| ENG/BTC | 18 | 1.36 | 24.54 | 0.00122807 | 12.27 | 2:50:00 | 8 | 0 | 10 |
| EOS/BTC | 36 | 0.08 | 3.06 | 0.00015304 | 1.53 | 3:34:00 | 16 | 0 | 20 |
| ETC/BTC | 26 | 0.37 | 9.51 | 0.00047576 | 4.75 | 6:14:00 | 11 | 0 | 15 |
| ETH/BTC | 33 | 0.30 | 9.96 | 0.00049856 | 4.98 | 7:31:00 | 16 | 0 | 17 |
| IOTA/BTC | 32 | 0.03 | 1.09 | 0.00005444 | 0.54 | 3:12:00 | 14 | 0 | 18 |
| LSK/BTC | 15 | 1.75 | 26.26 | 0.00131413 | 13.13 | 2:58:00 | 6 | 0 | 9 |
| LTC/BTC | 32 | -0.04 | -1.38 | -0.00006886 | -0.69 | 4:49:00 | 11 | 0 | 21 |
| NANO/BTC | 17 | 1.26 | 21.39 | 0.00107058 | 10.70 | 1:55:00 | 10 | 0 | 7 |
| NEO/BTC | 23 | 0.82 | 18.97 | 0.00094936 | 9.48 | 2:59:00 | 10 | 0 | 13 |
| REQ/BTC | 9 | 1.17 | 10.54 | 0.00052734 | 5.27 | 3:47:00 | 4 | 0 | 5 |
| XLM/BTC | 16 | 1.22 | 19.54 | 0.00097800 | 9.77 | 3:15:00 | 7 | 0 | 9 |
| XMR/BTC | 23 | -0.18 | -4.13 | -0.00020696 | -2.07 | 5:30:00 | 12 | 0 | 11 |
| XRP/BTC | 35 | 0.66 | 22.96 | 0.00114897 | 11.48 | 3:49:00 | 12 | 0 | 23 |
| ZEC/BTC | 22 | -0.46 | -10.18 | -0.00050971 | -5.09 | 2:22:00 | 7 | 0 | 15 |
| TOTAL | 429 | 0.36 | 152.41 | 0.00762792 | 76.20 | 4:12:00 | 186 | 0 | 243 |
2019-01-24 14:24:38 +00:00
========================================================= SELL REASON STATS =========================================================
| Sell Reason | Sells | Wins | Draws | Losses |
|:-------------------|--------:|------:|-------:|--------:|
| trailing_stop_loss | 205 | 150 | 0 | 55 |
| stop_loss | 166 | 0 | 0 | 166 |
| sell_signal | 56 | 36 | 0 | 20 |
| force_sell | 2 | 0 | 0 | 2 |
2019-01-24 14:24:38 +00:00
====================================================== LEFT OPEN TRADES REPORT ======================================================
| Pair | Buys | Avg Profit % | Cum Profit % | Tot Profit BTC | Tot Profit % | Avg Duration | Wins | Draws | Losses |
|:---------|-------:|---------------:|---------------:|-----------------:|---------------:|:---------------|------:|-------:|--------:|
| ADA/BTC | 1 | 0.89 | 0.89 | 0.00004434 | 0.44 | 6:00:00 | 1 | 0 | 0 |
| LTC/BTC | 1 | 0.68 | 0.68 | 0.00003421 | 0.34 | 2:00:00 | 1 | 0 | 0 |
| TOTAL | 2 | 0.78 | 1.57 | 0.00007855 | 0.78 | 4:00:00 | 2 | 0 | 0 |
```
2019-09-21 08:09:14 +00:00
The 1st table contains all trades the bot made, including "left open trades".
2019-09-21 08:09:14 +00:00
The 2nd table contains a recap of sell reasons.
2019-12-25 08:39:44 +00:00
This table can tell you which area needs some additional work (i.e. all `sell_signal` trades are losses, so we should disable the sell-signal or work on improving that).
2019-01-24 14:24:38 +00:00
2019-09-21 08:09:14 +00:00
The 3rd table contains all trades the bot had to `forcesell` at the end of the backtest period to present a full picture.
This is necessary to simulate realistic behaviour, since the backtest period has to end at some point, while realistically, you could leave the bot running forever.
These trades are also included in the first table, but are extracted separately for clarity.
The last line will give you the overall performance of your strategy,
here:
2018-06-13 04:52:17 +00:00
```
2019-01-24 14:24:38 +00:00
| TOTAL | 429 | 0.36 | 152.41 | 0.00762792 | 76.20 | 4:12:00 | 186 | 243 |
```
2019-09-21 08:09:14 +00:00
The bot has made `429` trades for an average duration of `4:12:00`, with a performance of `76.20%` (profit), that means it has
2019-01-24 14:24:38 +00:00
earned a total of `0.00762792 BTC` starting with a capital of 0.01 BTC.
The column `avg profit %` shows the average profit for all trades made while the column `cum profit %` sums up all the profits/losses.
2019-09-21 08:09:14 +00:00
The column `tot profit %` shows instead the total profit % in relation to allocated capital (`max_open_trades * stake_amount`).
In the above results we have `max_open_trades=2` and `stake_amount=0.005` in config so `tot_profit %` will be `(76.20/100) * (0.005 * 2) =~ 0.00762792 BTC`.
2018-12-30 16:14:56 +00:00
2019-09-21 08:09:14 +00:00
Your strategy performance is influenced by your buy strategy, your sell strategy, and also by the `minimal_roi` and `stop_loss` you have set.
For example, if your `minimal_roi` is only `"0": 0.01` you cannot expect the bot to make more profit than 1% (because it will sell every time a trade reaches 1%).
2018-06-13 04:52:17 +00:00
```json
"minimal_roi": {
"0": 0.01
},
```
On the other hand, if you set a too high `minimal_roi` like `"0": 0.55`
(55%), there is almost no chance that the bot will ever reach this profit.
Hence, keep in mind that your performance is an integral mix of all different elements of the strategy, your configuration, and the crypto-currency pairs you have set up.
2019-09-21 08:09:14 +00:00
### Assumptions made by backtesting
Since backtesting lacks some detailed information about what happens within a candle, it needs to take a few assumptions:
- Buys happen at open-price
2019-10-18 08:00:43 +00:00
- Sell signal sells happen at open-price of the following candle
- Low happens before high for stoploss, protecting capital first
2019-12-07 14:26:10 +00:00
- ROI
- sells are compared to high - but the ROI value is used (e.g. ROI = 2%, high=5% - so the sell will be at 2%)
- sells are never "below the candle", so a ROI of 2% may result in a sell at 2.4% if low was at 2.4% profit
- Forcesells caused by `<N>=-1` ROI entries use low as sell value, unless N falls on the candle open (e.g. `120: -1` for 1h candles)
2019-09-21 08:09:14 +00:00
- Stoploss sells happen exactly at stoploss price, even if low was lower
- Trailing stoploss
- High happens first - adjusting stoploss
- Low uses the adjusted stoploss (so sells with large high-low difference are backtested correctly)
- Sell-reason does not explain if a trade was positive or negative, just what triggered the sell (this can look odd if negative ROI values are used)
- Stoploss (and trailing stoploss) is evaluated before ROI within one candle. So you can often see more trades with the `stoploss` and/or `trailing_stop` sell reason comparing to the results obtained with the same strategy in the Dry Run/Live Trade modes.
2019-10-18 08:00:43 +00:00
Taking these assumptions, backtesting tries to mirror real trading as closely as possible. However, backtesting will **never** replace running a strategy in dry-run mode.
Also, keep in mind that past results don't guarantee future success.
In addition to the above assumptions, strategy authors should carefully read the [Common Mistakes](strategy-customization.md#common-mistakes-when-developing-strategies) section, to avoid using data in backtesting which is not available in real market conditions.
2019-03-07 20:11:14 +00:00
### Further backtest-result analysis
To further analyze your backtest results, you can [export the trades](#exporting-trades-to-file).
2019-06-22 14:18:22 +00:00
You can then load the trades to perform further analysis as shown in our [data analysis](data-analysis.md#backtesting) backtesting section.
2019-03-07 20:11:14 +00:00
## Backtesting multiple strategies
2019-09-21 08:09:14 +00:00
To compare multiple strategies, a list of Strategies can be provided to backtesting.
2020-06-02 08:06:26 +00:00
This is limited to 1 timeframe value per run. However, data is only loaded once from disk so if you have multiple
2019-09-21 08:09:14 +00:00
strategies you'd like to compare, this will give a nice runtime boost.
All listed Strategies need to be in the same directory.
``` bash
2020-06-02 08:06:26 +00:00
freqtrade backtesting --timerange 20180401-20180410 --timeframe 5m --strategy-list Strategy001 Strategy002 --export trades
```
This will save the results to `user_data/backtest_results/backtest-result-<strategy>.json`, injecting the strategy-name into the target filename.
2018-07-29 11:13:03 +00:00
There will be an additional table comparing win/losses of the different strategies (identical to the "Total" row in the first table).
2019-09-21 08:09:14 +00:00
Detailed output for all strategies one after the other will be available, so make sure to scroll up to see the details per strategy.
2018-07-29 11:13:03 +00:00
```
=========================================================== STRATEGY SUMMARY ===========================================================
| Strategy | Buys | Avg Profit % | Cum Profit % | Tot Profit BTC | Tot Profit % | Avg Duration | Wins | Draws | Losses |
|:------------|-------:|---------------:|---------------:|-----------------:|---------------:|:---------------|------:|-------:|-------:|
| Strategy1 | 429 | 0.36 | 152.41 | 0.00762792 | 76.20 | 4:12:00 | 186 | 0 | 243 |
| Strategy2 | 1487 | -0.13 | -197.58 | -0.00988917 | -98.79 | 4:43:00 | 662 | 0 | 825 |
2018-07-29 11:13:03 +00:00
```
## Next step
2018-06-13 04:52:17 +00:00
Great, your strategy is profitable. What if the bot can give your the
2018-12-30 16:14:56 +00:00
optimal parameters to use for your strategy?
2018-12-31 12:36:27 +00:00
Your next step is to learn [how to find optimal parameters with Hyperopt](hyperopt.md)