ab9e2fcea0
fix guard names to match search space
2018-06-23 15:47:19 +03:00
136456afc0
add three triggers to hyperopting
2018-06-23 15:44:51 +03:00
e8f2e6956d
to avoid pickle problems, get rid of reference to exchange after initialization
2018-06-23 14:37:36 +03:00
a525cba8e9
switch signal handler to try catch. fix pickling and formatting output
2018-06-23 14:37:36 +03:00
8272120c3a
convert stoploss and ROI search spaces to skopt format
2018-06-23 14:37:36 +03:00
8fee2e2409
move result logging out from optimizer
2018-06-23 14:37:36 +03:00
c415014153
use multiple jobs in acq
2018-06-23 14:37:36 +03:00
964cbdc262
increase initial sampling points
2018-06-23 14:37:36 +03:00
a46badd5c0
reuse pool workers
2018-06-23 14:37:36 +03:00
0cb1aedf5b
problem with pickling
2018-06-23 14:37:36 +03:00
b485e6e0ba
start small
2018-06-23 14:37:36 +03:00
78f50a1471
move logic from hyperopt to freqtrade.strategy
2018-06-23 14:37:36 +03:00
5aae215c94
wrap strategies with HyperoptStrategy for module lookups with pickle
2018-06-23 14:37:36 +03:00
a68c90c512
avoid calling exchange.get_fee inside loop
2018-06-23 14:37:36 +03:00
251f7db3ca
require exchange object to delete pairs
2018-06-17 23:38:07 +02:00
21edcbdc27
Refactor exchange to class
2018-06-17 23:38:07 +02:00
a5511e2e30
Merge pull request #894 from freqtrade/feature/force_close_backtest
...
Display open trades after backtest period
2018-06-16 12:49:08 +02:00
0c85febe76
remove all mongodb related code
2018-06-16 09:09:28 +03:00
c1f8f641e6
remove use of hyperopt_conf.py
2018-06-16 09:09:28 +03:00
c0289ad844
use list comprehension to build list
2018-06-13 19:53:12 +02:00
e600be4f56
Reduce force-sell verbosity
2018-06-13 19:44:00 +02:00
6357812743
fix backtest report able
2018-06-13 06:57:49 +02:00
e3ced7c15e
extract export from backtest function
2018-06-12 22:29:30 +02:00
bfde33c945
Use timestamp() instead of strftime
...
this will avoid a bug shifting epoch time by 1 hour:
https://stackoverflow.com/questions/11743019/convert-python-datetime-to-epoch-with-strftime
2018-06-12 21:12:55 +02:00
335d1fbbbc
Check if no backtest data is found and fail gracefully
2018-06-11 19:50:43 +02:00
12e455cbf5
add buy/sell index to backtest result
2018-06-10 20:52:42 +02:00
4710210cff
fix hyperopt to use new backtesting result tuple
2018-06-10 13:56:10 +02:00
27ee8f7360
make flake happy
2018-06-10 13:55:48 +02:00
1cd7ac55a8
Added "left open trades" report
2018-06-10 13:45:16 +02:00
b81588307f
Add "open_at_end" parameter
2018-06-10 13:37:53 +02:00
31025216f9
fix type of open/close timestmap
2018-06-10 13:32:07 +02:00
322a528c12
fix bug with backtestResult
2018-06-10 13:25:16 +02:00
9c57d3aa8b
add BacktestresultTuple
2018-06-10 13:15:46 +02:00
c1b2e06eda
simplify return from _get_sell_trade_entry
2018-06-10 09:07:04 +02:00
3094acc7fb
update comment
2018-06-10 08:58:28 +02:00
24a875ed46
remove experimental parameters - they are read by analyze.py anyway
2018-06-09 21:44:57 +02:00
5623ea3ac6
Add forcesell at end of backtest period
2018-06-09 21:44:20 +02:00
8effc5f929
fix windows-specific init issue with named tuple
2018-06-08 19:46:07 +02:00
b4ae5a36a8
use .copy() to avoid Pandas mistake. drop first row because of shifting
2018-06-07 17:29:40 +03:00
7f8e0ba25f
use buy/sell signal from previous candle, not current to avoid seeing to the future
2018-06-07 17:28:40 +03:00
f37c5b70ba
Fix tests - read optional argument
2018-06-05 23:53:49 +02:00
7a34578b4d
refactor timerange to named tuple
2018-06-05 23:34:26 +02:00
e3227a741c
add --export-filename for backtesting
2018-06-03 19:36:53 +02:00
50fc5f91ca
Merge branch 'develop' into mypy_typecheck
2018-06-03 10:35:56 +02:00
a657e3d24a
Merge pull request #826 from gcarq/fix/hyperopt-stake_currency
...
Fix stake_currency returned by Hyperopt …
2018-06-03 07:19:24 +03:00
2cd8782a88
Merge pull request #825 from gcarq/fix/hyperopt-in-progress
...
Fix the in-progress dot that does not show up during a Hyperopt run
2018-06-03 07:16:39 +03:00
fe8ff1b929
Fix stake_currency return by Hyperopt
...
Hyperopt had BTC hard coded in the result. This commit will display
the real stake_currency used.
If you used `"stake_currency": "USDT",` in your config file.
Before this commit you saw a message like:
"2 trades. Avg profit 0.13%. Total profit 0.00002651 BTC (0.0027Σ%). Avg duration 142.5 mins."
Now with the commit, we fix the wrong BTC currency:
"2 trades. Avg profit 0.13%. Total profit 0.00002651 USDT (0.0027Σ%). Avg duration 142.5 mins."
2018-06-02 14:07:31 -07:00
127cf5d619
Backtesting: Add the Interval required when data is missing
...
Change the message:
"No data for pair ETH/BTC, use --refresh-pairs-cached to download the data"
for:
"No data for pair: "ETH/BTC", Interval: 5m. Use --refresh-pairs-cached to download the data"
The message structure is unified with the download message:
"Download the pair: "ETH/BTC", Interval: 5m"
2018-06-02 13:55:05 -07:00
dc65753a64
Fix the in-progress dot that does not show up during a Hyperopt run
2018-06-02 12:35:07 -07:00
2791d543ea
Make backtesting report markdown shareable
...
Small tweak to make the backtesting report markdown ready and much easier to share reports on many markdown publishing tools and editors that already support Markdown Extra with just a copy and paste
Example:
2018-06-02 19:52:16 +02:00
Janne Sinivirta
gcarq
xmatthias
Gerald Lonlas
Raymond Luo
