35c51c73f7
Merge pull request #518 from gcarq/cleaning_up_backtesting
...
Cleaning up backtesting/hyperopt
2018-02-18 10:18:00 +02:00
fac122891f
remove stoploss parameter from backtest, it is loaded from strategy
2018-02-17 11:14:03 +02:00
d1bdbcd273
Fix wrong duration calculation in hyperopting
2018-02-16 22:08:20 +02:00
bf72b5bc37
make args available for optimizer and use them instead of guessing from params
2018-02-16 14:00:12 +02:00
e3d222912d
Merge pull request #511 from gcarq/hyperopt_selectable_spaces
...
Allow selecting Hyperopt search space
2018-02-12 08:28:24 +02:00
3e07d41fa9
remove mention of sell space
2018-02-12 07:01:51 +02:00
2ce03ab1b5
make Strategy store roi and stoploss values as numbers to avoid later casting
2018-02-11 15:25:30 +02:00
f14d6249e0
allow selecting hyperopt searchspace
2018-02-09 20:59:06 +02:00
8c7b29734e
use date info to calculate trade durations
2018-02-06 11:34:00 +02:00
321e3ede30
Fix hyperopt ticker interval download
2018-01-29 22:53:28 -08:00
a6a479f7aa
balances to min roi hyperopt settings
2018-01-28 10:46:22 +02:00
67ddb2e7f8
lower precision for most search space variables
2018-01-27 09:51:06 +02:00
f33923c784
fix typings for hyperopt code
2018-01-26 18:32:45 +02:00
4fe6ae0bae
fix search space for min ROI
2018-01-25 22:32:46 +02:00
42087c9bfe
let hyperopt optimize ROI table
2018-01-25 11:12:00 +02:00
5007165908
add search space for ROI table
2018-01-25 09:34:26 +02:00
c400d15ed1
rip out hyperopt things from strategy, add indicator populating to hyperopt
2018-01-23 16:56:12 +02:00
1c7da95fed
Move hyperopt_trials.pickle to user_data/
2018-01-22 20:51:39 -08:00
baae374899
Move hyperopt_conf.py into user_data/
2018-01-22 20:51:39 -08:00
be75522507
Fix flake8
2018-01-22 20:51:39 -08:00
c46d78b4b9
Decouple strategy from analyse.py
2018-01-22 20:51:39 -08:00
4a9e1cb345
Merge branch 'develop' into backtest-export
2018-01-19 07:02:38 +01:00
14d16f2574
Merge pull request #357 from kryofly/timeperiod
...
Timeperiod
2018-01-18 20:26:44 -08:00
c670ccfd37
add trigger +DI crossed above -DI
2018-01-16 18:52:06 +02:00
8896b39231
add heikenashi reversal bullish trigger to hyperopt
2018-01-16 18:52:06 +02:00
ce963aae58
add macd < 0 guard to hyperopt
2018-01-16 18:52:06 +02:00
dc01807b3c
switch ema5 trigger to ema3 cross trigger
2018-01-16 18:52:06 +02:00
fadac5fe4a
remove too aggressive trigger
2018-01-16 18:52:06 +02:00
99260735ae
remove broken bbands trigger from hyperopt. add two working bbands triggers
2018-01-16 18:52:06 +02:00
501be8a3bc
adjust the hyperopt objective function to emphasize profit and allow more variation in trade counts
2018-01-16 16:36:50 +02:00
38fe7ec7cd
adjust default target values for hyperopt
2018-01-16 16:35:48 +02:00
0e58ab7e01
more advanced use of --timerange
2018-01-16 00:15:49 +01:00
71bb348698
rename --timeperiod to --timerange
2018-01-15 21:49:06 +01:00
ec7bfba8df
add comment about checking the new total profit logging
2018-01-14 13:11:19 +02:00
f1e176d35c
log total profit in percentages also
2018-01-14 13:10:25 +02:00
92241baade
log the loss value
2018-01-14 13:09:39 +02:00
153e11f045
Merge branch 'develop' into timeperiod
2018-01-11 19:45:47 +01:00
4781a23809
Merge branch 'develop' into backtest-export
2018-01-11 19:40:42 +01:00
27769f0301
uncomplex backtest
2018-01-11 17:45:41 +01:00
86db6c9084
sort imports
2018-01-11 07:08:56 +02:00
1b6b0ad9d2
autopep8
2018-01-11 06:50:36 +02:00
b0f3fd7ffb
timeperiod argument to backtesting and hyperopt
2018-01-10 23:48:59 +01:00
feca87345f
refactor
2018-01-10 23:00:40 +01:00
e67c652988
use os.path.join, fix docstrings
2018-01-10 11:50:00 +02:00
ffae0b2cd5
hyperopt: prettyfie best values when receiving SIGINT, use the global TRIALS
2018-01-09 12:37:56 +02:00
1647e7a0c1
update fix failing tests, unitest that resume hyperopt functionality works
2018-01-09 12:26:52 +02:00
b35fa4c9f6
hyperopt: show the best results so far
2018-01-09 12:25:58 +02:00
a48840509b
Hyperopt: use results from previous runs
2018-01-09 12:25:58 +02:00
ca8cab0ce9
Hyperopt to handle SIGINT by saving/reading the trials file
2018-01-09 12:25:58 +02:00
dd2ccea6e5
fix wrong range in stoploss search space
2018-01-08 21:59:46 +02:00
9dd38aebe0
add stoploss to the hyperopt parameters
2018-01-07 21:08:12 -08:00
9c21077dc1
Fix hypeopt issue when no result found
2018-01-07 17:53:21 -08:00
890083ce7f
Merge branch 'develop' into datadir
2018-01-07 10:00:35 +01:00
b3ea0f4ec5
Make readable hyperopt best parameters result
2018-01-06 17:19:48 -08:00
60ed4b9d1e
--datadir <path> argument
...
This argument enables usage of different backtesting directories.
Useful if one wants compare backtesting performance over time.
2018-01-06 23:24:35 +01:00
a36fd00f6a
also print dot when hyperopt eval result is fail
2017-12-28 06:40:11 +02:00
7f44ba6df4
unit tests for optimize.hyperopt
2017-12-28 06:39:56 +02:00
7b0beb0afa
cleanups
2017-12-28 06:36:18 +02:00
de33d69eed
Lint fixes ( #236 )
...
* correct docstring
* add type annotation to trade_count_lock
* fix indentations
* allow globals in hyperopt.py
* fix import order
* simplify asserts
* use proper variable name
* simplify condition
* fix path operation that fails on windows
2017-12-25 12:07:50 +01:00
9959d53f5e
Logging improvements to Hyperopt ( #235 )
...
* make log texts go on new line
* remove unnecessary fields from hyperopt log messages
* shorten log text in hyperopt
* consider making zero trades a failed hyperopt eval
* only log from hyperopt when result improves
* remove unnecessary temp variables
* remove unused result data variables
* remove unused import
* fix an outdated comment
2017-12-25 08:18:34 +01:00
353b0d2d34
balance hyperopt objective to adjusted profit calculations
2017-12-23 19:18:28 +02:00
e644d57dbe
log should state profit is in BTC to avoid confusion
2017-12-23 19:00:49 +02:00
50e7cef5f3
remove commented-out code
2017-12-23 19:00:49 +02:00
1058820e1b
just pass stake_amount instead of the whole config
2017-12-23 19:00:49 +02:00
24bc3a8390
show more digits for profits
2017-12-23 15:11:19 +02:00
5309ea3820
use newline for each log result for readability
2017-12-23 15:11:19 +02:00
a063680d32
calculate log line only if really logging
2017-12-23 15:11:19 +02:00
10cf2ce853
remove unnecessary confusing division
2017-12-23 15:11:19 +02:00
871357a2e3
just require positive results
2017-12-23 15:11:19 +02:00
8d93363655
filter nan values from total_profit and avg_profit
2017-12-23 09:21:04 +02:00
f300af0fe2
Merge pull request #200 from glonlas/fix_fees_calculation
...
Fix the fee calculation
2017-12-22 13:55:02 +02:00
41e22657e4
Fix hyperopt when using MongoDB
2017-12-21 19:20:47 -08:00
d258118b0a
Fix the fee calculation, backtesting, and hyperopt fee calculation and avg_profit
2017-12-20 20:18:41 -08:00
c8fb6c4661
More lint fixes ( #198 )
...
* autopep fixes
* remove unused imports
* fix plot_dataframe.py lint warnings
* make pep8 error fails the build
* two more line breakings
* matplotlib.use() must be called before pyplot import
2017-12-18 17:36:00 +01:00
80ef2cfed4
Merge pull request #193 from gcarq/feature/ci-enforce-pep8
...
CI: enforce PEP8 conform code
2017-12-17 07:42:23 +02:00
512fcdbcb1
Allow user to update testdata files with parameter --refresh-pairs-cached ( #174 )
2017-12-16 15:42:28 +01:00
95fe0f4dec
fix pep8 warnings
2017-12-16 03:39:47 +01:00
a7cca4985e
omit hyperopt output if total_profit doesn't go pass threashold (3)
2017-12-02 01:32:23 +02:00
965c075362
disable info logging on hyperopt.tpe
2017-12-02 00:21:46 +02:00
0c35e6ad19
minor changes
2017-11-25 03:28:52 +01:00
e27a6a7a91
add mongodb support for hyperopt parallelization
2017-11-25 02:04:37 +01:00
5bf583cba4
remove unused imports
2017-11-25 01:23:18 +01:00
a23fce519d
pretty print hyperopt results
2017-11-25 01:22:36 +01:00
9ff1f05e66
add --epochs to hyperopt subcommand
2017-11-25 01:12:44 +01:00
b9c4eafd96
integrate hyperopt and implement subcommand
2017-11-25 01:04:11 +01:00
7fa5846c6b
move hyperopt to freqtrade.optimize.hyperopt
2017-11-25 00:30:39 +01:00
Samuel Husso
Janne Sinivirta
Gerald Lonlas
kryofly
gcarq