I use Amibroker to develop systems that are then validated trough a walk forward analysis.
So they're live, constantly adapting to the latest market conditions.
I'm just looking to execute the signals in a crypto exchange 24/7.
The signals could be adapted to the software.
Could Gekko be useful for that?
And if so, how should I send the alerts?
I'm busy adding support to a new exchange (lykke.com) and got more or less everything setup. The market watcher is getting the data and showing me a graph, but when I start a strat runner the Runtime part of the strat running keeps on showing the loading symbol. I tested with Kraken and there the Runtime part show this information quite quickly:
Watching since
2018-05-18 14:38 Received data until
2018-05-18 15:34 Data spanning
56 minutes Amount of trades
0
CC API has some limitations - only last 7 days are available to import, dont provide trades amount and wvp. Be sure does Your strategy dont use trades and wvp values!
node-pre-gyp ERR! Tried to download(404): https://tulipnode.s3.amazonaws.com/tulin...x64.tar.gz
node-pre-gyp ERR! Pre-built binaries not found for tulind@0.8.10 and node@8.11.1 (node-v57 ABI, unknown) (falling back to source compile with node-gyp)
make ✖ ERR Missing Makefile / Bakefile
make ℹ info Run "make init" to generate a Makefile.
+ talib@1.0.2
+ tulind@0.8.10
updated 2 packages in 12.588s
I created a strategy in Tradeview because it's more user friendly. When I run it in Gekko the results are horrible. The data is not the same as in Tradeview.
For example, the Bollinger Bands top & bottom numbers on daily is not the same in Gekko as in Tradeview, why?
Anybody using sar indicators? I'm triying to config talib or Tulip sar indicators to get the same values I get in https://www.tradingview.com/ with the sar indicator with default config SAR (0.02, 0.02, 0.2) but no way to get the same values in gekko.
I'd like to be able to have gekkoga do multiple date-ranges like it does multiple parameters for other parts of the config; i.e. I imagine it would be done like this (and maybe it already can do this and I'm just not doing it right? let me know):
Instead of this:
Code:
daterange: {
from: '2018-02-13 05:54:00',
to: ' 2018-02-25 04:12:00'
},
It might look pointless, but I think sometimes the difference between a bot started at 5:04pm and a bot started at 5:08pm can be huge, and that by backtesting like this, you could determine where the "starting points" are that work well for your strategy for a given data set. It might be that your strategy always does better when started right on the hour, or 11 minutes after the hour, or 16 minutes after the hour, etc. -- I think that it's totally possible that "lining it up" by checking different start times could make a huge difference in performance.
What do you guys think?
Attached photo is two bots with identical config, 11 minute candles, one performing better than the other, seems to have caught the right cycle, while the other one missed the timing, and so took bigger losses. One started just a few minutes before the other.
node-pre-gyp ERR! Tried to download(403): https://mapbox-node-binary.s3.amazonaws.com/sqlite3/v3.1.13/node-v57-linux-arm.tar.gz
node-pre-gyp ERR! Pre-built binaries not found for sqlite3@3.1.13 and node@8.11.2 (node-v57 ABI) (falling back to source compile with node-gyp)
make: Entering directory '/home/pi/gekko/node_modules/sqlite3/build'
ACTION deps_sqlite3_gyp_action_before_build_target_unpack_sqlite_dep Release/obj/gen/sqlite-autoconf-3150000/sqlite3.c
TOUCH Release/obj.target/deps/action_before_build.stamp
CC(target) Release/obj.target/sqlite3/gen/sqlite-autoconf-3150000/sqlite3.o