High Frequency Trading Framework and Logics for Crypto Traders
NDB Technologies
NDB-Tech steer its customers to provide high frequency trading framework and strategy/logic for cryptocurrency markets.
Prologue:
We provide a framework that can run on BTC automating trading for Bitmex, and Liquid exchanges. Also, we aimed to mitigate redundent workload to develop trading infrastructure. Our framework is robust and highly sustainable that’s mean participations can be focus only trading logic.
NDB framework has been developing over 2 years and we’re still updating constantly. The simple logic comes with 20 type of different index and 2 types of MM (market-making) logic. It’s generally hard to obtain winnig trades through simple approach but using those provides an additional idea and using one of those traders discreation needed.
Features:
Whereby our framework traders can utilize 3 different mode for trading events.
Event-driven HFT mode (based on execution)
Time-driven Scalping mode (based on unit)
Normal mode (time axis of 1 min. or more)
Our framework has quite extensive functions and especially for Bitmex functions are substaintial.
Supporting Exchanges:
We support Bitmex and Liquid exchanges however, we planned to increase number of exchanges in order to run on our framework.
Servers:
AWS Cloud9 is recommended by our side however, for ultra-low latency demands please contact us.
Bot Functions and Features:
It is possible to develop without realizing API logic of exchanges.
NDB framework absorption care of, various restriction (API number of times and error, etc.)
Can be simultaneously run on multiple logic under-one server environment.
Candlestick (bar value) is 1 minute to correspond to the time axis of at least one day in
the high-speed scan corresponding to the bot using the BitMEX for
timing can not promise plan the LiquidTap corresponding
commitments to friendly. BitMEX Supports event-driven bots from history.
Supports logic using multi-timeframes (multiple time axes).
Basic parameters are configured and operation can be customized.
Changes in various parameters are immediately reflected during BOT operation’s possible behavior change.
The-situation analysis and easy to profit and loss situation periodically notified on Twitter.
Back testing program, and a data collection program is included.
in the Relay function and each other can communicate bidirectionally (see below), genetic algorithm optimal parameter search comes with backtester.
Following Logic:
Existing and written currently 20 kinds logic (described in detail later)
- Extended functions ( shall be sold separately).
Logic forward test (air training) function for BitMEX Judges the execution
history in real time and simulates the actual bullet result with high accuracy. Logic verification is possible without consuming the actual product.
It is difficult to explain the relay function and its value because we have not heard of BOTs that have the same function at present, but we will extract a part of the manual published in the paid part.
Conceptual diagram is as shown:
The role of the relay function is the following three points. By using this function, you can develop advanced logic that cannot be realized by other systems.
1- A function to communicate between multiple activated NDB bots.
2- A function to distribute candlesticks to each logic from RelayServer.
3- A function to operate more than 10 parallel air tray functions at the same time.
Necessary knowledge and target users:
Target layer:
NDB framework and logics does not guarantee the profit just by operating the attached strategies. Unlike the popular bot sales of informational products, NDB is a tool that helps you to realize your ideas.
If you are expecting to make a full profit just by participating or purchasing please do not utilize our services.
Necessary knowledge:
BitMEX is an exchange that is much more difficult to obtain a profit when we compared to domestic exchanges such as bitFlyer and so on due to fees and minimal spread. On the other hand, the trading volume is much bigger and you will get a big return once you obtained it.
In such a situation, I summarized the necessary conditions to use NDB bot. The conditions described above are more important.
Continuity with daily profit and loss, turn to PDCA of log analysis & logic improvement — Indomitable spread that never gives up (BitMEX is really difficult)
- Have a logic ideas (knowledge of indicators, knowledge of market distortion) Etc.)
About support:
We are providing support under few conditions only.
- Bug fixes
- For sale features
About ‘’Bug’’ fixing please follow-up our Twitter account.
At a glance of NDB bot configurations:
The composition of programs and documents at the time of writing is as follows. You need above-average programming skills to understand the full framework.
-
NDB bot provides the following logic development templates. By adding an indicator, calculation and buying / selling logic in the vicinity of “# Add processing here”, the style of developing the logic will be a style that will greatly reduce the man-hours compared to the case of developing from scratch.
- Although the image is small and hard to see, the SMA logic will be completed with the following changes compared to the template (the SMA code on the left and the template on the right).
Introduction of attached sample logic(s):
Introducing the sample logic that comes with NDB bot. For your reference, here are the results of backtesting on testable logic.
Comes with sample logic is basically because there are a lot of those tuning in at that time logic development, performance comes in the fiscal 2018 rate you many things but, dare at the time of the parameter 2019/01 / 01–2019 / 11/04 period It is also true that there are many logics of profit and loss because the back test is performed with the data of. This is described without tuning the parameters in order to help you understand that PDCA is indispensable .
Don’t be afraid to misunderstand, the parameters are a thing of the past and the results of the backtests described below are only for a certain period of time and do not indicate the superiority of the logic itself. The performance changes if parameters are firmly adjusted to the market, and it is also true that some parameters do not match the market at that time. In that sense, I would like to use sample logic as a programming material to create your own logic.
NDB bot graphs are also very important in backtesting and production. Understand what kind of position and lot you will have in what market and use it as the basis for your own logic improvement. In the following logic introduction, graphs are not shown due to the appearance of the paper, but for example, the following back test graph is output in MFI logic. Similar graphs are periodically notified to Twitter etc.
Random.py:
Timeframe : 60sec
Description: Random trading logic for confirming normal operation after installation.
This logic is provided in BitMEX version and Liquid version.
Profit and loss:
Channel_breakout.py:
Timeframe: 3600sec
Description: Learning logic that performs maintenance based on the channel breakout index
This logic is provided in BitMEX version and Liquid version
Profit and loss:
-
Mfi.py:
Timeframe : 1800sec
Description: Logic for learning that is detonated on the upper and lower limit lines based on the MFI index.
Profit and loss:
Sma.py:
Timeframe : 3600sec
Description: Logic for learning that crosses over based on two SMA indicators
This logic is provided in BitMEX and Liquid versions
Profit and loss:
Vix_rci.py:
Timeframe : 3600sec
Description: Learning logic to buy and sell based on the ViXFix indicator that became popular in 2018.
Profit and loss:
Piano.Py:
Timeframe: 1800Sec
Description: two SMA / EMA on the basis of the index logic to Doten in cross
This logic is BitMEX version has Liquid version is provided.
Profit and loss:
Crocodile.py:
Timeframe: 1800sec
Description: Logic to contradict the upper and lower limit line based on RSI indicator ・
Profit and loss:
frog.py:
Timeframe: 7200sec
Description: Logic based on Bollinger band index.
Profit and loss:
Delta.py:
Timeframe: 86400sec
Description: Logic based on index calculated from fluctuation range of daily chart.
Profit and loss:
Heikin_kun.py:
Timeframe: 3600sec
Description: Logic based on index calculated from fluctuation range of average bar.
Profit and loss:
New_genbu.py:
Timeframe: 3600sec
Description: Logic cross-based on two WMA indicators.
Profit and loss:
New_seiryu.py:
Timeframe: 900sec
Description: ATR, RSI, mustache based on the indicator for instantaneous mustache logic.
Profit and loss:
Shenlong.py:
Timeframe: 10800sec
Description : Elaboration of contrarian logic based on Granville’s law.
Profit and loss:
MA_diff_cross.py:
Timeframe: 14400sec
Description: Logic based on short- and long-term indicators obtained from linear regression analysis.
Profit and loss:
Perfect_order.py:
Timeframe: 14400sec
Description: Logic based on the perfect order method that is major in the exchange area.
Profit and loss:
Swallow.py:
Timeframe: 14400sec
Explanation: Logic based on high and low and linear regression analysis index.
Profit and loss:
Trend_surfing.py:
Timeframe: 3600sec
Description: Logic to follow trend from linear regression analysis.
Profit and loss:
This logic: which was the best on the back test, also includes a graph. We have been able to achieve stable results throughout the year.
trend_surfing2_py:
Time bar: 3600sec
Description: Positioning logic of another version of trend_surfing
Profit and loss:
doten.py:
Timeframe: 3600sec
Description: True channel break logic rewritten using Stop order.
Profit and loss:
Other notes:
A cryptocurrencies are a product that has the risk of sudden price fluctuations and may cause losses to users due to changes in the supply and demand balance and market trends. We are not responsible for any problems or loss that may occur with the software.
Social Channel:
Twitter: @NDB_Tech
