Skip to content

coreysan/awakoining

BranchesTags

Repository files navigation

Requirements

  • Install nodejs and yarn latest versions on your Mac
    • Visit the official sites to see how.
  • Install elastic search
    • brew install elasticsearch
  • Install Kibana (Elastic Search visualization)
    • Follow Kibana installation instructions

To get running

  • Copy credentials.example.js to credentials.js and change details to match your API details.
    • KEEP THESE SECRET!
  • Use the terminal and cd to/awakoining
  • Run yarn to install the project dependencies
  • Start elastic search
    • brew services start elasticsearch
  • Start kibana: open /your/path/to/kibana/bin/kibana

Usage

  • yarn run ping to ping the API and verify connectivity
  • yarn run export -s [Symbol A]/[Symbol B] --startTime "2018-01-10 09:30:00" to export the last 500 candlesticks for those symbols from that startTime to now
    • Where [Symbol A] is the 3-digit ticker symbol for the coin you're interested
    • And [Symbol B] is its trading pair
    • Get all candlesticks since January 1st 2017 at 9:30am
    • e.g. yarn run export ETH/BTC to get the candlesticks for ETH as traded with BTC
  • yarn run store -s [Symbol A]/[Symbol B] --startTime "2018-01-10 09:30:00" to populate ElasticSearch with the candlesticks since x time
  • yarn run store-all --startTime "2018-01-10 09:30:00" to populate ElasticSearch with the candlesticks since x time

Project Architecture

  1. Execution begins in index.js, and interprets the command line arguments.
  2. The BinanceApi class is called to process the given command.

The Goal

Identify dormant crypto stocks, and await explosions.

Cryptocurrencies sometimes enter a period of stabilization before exploding upwards on a run:

                           /
-__-_                     /
     --__                /
         -_-___-_-^--_--'

The goal of this project is to scan for coins in a dormant period, and await an upturn.

The algorithm:

Variables

To begin, the following variables are used:

`I    - Interval` is the candlestick interval
`W    - Window` the number of intervals required to recognize a dormancy period
`DP   - Dormancy Period` the period over which dormancy is confirmed
`V    - Variance` the variance, over the dormancy period, expressed as a percentage
`Va   - Variance Average` the average variance of the `DP`
`MaxV - Max Variance` the variance maximum. The highest sell price during the `DP`
`MinV - Min Variance` the variance minimum. The lowest sell price during the `DP`
`BT   - Buy Threshold` the buy threshold
`S    - Stabilization period` the last % of the `W` during which the sales should have stabilized.
                              Note that this may be turned off to nullify its effects.

The following states exist in the script:

Analysis

The algorithm begins here:

  1. Get the candlesticks over some period, and scan each window for a dormancy period DP. Dormancy period is defined as follows, relative to the current price:
    • No swings of greater than V (5%?) within the DP
    • The V in last ~10% of the DP is less than the average V throughout (The line-of-best-fit may be of either slope up or down, no matter)

Dormancy

When a DP is found, the coin goes into watch mode. and MaxV and MinV are recorded.

  1. If a dip below MinV is encountered, the DP is rejected and forgotten.
  2. If a surge above MaxV is encountered, the algorithm changes states and enters its buy state

Buy and Protect

Buy

Pz - Price zero is the price we have launched from, also known as the high of the bullish candlestick-zero Cz SL - Stop Limit is the trailing stop limit. Q - Quantity, quantity Given some Cz that breaks the MaxV, we place a buy order for some Q purchased with Ethereum.

For example, suppose a Cz breaks out of its DP and establishes a price of .100 ETH. We buy at this price Pz.

  • Set a sell limit at Pz - 1% with order price Pz - 1.5% for the same Q we purchased with.
    • If the next candlestick is bearish, we may initiate the sell order immediately and lose 1.5%.
    • If the next candlestick is bullish, we will hold and ride the rise, entering the next state.

Protect

Assuming our stock is rising, we enter the Protect State. Here, we seek to preserve some of the gains acquired during this upturn.

  • With the first trend-breaking candlestick, Cz, we purchase at price Pz and set sell limits as outlined above.
  • Let's take the following concrete numbers:
    • Pz = 0.100 ETH
    • SLs - Stop Limit stop = 0.990 ETH
    • SLl - Stop Limit limit = 0.985 ETH
    • SLd - Stop Limit difference = SLs - SLl = 0.005 ETH, recorded and reused throughout
  • Suppose the next candlestick is bullish and increases by a Delta D of 0.01 ETH. Here use D to set new a new gain cieling C and reset our stops:
    • Pz = 0.100 ETH
    • SLt - Stop limit threshold = The threshold at which to sell the gains - set at 30%
    • Pc - Price, current = Gprevious + D
    • C - Ceiling = Max(Gprevious, Pc) = 0.1 ETH + 0.01 Eth = 0.11 ETH
    • Pz = 0.100 ETH, this is fixed until we sell.
    • SLs - Stop Limit stop = ( [net gain](C - Pz) * SLt ) + Pz = (0.01 * 30%) + 0.1 = 0.103 ETH
    • SLl - Stop Limit limit = SLs - SLd = 0.103 - 0.005 = 0.908 ETH
    • This step continues until we hit the stop limit.
    • We may consider allowing a sliding SLt threshold to preserve 50% of large gains.. The 30% threshold is intentionally low to prevent early exits.

Completion

The algorithm completes under one of the following conditions:

  • The script it terminated manually, leaving the stop orders in place.
  • The script hits a stop limit sells the stock for Ps and prints the results of the transaction.

Visual Interface

A simple command-line interface should suffice to:

  • Show Coins in any of the above states
  • Show variables pertaining to each of the states.

Example interface:

1000 coins currently being analyzed each minute...

Dormancy:
ETH 1000 USD +/- 4.3% over I=30 - Last interval: -1.5

Protect:
40  ETH Pz: $1000.00 @ $1051    Sells at $1015
180 XLM PZ: $0.7515  @ $0.8131  Sells at $0.775

Sold:
40 BTC  @ Pz $17015 for Ps $17055  Total $1600    at 3:30pm January 15th
320 VEN @ Pz $4.30  for Ps $4.75   Total $144     at 3:30pm January 15th

Requirements:

Using Elasic Search (ES)

Note that you interact with ES through Kibana (see below)

Using Kibana

Use the kibana console for most interactions

# Create and list candlestick index
PUT /candlestick?pretty
GET /_cat/indices?v

# Delete the candlestick index, show it's deleted, then put it back
DELETE /candlestick?pretty
GET /_cat/indices?v
PUT /candlestick?pretty

# Put a doc in the index
PUT /candlestick/doc/1?pretty
{
  "openTime": 1515260700000,
  "openPrice": "0.00716500",
  "highPrice": "0.00716500",
  "lowPrice": "0.00716500",
  "closePrice": "0.00716500",
  "volume": "44.00000000",
  "closeTime": 1515260759999,
  "quoteAssetVolume": "0.31526000",
  "numberOfTrades": 1,
  "takerBuyBaseAssetVolume": "0.00000000",
  "takerBuyQuoteAssetVolume": "0.00000000",
  "ignored": "0"
}

# Get that newly created doc:
GET /candlestick/doc/1?pretty

response:
{
  "_index": "candlestick",
  "_type": "doc",
  "_id": "1",
  "_version": 1,
  "found": true,
  "_source": {
    "openTime": 1515260700000,
    "openPrice": "0.00716500",
    "highPrice": "0.00716500",
    "lowPrice": "0.00716500",
    "closePrice": "0.00716500",
    "volume": "44.00000000",
    "closeTime": 1515260759999,
    "quoteAssetVolume": "0.31526000",
    "numberOfTrades": 1,
    "takerBuyBaseAssetVolume": "0.00000000",
    "takerBuyQuoteAssetVolume": "0.00000000",
    "ignored": "0"
  }
}

Bulk insert a couple:

POST /candlestick/doc/_bulk
{ "index" : { "_index" : "candlestick" } }
{ "openTime": 1515260700003 }
{ "index" : { "_index" : "candlestick"} }
{ "openTime": 1515260700004 }

To get results/search:

GET /candlestick/_search

To calculate a moving average (broken rolling window)

GET candlestick/_search?size=0
{
  "sort": [
      { "openTime": "asc" }
  ],
  "aggs": {
    "histogram_by_date": {
      "date_histogram": {
        "field":     "openTime",
        "interval":  "2m"
      },
      "aggs":{
        "the_sum":{
          "sum":{ "field": "volume" }
        },
        "the_movavg":{
          "moving_avg":{
            "buckets_path": "the_sum",
            "window": 3
          }
        }
      }
    }
  }
}

Search for candlesticks of a certain symbol/interval

POST /candlestick/_search
{
  "query": {
    "bool": {
      "must": [
        { "term": { "symbols.keyword": "NEOETH"}},
        { "term": { "interval": "15m" }}
      ]
    }
  }
}

Read more about in the ES docs!

Troubleshooting ES/Kibana Errors

If you get an error like this:

({
  "type"=>"cluster_block_exception",
  "reason"=>"blocked by: [FORBIDDEN/12/index read-only / allow delete (api)
])

Then you'll have to run this in the kibana console change some settings:

PUT candlestick/_settings
{
  "index": {
    "blocks": {
      "read_only_allow_delete": "false"
    }
  }
}

Changing Config Settings

To find where your config settings live:

$ brew info elasticsearch
elasticsearch: stable 6.1.1, HEAD
Distributed search & analytics engine
https://www.elastic.co/products/elasticsearch
/usr/local/Cellar/elasticsearch/5.5.2 (103 files, 35.6MB)
  Built from source on 2017-09-08 at 23:48:40
/usr/local/Cellar/elasticsearch/6.1.1 (107 files, 30.2MB) *
  Built from source on 2018-01-07 at 22:03:51
From: https://github.com/Homebrew/homebrew-core/blob/master/Formula/elasticsearch.rb
==> Requirements
Required: java = 1.8 ✔
==> Caveats
Data:    /usr/local/var/lib/elasticsearch/elasticsearch_coreysanford/
Logs:    /usr/local/var/log/elasticsearch/elasticsearch_coreysanford.log
Plugins: /usr/local/var/elasticsearch/plugins/
Config:  /usr/local/etc/elasticsearch/

The last line above points to /user/local/etc/elasticsearch/ where you'll find an elasticsearch.yml file.

To Enable CORS

Add this to the config you found above:

# ENABLE CORS
http.cors.enabled : true
http.cors.allow-origin : "*"
http.cors.allow-methods : OPTIONS, HEAD, GET, POST, PUT, DELETE
http.cors.allow-headers : X-Requested-With,X-Auth-Token,Content-Type, Content-Length

About

Predict crypto launches

Resources

Readme
Activity

Stars

3 stars

Watchers

1 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages