Skip to content
Multi-bucket range trading for the GDAX exchange
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
README.md
account.js
buckets.js
credentials.json
db.js
gdax.js
multipong-eth-1.PNG
multipong-eth-2.PNG
multipong.js
package.json
settings.json
trades.js
ui.js

README.md

gdax-multipong

This is a tradebot for BTC, ETH and LTC written in NodeJS for GDAX. It is designed for flat or oscillating markets.

You will need

Setup

  1. clone git@github.com:msolters/gdax-multipong && cd gdax-multipong
  2. npm install
  3. Put your GDAX API passphrase, key, and secret into credentials.json
  4. Set your currency and bucket parameters in settings.json (more on configuration below)
  5. node multipong.js

To begin buying and selling, press the "b" and "s" keys.

Commands

  • r - Reload and apply settings.json
  • b - Toggle buying on and off
  • s - Toggle selling on and off
  • c - Cancel all pending buys. Note, if buying is "on", these orders will be almost instantly-replaced. Turn off buying before canceling any orders!
  • q, ESC, or CTRL+C - Exit Multipong

Strategy

Multipong uses a bucket strategy to perform high-frequency range trading between a minimum and maximum currency price in many smaller sub-ranges.

Here's an example with Ethereum. In this example, buckets are a little less than $0.10USD wide. Multipong will place a few buy orders, always less than the current mid-market price. Each buy order is one bucket-width apart. As soon as one of these buy orders is filled, Multipong will automatically place a sell order one bucket-width higher. Thus every filled buy will yield exactly one sell, and always at a profit:

As the price of Ethereum falls, more buy orders are placed and filled. As the price bounces back up, the sell orders are filled, yielding a profit; and the buys are then re-placed.

As the price of a cryptocurrency oscillates throughout the day, many small profitable buy/sell pairs will be executed. Especially for oscillating markets, this can produce hundreds or thousands of trades a day, all of them favorable.

Each trade will be trade_size big; in the images above, the trade_size was 0.1ETH. Clearly, deeper trade sizes will yield more profit per trade, but will drain your fiat capital faster as each trade is more expensive.

Wider buckets can yield a larger profit per trade as well, as the buy-sell spread is larger, but, the price would have to vary much more to make each trade happen, and are only good for more volatile market conditions. This approach can also be useful if the price is varying a great deal during the day, and you don't want to spend all your fiat in a small price range.

The bucket strategy is helpful in case the currency price falls after buying some crypto. Say the price of BTC is 5995USD and we buy 0.01BTC; while we wait for the price to hit 6000USD and make a sale, what if the price falls again? Since each bucket can only buy once, we are guaranteed to still have buckets at lower prices that will continue to trade, making money, while our earlier gamble is waiting to be realized. This allows us to keep making money across a wide price range instead of putting all our eggs into one basket.

Multipong performs best with thin (price width), deep (trade size) buckets, and works best in mostly-flat or oscillating markets. Note, you may win or lose. Multipong is provided "as-is," with no guarantees or promises! Multipong should never place an unfavorable trade, but it could place a favorable trade that never happens. As with all things crypto, be safe!

Configuration

Here's an example settings.json:

{
  "tz": "America/New_York",
  "multipong": {
    "coin": "BTC",
    "fiat": "USD",
    "initial_cash": 100,
    "greedy": true,
    "num_buckets": 10,
    "trade_size": 0.01,
    "min_price": 7000,
    "max_price": 7400,
    "bucket_runway": 3,
    "midmarket_price_period": 1,
    "trade_period": 50
  }
}
  • Multipong will trade between fiat currency for coin; in this example, it will buy and sell BTC with USD. Available options are only what GDAX provides: BTC, ETH, LTC, USD, GBP, EUR.
  • trade_size is the quantity of coin that the bot will buy or sell per bucket. Keep in mind GDAX's minimums: 0.01 for BTC.
  • Buy and sell BTC if it's between min_price and max_price (+/- 0.01 [fiat]). In other words, trading will stop if the price goes above max_price or below min_price.
  • num_buckets is how many trade ranges the total min-max price range will be subdivided into. A higher number here means more frequent trades, but less profitable ones. (It also increases liability for fees if the market has sudden shifts.) If buckets are only a few cents or dollars, a quickly moving price might move faster than trades can be requested (esp. depending on your internet connection) and some orders might end up being taker orders!
  • bucket_runway is how many buy orders to the left of the midmarket price the app will maintain at any given time. A value of 3 means that, at most, 3 buy orders will be placed at any given time. If the price moves up, the lowest buy order is deleted and a new, higher one is placed. This is not true as the price falls (or else how would buys be filled?)
  • initial_cash is the amount of cash you are telling Multipong is in your fiat wallet to use.
  • greedy determines whether or not initial_cash will be treated as your total fiat capital. If greedy: true, Multipong will make buys as long as your fiat wallet has enough funds, even if those funds are more than initial_cash. Conversely, if greedy: false, Multipong will stop placing buys when it has exhausted initial_cash, even if there are more funds in your fiat wallet.
  • midmarket_price_period how many ms between updating the current crypto midmarket price in the app
  • trade_period how many ms between placing new trades

You are encouraged to explore various permutations of min_price, max_price, trade_size and num_buckets. To avoid getting stuck out with high buys during large runs, we advise you to set a max_price below the expected maximum for a given trading period. You want min_price and max_price to wrap around the range where the price is expected to see the most volume of trading!

Updating Buckets/Resetting

As the market moves, you may find you want to adjust your min_price, max_price, num_buckets, etc. without stopping Multipong. Simply edit your settings.json and save the file. Then, in the app, press the r key. Multipong will re-read the values of settings.json, and update the components of the app which require it.

If you are only changing your bucket distribution, ongoing trades will not be touched, but new trades moving forward will respect your new buckets. However, if you are changing the currency pair that you are trading, Multipong will have to undergo a slightly longer re-initialization process.

Either way, depending on your internet connection, this step should never take longer than a few seconds! Happy trading!

Thank You!

This software is provided free for your amusement, edification, and enrichment! However, if you enjoy multipong, you can show your appreciation by sending a BTC tip to the author here at 1Nhdd9UCsv9dsabLLNRue8ecDN41yrSRdk

If you have any bugs, try to get a screen capture and/or the contents of your <coin>-<fiat>.log file! (Don't share any sensitive information in your uploads though!)

You can’t perform that action at this time.