Skip to content
A less naive crypto trading bot
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

Files

Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
apis
stenographer
sweet_dreams
.gitignore
README.md
initial_setup.py
notes.md
run.py

README.md

sweet dreams

"If you don't find a way to make money while you sleep, you will work until you die."

  • Warren Buffet

Sweet dreams is a trading bot that allows you to make money even while you sleep. It is the next stage of evolution of polaroid.

Polaroid was the first version of the trading bot. It tried to make markets. It would check the mid-price, and place a buy slightly below and a sell slightly above. So every time a pair of trades are executed it earned a small amount. It only looked to grow the amount of BTC held. While buying, it was very aggressive. Rather than sitting and waiting with nothing, it would follow the price as it rose so that the time with unspent capital was minimum.

While selling, the same cannot be done, as then it would have faced a loss. So instead, it just tracked the price, and if it fell below a certain threshhold, then it would square off and sell according to the new mid-price.

On upswings, it was very profitable, as it would buy in on the small dips and keep making. But since it was chasing price to the top, on down swings, it would keep squaring off, and each square-off was neutralizing 10-15 successful trades.

Sweet Dreams, to fix this issue, now tries to capitalise on the down-swings using the reverse of the original strategy. Rather than first buying low and then selling high, it first sells high and then buys low. The thing it is betting on is small corrections in the direction opposite to the trend. Theoretically, one of the losses that sweet dreams would have are on multiple continous flips. Such a situation should be very rare, and it should be possible to ensure that this doesn't happen using the correct parameters for flipping. Because such fluctuations should be where the most money is to be made.

The other situation where losses would be faced are when there is a sharp movement without any corrections. Since the strategy is designed around exploiting the small corrections in the general trend, if there are two sharp movements consecutively, it could cause a lot of loss. But the likelihood of that seems low.

Technologies used

  1. Beanstalk - Queue
  2. MongoDB
  3. Python3
You can’t perform that action at this time.