Topic: Developing Market Maker Robot (Read 147 times)

Greetings to BitcoinTalk Community.

I am working on “market maker” robot and want to ask about your experience and share mine.
MM (in my understanding) is a trading robot that should not rely on the long-term price direction and earn from the price irregularities or jumps, in other words - provide high-frequency scalping trades.
What I am using: Self-made bot for Poloniex and Bittrex (written in Erlang)

Strategy:
1.   Place buy-limit order at a price smaller than last ask price by 0.1%
2.   As soon as buy order executed, place sell-limit order at a price higher than buy price by 0.5% + both order’s commissions (or in other way: (buy cost + target percent) / (received size - commission percent) to take actual filled size into account)
The idea is to earn as frequently as possible, with fixed percent + commission taken into account.

Pairs: LTC/BTC, XRP/BTC, XLM/BTC (STR/BTC for Poloniex)
Results: Getting about 10-20 trades a day during volatile periods. The best results to this point showed XRP/BTC pair, seems to be more volatile.
10 – 20 trades a day, make 1% profit per day on average.

Problems:
0.1% target was too big for the LTC/BTC pair recently. The coin is just going down, and even order with such small target could hang for a couple of hours or even leave hanging while the price is going down.
Tried to lower the target, as low as 0.01%, still getting similar results.

Plans: Scanning orderbook.
Idea #1: Post orders only close to prices with relatively high volume
Idea #2: Analyze Buy and Sell order summary weight and post trades only when the average weighted price is higher than the last one.
 
So, what your opinion on this approach?
Maybe you know some good MM bots to take a look at?