A trading strategy for prediction markets mediated by the logarithmic market scoring rule (LMSR) market maker
To run the simulation:
python my_bot.py
You can choose between plotting a single simulation and printing
aggregate statistics for multiple simulations by
commenting/uncommenting the appropriate lines in main().
Required Python packages:
- numpy
- matplotlib
In general, the line:
bots.extend(other_bots.get_bots(num_fundamentals, num_technical))
...will create a simulation with 1 + num_fundamentals + num_technical
traders including your bot. You can include multiple copies of your
bot or other bots in the simulation by adding them to the list with
bots.append (for a single bot) or bots.extend (for a list of bots).
other_bots.py includes one type of fundamentals trader and two types
of technical traders (get_bots() splits num_technical evenly between
these two types). The technical traders use the price history only,
and do not make money on average: they simulate noise traders which we
often see in real markets.