dyllan-to-you / hydra Goto Github PK

• Preprocess price graphs for each indicator
• pick arbitrary start time to simulate passage of time & buy/sell trades
• Enact strategy for buy/sell based on past price graph
• Basic Strategy to start: Finding best performing indicator in the past X minutes (e.g. 60) from indicator that have at least Y buy/sell signals during that time period

Details: 1;
Count # entries/exits in range to constrain combinations
Determine number of possible combinations to figure out whether the set of potential combinations need to be constrained further
Count # buy/sells for each indicator parameter combination to further constrain based on # of cycles
Determine %increase or decrease from each buy/sell combination
Compare the multiplication of the %I/D numbers
Set indicator to trade using best indicator parameters based on last line
Move range up one timepoint, throw out %I/D where buy times out (store buy time with %I/D), check for new buy/sell combinations and recalculate to compare again

• figure out how to classify the environments
• figure out how to tie successful indicators to relevant environments

Ideas:

• detrending
• fft/frequency
• volatility measures (std deviation)
• avg velocity/acceleration of price
• other indicators (rsi, psar, sma/ema)