avarice

(noun) extreme greed for wealth or material gain.

Greed captures the essence of the evolutionary spirit. -Gordon Gekko

Disclaimer

Please understand what you are doing before attempting to live trade at all. Any money you may lose is your own responsibility.
Please see LICENSE.txt for license info

Notes

This software is expected to run continuously to get valid data. we do however support resuming if the downtime is less than the configured CandleSize.
All testing is done on current Python3.5.

Contributing

Check docs/Contributing.md for details.

A few features

Can utilize a list of multiple indicators to trade off for more secure trade decisions.
Can easily customize the indicators used.
Many indicators support multiple trade strategies.
Simulation mode allows simulating trades without live trading, or requiring an exchange account.
Live trading mode at market prices.
Interactive graphs may be created in real time for selected trade indicators and price. Mouseover Examples: price, KDJ, MACD
Consistently pursuing a cleaner base to allow easier community involvement.

Indicators

SMA (Simple Movement Average)

NOTE: We support two SMA trade strategies specified in genconfig.SMA.IndicatorStrategy; "CD" (convergence/divergence), and "Diff" (waits to pass up or down diff threshold before trend is determined)
Does two calculations off SMAShortPeriod and SMALongPeriod
SMA = (Sum of last SMAPeriod candles) / SMAPeriod
SMADiff = 100 * (shortSMA - longSMA) / ((shortSMA + longSMA) / 2)

EMA (Exponential Movement Average)

NOTE: We support two EMA trade strategies specified in genconfig.EMA.IndicatorStrategy; "CD" (convergence/divergence), and "Diff" (waits to pass up or down diff threshold before trend is determined
We trade this as a crossover (convergence/divergence) indicator. This is one of our supported EMA trade strategies. On 10/21, when EMA10 < EMA21, we sell (and visa versa). Differs from MACD due to lack of third signal line.
NOTE: does two calculations using EMAShort and EMALong.
The first iteration uses SMA to generate the first EMA.
Multiplier = (2 / EMAPeriod) + 1
EMA = ((Current Close - Previous EMA) * Multiplier) + Previous EMA
EMADiff = 100 * (shortEMA - longEMA) / ((shortEMA + longEMA) / 2)

DEMA (Double Exponential Movement Average)

NOTE: We support two DEMA trade strategies specified in genconfig.DEMA.IndicatorStrategy; "CD" (convergence/divergence), and "Diff" (waits to pass up or down diff threshold before trend is determined
Similar points as the "EMA" section above, however with more of a weight on the last EMA (LOWER LATENCY THAN EMA).
DEMA = 2 * EMA – EMA(EMA)
DEMADiff = 100 * (shortDEMA - longDEMA) / ((shortDEMA + longDEMA) / 2)

FRAMA (Fractal Adaptive Moving Average)

NOTE: We support two FRAMA trade strategies specified in genconfig.FRAMA.IndicatorStrategy; "CD" (convergence/divergence), and "Diff" (waits to pass up or down diff threshold before trend is determined
N = (highest price - lowest price) / period ; split into 3 periods - first half, second half, full period.
D = (Log(N1 + N2) - Log(N3)) / Log(2)
Alpha = exp(-4.6 * (D-1))
FRAMA = Alpha * Price + (1 - Alpha) * LastFRAMA
FRAMADiff = 100 * (shortFRAMA - longFRAMA) / ((shortFRAMA + longFRAMA) / 2)

MACD (Moving Average Convergence-Divergence)

NOTE: We support two MACD trade strategies specified in genconfig.MACD.IndicatorStrategy; "CD" (convergence/divergence), and "Diff".
CD: When MACD < signal, we sell (and visa versa).
Diff: Traditionally, this MACD strategy would sell if MACD goes below zero line (and visa versa). We do the same, but for MACDDiffUp and MACDDiffDown for fewer false positives (recommend configuring as you see fit).
MACD = MACDShortEMA - MACDLongEMA
MACDSignal = MACDSignal period EMA of MACD

DMACD (Double Moving Average Convergence-Divergence)

W use MACDLong, MACDShort, and MACDPeriod settings for DMACD
NOTE: We support two DMACD trade strategies specified in genconfig.DMACD.IndicatorStrategy; "CD" (convergence/divergence), and "Diff".
Similar to the MACD section above, except we use DEMAs instead of EMAs (yes, even on signal). See DEMA above if unsure what this means.

RSI (Relative Strength Index Oscillator)

NOTE: avg gains andlosses are smoothed after first iteration
NOTE: Want to try out RSI(2) or RSI(3)? Set those periods (2 or 3), and run 90/10 or 95/5 as ask/bid thresholds.
RS = avg_gain / avg_loss
RSI = 100 - (100 / (1 + RSI))

FastStochRSI (Stochastic RSI Oscillator)

NOTE: lowest/highest are from RSIPeriod
FastStochRSIK = ((RSI - Lowest RSI) / (Highest RSI - Lowest RSI)) * 100
FastStochRSID = FastStochRSIDPeriod SMA of FastStochRSIK

FullStochRSID

FullStochasticRSID = FullStochRSIDPeriod SMA of FastStochRSID

FastStochK (Fast Stochastic Oscillator %K)

FastStochasticK = ((Current Close - Low) / (High - Low)) * 100

FastStochD

FastStochasticD = FastStochDPeriod SMA of %K

FullStochD

NOTE: FullStochK is not includes since it's equivalent to FastStochD
FullStochasticD = FullStochDPeriod SMA of Fast %D

KDJ

NOTE: Uses new FullStoch %K and FullStoch %D calculations.
NOTE: Supports both CD and Diff in genconfig.KDJ.IndicatorStrategy.
CD: When K < D, we sell (and visa versa)
Diff: When J is above KDJJAsk, we sell. When J is below KDJJBid, we buy. J may go above and below 100 and 0.
J = (3 * D) - (2 * K)

Aroon (Aroon Oscillator)

NOTE: Supports both CD and Diff in genconfig.Aroon.IndicatorStrategy.
CD: When AroonOscillator < 0, we sell (and visa versa). This is because when AroonOscillator is 0, AroonUp and AroonDown converge/diverge.
Diff: When Aroon is below AroonBid, we buy. When Aroon is above AroonAsk, we sell.
AroonUp = 100 * ((AroonPeriod - Candles since last AroonPeriod high) / AroonPeriod)
AroonDown = 100 * ((AroonPeriod - Candles since last AroonPeriod low) / AroonPeriod)
Aroon = AroonUp - AroonDown

Ichimoku (Ichimoku Cloud)

NOTE: Utilizes Ichimoku.IndicatorStrategy for "Strong" or "Weak" strategies.
NOTE: Chikou Span's cool and all, but we don't care. We want to trade in real time, and a price list 26 periods behind only confirms if we were right or wrong. Because proper Ichimoku cloud relies on Senkou Span A being plotted ChikouSpan periods in the future, we still set this integer.
Strong: if (price > Ichimoku cloud min) and (price < Kijun-Sen) and (price > Tenkan-Sen), sell. Buy on the inverse.
Weak: if (Tenkan-Sen > Kijun-Sen), sell. Buy on the inverse. Weak is more of a standard C-D strategy.
Tenkan-sen = (TenkanSenPeriod high + TenkanSenPeriod low)/2))
Kijun-sen = (KijunSenPeriod high + KijunSenPeriod low)/2))
Senkou Span A = (Tenkan-sen + Kijun-sen)/2)) ; Plotted ChikouSpanPeriods in the future.
Senkou Span B = (SenkouSpanPeriod high + SenkouSpanPeriod low)/2))

StdDev (Sampled Standard Deviation)

NOTE: Only here to be used by other volatility indexes
The following is ripped/edited from stockcharts.com since it summarizes the Std Dev calculations quite well:

Calculate the average (mean) price for the number of periods or observations.
Determine each period's deviation (close less average price).
Square each period's deviation.
Sum the squared deviations.
Divide this sum by the number of samples.
The standard deviation is then equal to the square root of that number.

BollBands (Bollinger Bands)

Middle Band = BollBandPeriod SMA
Upper Band = BollBandPeriod SMA + (BollBandPeriod StdDev * 2)
Lower Band = BollBandPeriod SMA - (BollBandPeriod StdDev * 2)

SROC (Simple Rate of Change AKA Movement)

SROC = (Close - Close n periods ago)
if current SROC > 0, and previous SROC <= 0, BUY. Sell during the inverse.

TODO

Add PeriodDivisor support, modify indicator periods based on a volatility index (bollinger bandwidth to start). In a lot of research, this gets rid of the need for "stop loss" in most cases for going into "uncharted" territory.
Support "Max Trade Slippage" using ADX and asset + currency trade volume, and bollinger bands with bollbandwidth = (PeriodBandHigh - PeriodBandLow) / (PeriodBandSum / Period) to verify
Record at candle market depth in a new sqlite table, break markets into child markets for better economic significance backtesting (cryptotrader "at market" style backtesting is a joke in the economics world).

Running

Dependencies
- simplejson for API dependency
- pygal for graphing (not required as long as genconfig.Grapher is disabled)
- lxml for pygal (not required as long as genconfig.Grapher is disabled)
Clone
Edit genconfig.py; find your own successful configuration
Run avarice.py - This software is meant to be run continuously, and will take awhile to generate valid info depending on configuration.

Donations

Donations are not at all required, or even really requested. If you want to show a "thank you," and can't contribute code, then feel free to donate. BTC: 18XCHsFFpnSWv2GZTbQbXtGTSQGrZABueZ

billyellow/avarice