Published
Edited
Nov 20, 2019
2 forks
Importers
1 star
mkrRollingBeta = {
return dailyCloses[selectedIdx].map((close, idx) => {
let cleaned = dailyPercentageChanges[selectedIdx]
let btc = btcPercentChange.slice(-cleaned.length)
if (idx < betaLookback) return 0
let startIdx = idx - betaLookback
return calculateBeta(cleaned.slice(startIdx, idx), btc.slice(startIdx, idx))
})
}
mkrForwardChange = getForwardChange(dailyPercentageChanges[selectedIdx], dailyCloses[selectedIdx])
mkrForwardChangeClean = betaChart.map(data => [data.beta, data.forwardChange]).filter(x => !isNaN(x[0]))
import {batExpectedReturn} from '@bgits/crypto-protocol-expected-returns'
permutatedRiskReturnsChartNoBat = vz(permutatedRiskReturnsNoBat)
.y('returns')
.x('portfolioStd')
.detail('details', 'n')
.width(900)
.height(600)
.plot();
permutatedRiskReturnsNoBat = {
return [
['DAI', 'MKR', 'SNT', 'ETH'],
['DAI', 'SNT', 'ETH', 'MKR'],
['MKR', 'DAI','ETH', 'SNT'],
['SNT', 'ETH', 'MKR','DAI']
]
.map(calculatePortfolios)
.reduce((pv,cv) => [...pv, ...cv.riskReturns],[])
}
mkrCollateralTypes = ['SNT', 'ETH', 'GNT', 'DGD', 'ZRX', 'BAT', 'OMG', 'REP']
makerCDPCovariances = coorelationMatrix(mkrCollateralTypes)
cdpKeys = Object.keys(makerCDPCovariances)
makerAvgCovariances = mkrCollateralTypes.map(sym => {
const values = cdpKeys.filter(key => key.includes(sym)).map(k => makerCDPCovariances[k])
return {[sym]: simpleStatistics.mean(values)}
})
calculatePortfolios = (syms = []) => {
const coorelations = coorelationMatrix(syms)
const weights = combinationSum(syms.length, 100).map(x => x.map(x => x/100))
const returns = weights.map(comb => simpleStatistics.sum(comb.map((weight,i) => weight * forwardExpectedReturn[syms[i]])))
//TODO look into use PMPT - sortino/ downside volatility only
const stds = syms.map(ticker => {
const pctChanges = getSymbolPercentageChanges(ticker).slice(-sharpeLookback)
const std = simpleStatistics.standardDeviation(pctChanges)
return std
})
const weightedStds = weights.map(comb => simpleStatistics.sum(comb.map((weight, i) => (weight * stds[i])**2)))
const weightedCors = weights.map(weights => {
let cors = []
for(let i=0;i<syms.length;i++){
for(let j=i+1;j<syms.length;j++){
const first = syms[i]
const second = syms[j]
const key = `${first}_${second}`
const cor = coorelations[key]
const result = 2*weights[i]*weights[j]*stds[i]*stds[j]*cor
cors.push(result)
}
}
return simpleStatistics.sum(cors)
})
const portfolioStds = weightedStds.map((std, i) => Math.sqrt(std + weightedCors[i]))
const details = weights
.map(w => w.map((w, i) => `${syms[i]}: ${w}`))
.map(w => w.reduce((pv,cv) => `${pv} ${cv}`))
.map((w,i) => `${w} sharpe: ${(returns[i] - riskFreeRate) / portfolioStds[i]}`)
const riskReturns = portfolioStds.map((p,i) => ({ portfolioStd: p, returns: returns[i], details: details[i]}))
return { weights,returns, portfolioStds, stds, weightedStds, coorelations, weightedCors, riskReturns, details }
}
Purpose-built for displays of data
Observable is your go-to platform for exploring data and creating expressive data visualizations. Use reactive JavaScript notebooks for prototyping and a collaborative canvas for visual data exploration and dashboard creation.
