dsr: Deflated Sharpe Ratio (DSR)

Description Usage Arguments Author(s) References Examples

View source: R/dsharpe.R

Description

Computes the deflated sharpe ratio vis-a-vis Extreme Value Theory

Usage

1
dsr(N, T, observed.sharpe, variance.sharpe, skew, kurtosis)

Arguments

N

The total number of trials (parameters) attempted

T

The total number of observations (returns) s

observed.sharpe

The observed Sharpe Ratio calculated from 'T' real returns

variance.sharpe

The observed variance of the Sharpe Ratios's across 'N' trials

skew

The observed skew of the 'T' real returns

kurtosis

The observed kurtosis of the 'T' real returns

Author(s)

Nathan Matare <email: nmatare@chicagobooth.com>

References

Bailey et al (2015) "The Deflated Sharpe Ratio: Correcting for Selection Bias, Backtest Overfitting and Non-Normality" https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2460551

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
## Not run: 
M               <- replicate(10, rnorm(1000, 0, 1))
trials          <- ncol(M) # number of trials (e.g., parameters attempted)
observations    <- nrow(M) # number of observations (e.g., returns)
evaluation.function <- sharpe.ratio <-
   function(x, rf = 0.02 / 252) mean(x - rf) /  sd(x - rf)

result <- dsr(
     N=observations,
     T=trials,
     observed.sharpe=
)
result


## End(Not run)

nmatare/quanttools documentation built on May 23, 2019, 9:32 a.m.