streaks: Up and Down Streaks

In PMwR: Portfolio Management with R

Up and Down Streaks

Description

Compute up and down streaks for time-series.

Usage

streaks(x, ...)

## Default S3 method:
streaks(x, up =  0.2, down = -up,
        initial.state = NA, y = NULL, relative = TRUE, ...)
## S3 method for class 'zoo'
streaks(x, up =  0.2, down = -up,
        initial.state = NA, y = NULL, relative = TRUE, ...)
## S3 method for class 'NAVseries'
streaks(x, up =  0.2, down = -up,
        initial.state = NA, bm = NULL, relative = TRUE, ...)

Arguments

x	a price series
initial.state	NA, "up" or "down"
up	a number, such as 0.1 (i.e. 10%)
down	a negative number, such as -0.1 (i.e. -10%)
y	another price series
bm	another price series. Mapped to ‘y’ in the default method.
relative	logical
...	other arguments passed to methods

Details

streaks is a generic function. It computes series of uninterrupted up and down movements (‘streaks’) in a price series. Uninterrupted is meant in the sense that no countermovement of down (up) percent or more occurs in up (down) movements.

There are methods for numeric vectors, and NAVseries and zoo objects.

The turning points (extreme points) are computed with the benefit of hindsight: the starting point (the low) of an up streak can only be determined once the streak is triggered, i.e. the up streak has already run its minimum amount. Vice versa for down streaks.

When ‘up’ and ‘down’ are not equal, results may be inconsistent: in the current implementation, streaks alternates between up and down streaks. Suppose up is large compared with down, i.e. it takes long to trigger up streaks, but they are easily broken. Down streaks, on the other hand, are quickly triggered but rarely broken. Now suppose that a down streak is broken by an up streak: it may then well be that the up streak would never have been counted as such, because it was actually broken itself by another down streak. The implementation for differing values of ‘up’ and ‘down’ may change in the future.

Value

A data.frame:

start	beginning of streak
end	end of streak
state	up, down or NA
return,change	the return over the streak. If y was specified, geometric excess return is computed (see Examples). If relative is FALSE, the column is named change.

Author(s)

Enrico Schumann <[e-mail]>

References

Schumann, E. (2023) Portfolio Management with R. http://enricoschumann.net/PMwR/; in particular, see
http://enricoschumann.net/R/packages/PMwR/manual/PMwR.html#drawdowns-streaks

See Also

drawdowns

Examples

streaks(DAX[[1]], t = as.Date(row.names(DAX)))

## results <- streaks(x = <...>, y = <...>)
##
## ===> *arithmetic* excess returns
##       x[results$end]/x[results$start] -
##       y[results$end]/y[results$start]
## ===> *geometric* excess returns
##       x[results$end]/x[results$start] /
##      (y[results$end]/y[results$start]) - 1

PMwR documentation built on Oct. 19, 2023, 9:09 a.m.