Trade-Analysis {PMwR}

Analysing Trades: Compute Profit/Loss, Resize and more

Description

Functions to help analyse trades (as opposed to profit-and-loss series)

Usage

scale_trades(amount, price, timestamp, aggregate = FALSE,
            fun = NULL, ...)
split_trades(amount, price, timestamp, aggregate = FALSE,
             drop.zero = FALSE)

limit(amount, price, timestamp, lim, tol = 1e-8)
scale_to_unity(amount)
close_on_first(amount)

tw_exposure(amount, timestamp, start, end, abs.value = TRUE)

Arguments

amount

notionals

price

a vector of prices

timestamp

a vector.

aggregate

TRUE or FALSE

fun

a function

lim

a maximum absolute position size

start

optional time

end

optional time

drop.zero

logical. If TRUE, trades with zero amounts are removed. See Examples.

abs.value

logical. If TRUE, the absolute exposure is computed.

...

passed on to fun

tol

numeric

Details

scale_trades takes a vector of notionals, prices and scales all trades along the paths so that the maximum exposure is 1.

The default fun divides every element of a vector n by max(abs(cumsum(n))). If user-specified, the function fun needs to take a vector of notionals (changes in position.)

split_trades decomposes a trade list into single trades, where a single trade comprises those trades from a zero position to the next zero position. Note that the trades must be sorted chronologically.

Value

Either a list or a list-of-lists.

Author(s)

Enrico Schumann

See Also

btest

Examples

n <- c(1,1,-3,-1,2)
p <- 100 + 1:length(n)
timestamp <- 1:length(n)

scale_trades(n, p, timestamp)
scale_trades(n, p, timestamp, TRUE)  ## each _trade_ gets scaled

split_trades(n, p, timestamp)
split_trades(n, p, timestamp, TRUE)  ## almost like the original series



## effect of 'drop.zero'
P <- c(100, 99, 104, 103, 102, 105, 104)  ## price series
S <- c(  0,  1,   1,   0,   0,   1,   0)  ## position to be held
dS <- c(0, diff(S)) ## change in position ==> trades
t <- seq_along(P)

#### ==> 1) with all zero amounts
split_trades(amount = dS, price = P, timestamp = t)

#### ==> 2) without zero-amount trades
split_trades(amount = dS, price = P, timestamp = t, drop.zero = TRUE)

#### ==> 3) without all zero-amounts
zero <- dS == 0
split_trades(amount = dS[!zero], price = P[!zero], timestamp = t[!zero])