Build columns from the expected dropback model. Will return NA on data prior to 2006 since that was before NFL started marking scrambles. Must be run on a dataframe that has already had clean_pbp() run on it. Note that the functions build_nflfastR_pbp() and the database function update_db() already include this function.

## Usage

add_xpass(pbp, ...)

## Arguments

pbp

is a Data frame of play-by-play data scraped using fast_scraper().

...

Additional arguments passed to a message function (for internal use).

## Value

The input Data Frame of the parameter pbp with the following columns added:

xpass

Probability of dropback scaled from 0 to 1.

pass_oe

Dropback percent over expected on a given play scaled from 0 to 100.