Penalty Strokes on the PGA Tour

There are few statistics available concerning the occurrence of penalty strokes in professional golf. This article provides the raw numbers as well as some adjusted figures to tell the story of penalty strokes on the PGA Tour.

The analysis uses shot-level data from 2007-2016. Due to data limitations, the majors, WGC events, and events held opposite to majors are excluded. To start, the figure below show the distribution of players’ average penalty strokes per round, along with lists of players with the 10 worst and 10 best penalty stroke rates.

It can be seen that the majority of players average between 0.25-0.50 penalty strokes per round.

The next figure details the average penalty strokes per round by course on the PGA Tour.

The course name is found by hovering over each data point. Courses in the top right corner of the figure indicate difficult courses that yield high numbers of penalty strokes, while those in the bottom left are courses that yield few penalty strokes and have low scoring averages. In general, there is only a weak relationship between penalty strokes and scoring averages by course.

To get a better sense of which players take unusually high numbers of penalty strokes, two adjustments are made.

First, we saw previously that some courses have higher numbers of penalty strokes incurred than others. Given that players have different schedules, differences in courses played can affect a players’ average penalty strokes per round. To correct for this, for each round played we subtract the course-specific penalty stroke average from the number of penalty strokes incurred that day for each player. Then, the average of this adjusted penalty stroke measure over all rounds played, for each player, will take care of any differences in playing schedules between players.

Second, we would like to adjust for the ability of players. Evidently, golfers who are playing poorly (as indicated by their relative scoring average) are expected to have more penalty strokes per round. For example, if it were the case that Phil Mickelson and Heath Slocum had the same average penalty strokes per round, we would say that Mickelson has an unusually high average (as compared to Slocum) due to the fact that Mickelson’s relative scoring average is better than Slocum’s.

This latter adjustment is made using a statistical tool called linear regression. Without getting into specifics, we can use regression to predict a player’s expected penalty strokes per round given their relative scoring average. Then, it is the difference between a player’s actual penalty strokes per round and their predicted penalty strokes per round that is the new object of interest (called the “residual” in the regression context). The prediction for penalty stroke rates as a function of relative scoring average is shown below.

You can hover your mouse over each data point to view the player name. Also, note that the penalty stroke rates are already adjusted for course effects (that is why Furyk no longer has the lowest penalty rate).

It is now the distance between a data point and the prediction line that is the measure of interest. Players below the line take fewer penalty strokes than what is predicted by their ability level, while those above the line take more penalty strokes than what is predicted by their ability level. For example, although Jim Furyk and J.J. Henry have almost identical raw penalty stroke rates, once we adjust for ability it can be seen that J.J Henry has a lower adjusted rate (i.e. Henry’s data point is further below the prediction line than Furyk’s is). These adjustments make an important difference in determining who has the lowest adjusted penalty rates, however it makes little difference for those with the highest penalty rates (i.e. the raw and adjusted penalty rate rankings are quite similar).

Finally, we take a closer look at (unadjusted) penalty strokes incurred as the result of tee shots. We classify tee shot penalty strokes into two categories; out-of-bounds (OB) tee shots, and tee shots that find hazards or are unplayable (Note: par 3s are included in this analysis). The following figure provides the number of tee shots that find OB or hazards per 1000 tee shots hit, for each player.

Again, explore the data by hovering over the data points. It is striking how few penalty strokes some players take on tee shots. Remarkably, in our sample Fred Funk hit 3204 tee shots, none of which found the OB, and only 13 of which found a hazard.