How to read a dispersion chart

A dispersion chart is a picture of where a set of shots with one club finished, plotted around the target you were aiming at. Almost everything worth knowing in it lives in two properties: where the middle of the pattern sits, and how wide the pattern is. The individual dots matter least.

What the pattern is made of

One of the most careful public models of amateur shot patterns comes from Mark Broadie and Soonmin Ko, who built a simulation of golf calibrated against Golfmetrics data - over 55,000 shots from more than 160 mostly amateur golfers recorded between 2005 and 2009 - plus PGA Tour ShotLink data from 2004 to 2008.

In their model a shot is the product of two independent errors. The first is directional: an angle between the line you aimed on and the line the ball actually finished on. They draw that angle from a normal distribution with a mean of zero, and they say plainly what the zero mean means - golfers do not systematically miss to the left or right of the target. They report that the normality and zero-mean assumptions are generally consistent with the data.

The second error is in distance, and it does not behave the same way. For shots other than putts, they note that the distance error often has a negative mean, meaning shots on average finish short of the target, along with negative skew. The reason is an asymmetry in what can go wrong: the furthest a club can go is capped by your physical ability, but a chunked shot can travel almost nowhere.

That is why a dispersion pattern is rarely a neat circle: symmetric side to side, lopsided front to back, with the long tail on the short side.

The middle matters more than the widest miss

Width is described by a standard deviation, which TrackMan defines plainly: a measure of how spread out numbers are, where the bigger the number, the wider the dispersion.

Broadie and Ko report that figure for driver tee shots by scoring group. For rounds scoring between 96 and 107, the standard deviation of directional error was 7.9 degrees. For rounds scoring between 78 and 82 it was 5.5 degrees. The 75th-percentile driver distance for those same two groups was 225 and 250 yards.

What that width costs you shows up in the scoring, and it shows up through the whole pattern rather than through single disasters. In their results, a golfer averaging 100 who reduced directional standard deviation by two degrees dropped about 2.6 shots per round. Fairways hit rose from about 43 percent to 53 percent, recovery shots fell from about 20 percent to 15 percent, and tee shots finishing out of bounds fell from 4.4 percent to 2 percent. The same golfer gaining 20 yards of driver distance dropped only about 1.2 shots. Their headline finding, contrary to earlier claims, is that for long tee shots directional accuracy has a greater impact on scores than distance.

Look at which things moved: fairways held, recovery shots avoided, balls kept in bounds - consequences of the spread as a whole, not of one wild swing. That is why width is the thing to plan around.

One caveat they supply themselves: those figures are for a typical course with trees, sand, water and out of bounds. On a modelled course with no trees, the same two-degree gain was worth 1.7 shots and the 20-yard distance gain was worth 2.0. Width costs you in proportion to how much trouble is out there.

How many shots make a pattern

A 2022 test-retest study published in Sensors put ten amateur golfers with a mean handicap of 31.2 through two sessions a week apart, fifteen swings each with a driver and a six iron, measured on a Doppler radar unit.

Ball speed, club speed, carry distance and total distance reached excellent reliability with only one or two shots. Smash factor, roll distance, spin rate and several flight variables needed between three and eight. Overall the authors recommend at least seven swings with the driver and ten with the six iron.

The exception is the one that matters most here. Variables related to side carry could not be reliably measured in these players in fifteen trials at all, with an intraclass correlation below 0.26.

Read that twice. The near-and-far axis of your chart settles almost immediately. The left-and-right axis, for players of that standard, had not settled after fifteen shots. The study does not tell you what number would have been enough, only that fifteen was not. The authors also note that variability in the swing performance of low-handicap players is lower than in high-handicap players, so a better player's pattern firms up sooner.

Reading a biased pattern

Because a mean of zero is both what the model assumes and what the data supports, a pattern whose centre sits consistently off your target is saying something different from a pattern that is merely wide. Width is a fact about your consistency, and you plan around it by moving the centre away from trouble. A displaced centre is a systematic bias in aim, setup, or how the face arrives, and it is something to fix. The catch is the paragraph above: you need enough shots before you can be confident the centre really has moved.

Compare like with like. Broadie and Ko report that distance errors are larger from the rough than from the fairway or tee, larger again from sand, and that relative distance error is roughly constant beyond sixty yards but grows as the shot gets shorter. A pattern pooled across lies and distance bands is a pattern of nothing in particular.

What it will not tell you

A dispersion chart records where the ball finished, not what the swing did. Two identical-looking patterns can come from entirely different causes.

It also describes a player and a measuring device together. Hitting into a simulator screen is one instance of that arrangement: the launch monitor reading your swing is the measuring device. The reliability figures above came from one specific radar unit used with one group of players, so they are a property of that whole arrangement rather than of golfers in the abstract.

And it is quietest on the question golfers most want answered: whether this month's pattern is genuinely different from last month's. A pattern that has not settled cannot tell you it has moved.

GSCaddie fits a dispersion pattern to your own shots, club by club.