In partnership with

Consistently Inconsistent -
Strokes Gained Variance

Golf is such a funny, frustrating game. One day you can’t miss a putt if you tried and the next you couldn’t buy a putt if you were Elon Musk. 82 one day, 92 the next. What gives?!?

If I could just be consistent!

The reality is from round to round subtle differences in our swings, our mindset, or course conditions impact the way we play. We’re consistently inconsistent. Understanding which parts of our game are the most prone to variance is a good way to focus practice.

What do I mean by variance? Variance is simply a measure of (in)consistency. A way to quantify the distribution of possible outcomes for your golf game. Those are big statistical words, lets look at a pretty graphic.

If you navigate through the data tab to your player summary in the TANGENT Golf app, you’re going to find the strokes gained variance graphic in your player report once you have more than 3 rounds saved.

It might look intimidating at first, but its really rather simple. The graphic shows you the distribution of possible outcomes for each area of your game represented by each vertical bar (Driving, Approach, Short Game, Putting, Total) from left to right.

The height of the gray part of the bar represents the total spread of outcomes. From your best to your worst performance for each area.

The height of the green bar represents your typical performance for that particular skill. This is where most rounds are going to end up.

The goal is to make the bars as high (gaining strokes) and as short (consistent) as possible.

What I like about this plot is that it does a great job of showing your ceiling for your current skill level. If all the stars aligned and I had my best driving (+2.3), approach (+2.6), short game (+3.5), and putting (+2.4) day of my life, I could gain 10.8 strokes on an elite am! I’d be on tour!

On my actual best day in total, I gained +4.9 shots on an Elite Am! Thats motivating to know just how good I can be. That could win me the US Mid Am.

To make the stars align more often, I need to improve consistency. So the real value is the other side, not only how bad does it get (-6.5 on approach), but how much spread is there from day to day. My best approach day is +2.6 and my worst is -6.5, thats more than a 9 shot swing! I need to practice approach to make my game more reliable.

It’s just another way of looking at improvement and which parts of your game consistently show up… or which parts are driving your inconsistency. TANGENT has the answers.

Golf Podcasts I Recommend

Golf Smarts: Golf Data Unleashed

In the last newsletter I shared an experiment I recently performed with Chat GPT to see just how well the AI engine could understand the golf data from the Round Report in the TANGENT Golf app, but then I got an email from one of my favorite users, Zach, who took AI a step further.

TLDR: Zach exported data from TANGENT and imported into Claude AI to do an analysis including:
- Performance by distance bucket
- Performance by club
- Club gapping
- Wedge lofts
- Best / worst clubs
- Performance in wind… and more
And the results were awesome.

Zach is like me in that we have big golf goals and believe TANGENT can help take us there. I recently laid out a big goal of mine to compete in and hopefully win the US Mid Am in the near future and similarly Zach is gearing up to qualify as well.

That means we both need to improve our approach game and when Zach laid out how he was using data from TANGENT to do so, I was more than a little interested.

He used Claude (an AI engine) to analyze his raw data from TANGENT and the result was impressive.

He started with the Strokes Gained analysis from his player report where he knew he was going to need to improve his approach game:

After digesting the rest of the Round Report, Zach wanted to take it a step further and compare his performance by distance from the hole to know specifically what distances he should be working on to improve.

Here’s where a happy accident happened. He was unable to find the specific graphic he was looking for in the TANGENT app, so he went looking elsewhere. An AI Engine called Claude. The happy accident is that the graphics are already in TANGENT, but this allowed us to now do a comparison between the results from AI and the results from TANGENT.

I’m a big believer in users owning their data and I couldn’t possibly predict all the things golfers will want to do with it, which is why I enable a full shot by shot export from TANGENT that you can use anywhere you want.

Zach exported the shot table and fed that data into Claude. He then started asking Claude for analysis of his golf game. The first thing he asked was for Claude to break down his shot performance into 25 yard buckets. Claude did exactly that and came back with key recommendations:

1. Primary Focus: 175-225 yard approaches
   Consistent struggles regardless of lie. High frequency of shots. Consistently losing 0.3-0.4 strokes per shot.

2. Secondary Focus: 50-75 yards
   Losing 0.25 strokes per shot and even more from Fairway lies.

Pretty strong stuff. Gives clear focus on where Zach can improve, but since the analysis is also already in the TANGENT app, let’s compare to the results. Under the data tab in your player summary, there is a section called STROKES GAINED where you can find your Strokes Gained by Distance. You can look at the data two ways, per shot and cumulative.

At a glance, Claude nailed it. Zach struggles with shots over 175 yards (sg per shot < -0.26) and under 74 yards (sg per shot < -0.2) according to the red line. He is much better at the yardages in between. The TANGENT app also allows you to slice the data aggregating the strokes gained to see the total strokes gained or lost in the data set.

The data clearly agrees with Claudes analysis and before we continue to go through how AI can help you with golf practice plans, I think its also important to point out just how powerful TANGENT already is. You can go as deep as you want to go to get specific areas of your game to work on to achieve your goals.

Now the one area where I would push back on Claude’s analysis is the order of PRIMARY versus SECONDARY. I think Zach should focus on the shorter shots first and then the longer shots. Why?

Longer shots are also impacted strongly by speed. How far you can hit the ball. There his a big difference between hitting a 5 iron 200 yards and a 3 Hybrid. So to really improve the longer shots, we would need to cut our dispersion in half! Which likely means we need to hit the ball further and turn that 3H into a 5 iron.

Wedge shots don’t require as many physical attributes and just due to the raw number of attempts he’ll have from that bucket… some serious improvement can be had. He’s losing almost 50 strokes in the above graphic!

Zach went further and asked Claude to do the split by club. Luckily, TANGENT can already do this too.

And again, TANGENT and Claude agreed on best clubs (PW), worst clubs (3H), and opportunities (7I @ -0.28).

Next, Zach asked Claude for a gapping analysis. This is a feature we’ve discussed adding to TANGENT, but have not yet implemented. So I was extremely interested in how Claude would tackle it and no surprise, Claude used a combination of distance and dispersion data from TANGENT to go through Zach’s bag and give a pretty good report including notable gapping issues:

This was great analysis and not far from how I would do it myself.

One thing Claude did not understand initially is that wedges are often used for less than full shots, so the inconsistency in distances is expected. However, after Zach explained this concept to Claude, it was able to do a new analysis focusing on the 70-90th percentile. This is actually pretty close to how TANGENT already does distances for wedges.

I’ll stop before we get any deeper into the chat with Claude, but I will point out the further things it was able to do with Zach’s TANGENT data…

• Based on the adjusted wedge distances, Claude gave a suggestion of wedge lofts to properly gap his bag (46°→50°→55°→60°), including pros and cons of other lofts.

• Claude justified keeping both a 3W and a 3H in the bag despite the clubs having overlap.

• Claude compared off the tee versus fairway performance for 3W and 3H.

• Claude compared approach performance based on wind conditions from light winds to strong winds.

As a developer, I’ve spent a significant amount of time predicting what features users want and trying to add graphics and analysis to support it, which requires a lot of code and effort.

Club fitters agonize over data to figure out how to properly fit and adjust your bag.

Claude can do it with a few prompts. You just have to have the data… and TANGENT has the data.

If you want to learn more about how AI can help you with more than just TANGENT, check out this letter’s sponsor. I subscribe to Rundown AI’s newsletter keep on top of the fast moving industry and I’m actively taking one of their courses to learn new AI skills. Highly recommend.

Start learning AI in 2025

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Sign up to start learning.