Is Tiger Woods Back?
Tiger Woods was in contention in the Valspar Championship this past week. And the buzz was incredible. The final round of this relatively unknown, early season tournament had an overnight rating of 5.11, which is higher than any golf telecast (outside of the Masters) since 2015. Tiger Woods is back. But is he back, back? Are we about to see the Tiger Woods of old winning Majors with unbelievable shots and clutch putting?
I’m going to preface this article with saying that I think I’m probably biased in answering “No” to this question. I’m a huge regression to the mean kind of guy (it’s probably due to my poker background). When we see a performance that is out of the ordinary, I’m almost always skeptical. But I tried to keep an open mind when doing research for this.
Most of the work I reference for this article has to do with a Strokes Gained (SG) Model that my brother and I created using yearly SG data and linear regression. I assume that you probably know what Strokes Gained is if you’ve found this article (but if you haven’t, you can find more info on the PGA tours website). I’m also going to make some educated guesses, and an outside resource that has made a very useful study on aging curves in golf (more on that later).
Back to Strokes Gained. There are 4 elements of SG, Off the Tee (OTT), Approach (APR), Around the Green (ARG), and Putting (Putt). One interesting thing I discovered when using linear regression to model SG was that each of these elements have different “stability” year over year. What this means is if we just simply take the last 2-3 years of data, how well can we make a model that predicts the future year. For OTT and APR, it’s actually quite stable. My model has an r-squared of .75 using the last 3 years of SG data. ARG however, is a little less stable, a solid .52. And putting, quite a bit worse, with an r-squared of close to .2. What this means for Tiger Woods is that with small samples of SG OTT and APR, we can be somewhat confident that there’s something there, but with small samples of SG APR and Putt, we can not be as confident.
And what we have for Tiger Woods right now is quite a small sample. He hasn’t been healthy since 2013. We currently have 13 total rounds this season that have shot link data (a full season is around 60-80 rounds). This is pretty imperfect data for doing any sort of statistical analysis.
So what we can do is gerry-rig a few things and then make some guesses.
The first thing we can look at is SG:OTT. In 2013, Tiger had a SG:OTT avg of -.142 (meaning he’s actually worse off the tee than the average golfer). Thinking back on the “old Tiger”, this makes some sense. Tiger has been known for being wild with his driver and using irons off the tee. This year, we’ve seen him use his Driver a bit more, but his SG:OTT has been .044. I think we can make the assumption that a 42 year old Tiger Woods probably isn’t going to be better off the Tee this year, but I’ll be generous and give him a -0.05.
Best guess: SG OTT -0.05.
Now, let’s look at SG:APR. This is Tiger’s bread and butter. in 2013, his SG:APR was 1.533, by far the best in golf. This year, his SG:APR is .672. One could argue he’s started off slow this year and his actual SG:APR is closer to this 2013 number. But coming off multiple back surgeries, I think it makes sense to be skeptical. Even at the Valspar, his SG:APR avg was 1.24, and this was the best golf we’ve seen him play in forever. I think keeping him at that .672 number is fair if not generous, he would be a top 5 golfer in SG:APR with that number in 2017.
Best guess: SG APR .672
For SG:ARG, I want to rely on earlier data since I simply do not trust 13 measured rounds when our model does not have a fantastic r-squared. In 2013, his SG:ARG was .247. in 2012, it was .194. Inserting those numbers into my linear model, it guesses his SG:ARG is .175. You might find it strange that the prediction for his SG:ARG is lower than either numbers in 2013 or 2012, but remember, since my SG:ARG model is only around .5, it takes a safe approach to it’s prediction because there’s a chance this could be random. In 2018, Tigers SG:ARG is .347. Given how long Tiger has been away from golf, I doubt this number is sustainable, but it’s safe to say he will at least be above average in this area if he stays healthy. I’m again going the conservative route and giving him slightly under what my linear model predicts given is 2013 and 2012 numbers.
Best guess: SG ARG .150
Lastly, in terms of putting, Tiger is at .541 SG:Putting this year. In 2013 and 2012, he was around .43 and .32 respectively. And remember, modeling SG:Putting on 2 years of data is not very reliable. This probably doesn’t “feel” right, as you’re probably thinking eyeballing this data that “Of course Tiger is a good putter!” The issue is the multiple comparisons problem. We have a set of pretty random data, sort of like a flip of the coin. By sheer randomness, some golfers will be good at putting over a few years and some will be bad. It’s possible, Tiger has simply been somewhat lucky. That being said, he has been playing golf for almost 20 years, and even before shot link was available we all watched and saw he was a good putter. But on the flip-side, he’s been injured for a while, and mostly working on getting back healthy, not working on his putting stroke. I definitely will give him above average putting, but I can’t give him something significant. Danny’s linear model rarely gives anyone over .340.
Best guess: SG Putt .250
Best guess: SG Total 1.067
This number would put Tiger as a top 20 golfer in 2017. In summary, pretty darn good. But, there are a couple of issues.
One issue with using 2013 and 2012 data for this is the way strokes gained is calculated. The average player that strokes gained (or lost) is calculated from is the average from the previous year. Tiger’s past numbers that we’re using are against the avg player in 2012 and 2011, while currently others are going off 2017 and 2016. Over the 5 year stretch, with the emergence of guys like Spieth, Dustin Johnson, Justin Thomas, and Rory Mcilroy, overall play has improved at least 5-10%. So the SG we see from Tiger 5 years ago has to be knocked down a bit.
Revised best guess: SG Total 1.00
But there’s an even bigger issue. Tiger is hitting a really bad part of the aging curve. 5 years ago, Tiger was 37. Today, he’s 42 and coming off multiple back surgeries and personal issues. Golf has a slightly different aging curve than most sports, as shown below in a well down analysis provided here: https://golfanalytics.wordpress.com/2013/10/11/the-aging-curve-for-pga-tour-golfers/. There actually isn’t much movement from age 21-37, golfers skill remain relatively the same. However, once a golfer hits 37, his numbers start going way downhill. The combination of injuries and age are really worrisome.
Tiger hasn’t won a tournament in the past several years, and has not been playing very well at all. Seeing a couple of rounds of “Old Tiger”, people are ready to chalk up those past years to injury and personal issues. But his drop in performance is perfectly in line with other golfers once they hit 37. It would be nice to think that the old Tiger is back, but it’s more likely that we’re just seeing glimpse’s of outstanding play in what will likely be a mediocre year that could be stopped short once again to do injury.
I saw on a sportsbook today that Tiger Woods is +300 to win a major this year, better than almost every golfer except Dustin Johnson. I think we’re being way too optimistic. I’m on -400, he will not win a major.
Final best guess: SG Total .75 (Top 40 golfer if healthy).View all posts by Max J Steinberg