For years, I’ve been lamenting the relative lack of meaningful or even vaguely useful bout-level statistics in wrestling. GoPSUSports publishes takedown data sometimes in its match notes, but they only include dual meets (with no tourney data), and they’re sealed away in PDF files. Not very user friendly for your neighborhood blogger.
So when I saw that Penn State fan and wrestling message board frequenter Brian Jones (LemonPie on the boards; @LemonsAndPie on the twitter) was charting some dope takedown stats in a google sheet earlier this year, I salivated.
Then I stalked, slowly and across the weeks of this season.
Then I asked nicely, and he agreed to share his data and explanations in this post.
Brian Jones / Lemon Pie:
When I took on the project of detailing every piece of available scoring data for the 2018-2019 Penn State Nittany Lions wrestling team, I hoped to provide an appealing opportunity for wrestling fans to analyze data of some of the country’s most prolific scorers. But the aim was not just to offer raw totals; I also wanted to explore and propose different methods in which we view traditional wrestling statistics such as points, takedowns, nearfall, and riding time.
In doing so, I applied a formula that acts as the foundation for many of the statistics featured in the spreadsheet. My hope (and argument) is that utilizing this metric will give us a more accurate representation of wrestling scoring data than totals alone.
The Seven-Minute Metric
If you’re not familiar with the basketball statistic, Per 36 Minutes, here’s a quick rundown: it’s a largely ignored NBA metric that adjusts every player’s stats to reflect how they would look with typical starter’s minutes (36 minutes). And though it’s a perfectly reasonable and unbiased metric, the problem is that the best players are already playing 36+ minutes and it’s hard to get fans to care about role players like Boban Marjanovic, who is a top 40 scorer when his stats are adjusted threefold.
But in wrestling, when the best competitors are compiling pins and tech-falls—and therefore competing less in overall match time—adjusting stats actually makes the most sense. So instead of 36 minutes, say we applied the same formula with a standardized time of seven minutes (or 420 seconds). The decision to use seven minutes (or 420 seconds) isn’t arbitrary. Seven minutes is the standard length of a collegiate wrestling match and it is my belief that the vast majority of matches at the Division-1 level end at 420 seconds.
I’d expect that the experience of a wrestler like Brady Berge, who has seen 15 of 18 matches go seven minutes, is more common than that of Jason Nolf, who reached seven minutes just six times. I trust that anyone who has wrestled or is a fan of wrestling should have a decent feel for seven minutes and the amount of action that occurs within those moments. It wouldn’t make sense to use any other standard of time.
The argument is that by applying this per-7-minute metric to totals, we normalize wrestlers who finish matches early (and are therefore denied the same opportunity to accumulate raw totals as those who infrequently secure a pin or tech) are not punished for doing so.
And no, there’s no actual punishment for flattening one’s opponent; I’m being hyperbolic here. But as an example, Bo Nickal, at his pin rate, could never lead the country in overall takedowns. But if using an adjusted metric, and we knew exactly how many takedowns he averaged per seven minutes, then hypothetically there is nothing to stop him from consideration as the country’s best neutral wrestler. Especially if we had comparative data for every other wrestler.
The best way to demonstrate the relevance of a 7-minute metric is to first break down the available takedown data. Below is Penn State’s individual takedown data in its most basic form. You may even find this type of data on your favorite team’s webpage if you look around.
Now, this raw data is certainly useful and fun to keep track of, but without the proper context it is possibly misleading. For example, we know 2018 All-American Shakur Rasheed is a very capable neutral wrestler, so how is it that his takedown numbers resemble closely to that of a 6-14 wrestler in Devin Schnupp?
The first reason is obvious—Shakur hasn’t wrestled as many matches. Devin Schnupp has wrestled 20 matches with Rasheed only having wrestled 17. To counter this imbalance of matches wrestled, the instinct here might be to calculate the number of takedowns each wrestler earned per match.
For example, if we did that for each PSU wrestler, it would look like this:
- Schnupp’s 27 takedowns divided by number of matches (20) = 1.35 takedowns per match
- Shak’s 34 takedowns divided by number of matches (17) = 2 takedowns per match
Alright, the All-American has a better takedowns per match ratio than the 6-14 wrestler (2 > 1.35), so perhaps any alarm over Rasheed’s takedown total is a bit misguided. However, while this method may have shown value for this specific example, examining the data further will demonstrate that the “per-match” method is in fact a misstep in the analytic process.
While we might consider both Anthony Cassar and Bo Nickal to be elite neutral wrestlers, objectively we might not naturally presume that Anthony Cassar is the superior neutral wrestler to the 2x NCAA Champion, as the chart above indicates. But with both Cassar and Nickal having wrestled 22 matches and Cassar having obtained 36 more total takedowns, Anthony’s takedowns per match rate well ahead of his teammate (to 3.45 to 4.64). Heck, looking at the table above, Bo is sitting behind Brady Berge using the “per match” method. So what seems to be the issue?
I know you are all way ahead of me here. As alluded to at the beginning of the article, the answer lies in the opportunity. Or in Bo’s case, the lack thereof.
With the amount of pins and tech-falls that Nickal has secured this year (it’s a combined 17, by the way), he found himself spending significantly less time on the mat. By introducing the time (in match seconds) each PSU wrestler has spent in competition this year, we learn that Bo Nickal has wrestled 40% less than Anthony Cassar in raw match time. Therefore, Bo has had far less of an opportunity to accumulate takedowns in the same volume as the heavyweight wrestler.
Keeping track of cumulative match time is essential. It is only when this data is made available that we can apply the 7-minute adjusted metric. In attempting to place each wrestler on a level playing field with a standardized time, I expect we might detect a more thoughtful comparative analysis of each wrestler’s scoring data.
Takedowns Per Seven Minutes (TD/7min)
In order to find the number of takedowns one might earn in a seven-minute time period, simply plug a pair of data points into the following formula:
TD/7min = (420/number of match seconds wrestled) * number of takedowns
Example: Nick Lee (using table above):
- Match Seconds Wrestled: 8691
- Takedowns: 98
- TD/7min= (420/8691)*98 = 4.74
When TD/7min is calculated for the rest of the PSU team, I think you’ll find—as hypothesized—that that the numbers are a much better depiction of takedown ability than either 1) raw totals or 2) a “per-match” basis.
Scanning the table above, Roman Bravo-Young is revealed as the median performer for Takedowns per Seven Minutes (TD/7min). In a talent-heavy lineup like Penn State’s, this is about where fans might view the true freshman and his skill set as a takedown artist. In contrast, on the “per-match” premise, he rates as third on the team.
Another finding is that Shakur Rasheed proves himself more efficient in the metric, gaining significant ground on Mark Hall and exhibiting a number more aligned with his All-American status. Finally, Hodge Trophy contenders Bo Nickal and Jason Nolf separate themselves from the pack, as most would expect.
The great thing about the 7-minute adjusted metric is that it isn’t just limited to takedowns. It also happens to be the basis of many other statistical categories displayed in the spreadsheet (e.g. points, nearfall, riding time) and I believe it proves superior, all the same.
But before we get to how the 7-minute metric is applied to other wrestling statistics, first I want to quickly introduce a metric I consider to be an even better measure for analyzing takedowns and neutral ability.
Takedowns Per 60 Seconds From Neutral Wrestling
The best argument against TD/7min is made with this simple case right here:
No knowledgeable wrestling fan would think Daniel Lewis’ TD/7min for this legendary 2017 NCAA Championships match (1.0) is any gauge of his neutral wrestling ability. The fact is, the conscious decision to ride (6:51 for Daniel here!) in excessive amounts can have a tremendous impact on TD/7min.
To combat this issue, I theorize that isolating the time spent in neutral position and applying the same adjusted metric (this time with sixty seconds) would serve as the best barometer for takedown ability.
Unlike TD/7min, the reason for choosing a 60-second standard is actually arbitrary. It’s not because the average time spent in neutral per match is one minute (it’s probably closer to three or four), but it’s because I find it reasonable for a wrestler to expect—or at least strive—to take his opponent down at least once when given a minute on their feet.
But it’s not as easy as one might expect, for even great wrestlers to do that.
Look at the following individual tournament data from three of the best teams in the country: Iowa results from the Midlands tournament and Oklahoma State and Penn State at the Southern Scuffle, respectively. Only 9 of 30 starters achieved what I consider to be a great benchmark for demonstrating success in neutral (> 1).
Unfortunately, Takedowns per 60 Seconds from Neutral is not without its flaws. Specifically the following pair of issues surface whenever attempting to apply the metric:
- It takes much longer for a person to record this data—either by mining FloArena or Trackwrestling dashboards (timestamps on box scores) or recording the neutral wrestling seconds manually with a scoreboard or stop-watch.
- It is much more susceptible to error than TD/7min—either on the part of the dashboard or the person collecting the data.
Now, in my spreadsheet I have safeguards in place to make sure every piece of time adds up to the total match time, but unfortunately there is no way to quantify an exact time spent in neutral without potential error. Similar to how a “Riding Time Clock” is vulnerable to human error, so is the case with tracking the time spent in neutral wrestling
Nonetheless, I feel confident in my numbers, and having isolated the neutral time for the Penn State wrestlers, their Takedowns Per 60 Seconds from Neutral stats are as follows:
Perhaps unsurprisingly (due to top-wrestling prowess), the big winners in applying this metric appear to be Nick Lee and Shakur Rasheed.
Nick Lee’s numbers brings us to a discussion regarding TD/60 seconds from neutral and how it relates to his wrestling style. With his absurd pace, Nick doesn’t waste any time on his feet. He will either take you down or get taken down trying.
Compare that to NCAA Champion Mark Hall, who might feel more comfortable staying on his feet with world-class defense and hand-fighting, and it might explain some of the disparity between them. So even though Nick’s takedown numbers here are over 50% better than Mark’s, I don’t think we can definitively say that Lee is better on his feet.
The knowledge of various wrestling styles is a much-needed bit of context in making sense of these numbers.
Also, I encourage you to look at the split statistics vs. ranked opponents (rankings via FloRankings) to see how greatly they affect the stats. You’ll find that Hall has wrestled ten ranked opponents to Nick Lee’s three. There simply aren’t enough matches in a wrestling season to ever diminish the impact of quality of competition.
One of the last key questions to ask of takedown statistics is: how often is our guy getting taken down? The answer is told first by mining and recording all takedowns against, which I’ve done for us, and which you can see in the Opponents section of the spreadsheet.
Rasheed is the standard-bearer, having never been taken down yet this season. But even the rest of that lineup’s takedown defense is ridiculous! Schnupp & Nick Lee are the only starters who have given up more than 5 takedowns all season! That’s unreal.
You can probably guess the next thing we would want to examine here: a comparison of takedowns for versus takedowns against, in a way that showcases takedown differentials.
If we take the three takedown metrics:
- TD raw counts
And rearrange them into For and Against columns, it was easy to drop in a differential column for each:
Jp: I’m not sure I can take all that dominance. It kinda feels like Kai in Kung Fu Panda 3, when Po gives him all the chi he’s been so greedily collecting, and he starts soaking it in before realizing it’s too much.
While I’m really surprised and impressed by Schnupp’s numbers (would you have guessed that he scores one and a half takedowns every 7 minutes?), it’s Nolf the Matrix who really knocks you out.
Not only is he going to score not one but two takedowns on you every freaking minute the two of you spend in neutral, but you would have to spend over 11 minutes in neutral with him, just to average a single takedown of your own.
LP: Finally, I just want to re-emphasize the importance of takedown statistics.
According to TrackWrestling dashboard data, of the 50 matches wrestled by last year’s NCAA champions in Cleveland, 44 of them saw the champion with a net-positive takedown differential (Track footnote).
It turns out superior neutral ability is a pretty significant predictor of match outcome.
Jp: Man, this stuff is so awesome for Wrestle Nerds like me. The Information Age has brought our favorite sport such a long way, but look at all this cool stuff that remains largely untapped!
I recruited LemonPie pretty hard for this post for a couple reasons. First, as noted, I personally eat it up. I really wanted to read the deeper stories that the data tells us about this unbelievable wrestling dynasty. And in no surprise to anyone, this data tells us even more about Penn State’s unreal levels of dominance.
Second, I wanted to showcase how much insane work LP has put in while charting this stuff this season. Read into the Appendices below to get a more detailed sense of what’s involved in deriving the seconds-spent in x position data—from a single bout. I’m sincerely personally grateful that he’s done the work and that, in the true Spirit of Internet, has shared it to all of us, for free.
Third, I wanted to help organize it in a way that it serves as an example for the individuals and services already out there, entering and keeping scores. I suspect LP’s sheets of formulas can serve as a Requirements Document of sorts for builders of score-keeping software. I’m not a paying consumer of any current for-fee offerings, so I can’t offer any takes on how much of LP’s stats the market currently offers or is missing. But if any of the big service providers that are currently posting totally free or free-ish bout results—or, especially, if any of the smaller indie services (say, @WrestleStat?), could incorporate a few of these formulas into their existing offerings, hoo boy, would the wrestling community benefit!
Lastly, I wanted to rope him in hard because I can see in his spreadsheet that he’s got more gold in there!
This post got long, so we held off on presenting and analyzing all the Riding Time and Nearfall data he’s mined and charted. I convinced my man to dust off his old SBN login (BrianJonesLP), so I’m hoping we can get him to build his next one out as a Fanpost, which we would undoubtedly bump to the front page.
I really cannot wait to read what you guys think of this. I hope you can give Brian/LemonPie enough reason to do more of these in the future. Thanks again, LP!
Appendix A: Sample Data Mining Using Trackwrestling
To give you an idea about this data collection & recordation process, say you want to isolate the neutral seconds for Zain Retherford’s 2018 Championship run. The 2018 NCAA brackets with underlying bout scoresheets can be found on the Trackwrestling website and when you pull up the 149 pound bracket it will look like:
Each bout number in the bracket (Zain’s bouts marked with eyeglasses) is linked to a box-score that reveals the recorded dashboard bout data.
Using Zain’s Championship Final as a specific example, here’s how you could monitor the timestamps (changes in action) to calculate his time spent in the neutral position:
Next up: transcribe your notes. Here’s how I did this bout:
- 3:00-2:06 = 54 seconds in neutral
- 2:06-1:17= 49 seconds in the top position
- 1:17-0:32= 45 seconds in neutral
- 0:32-0:00= 32 seconds in the top position
- 2:00-1:37 = 23 seconds in bottom position
- 01:37-0:00 = 97 seconds in neutral position
- 2:00-1:12 = 48 seconds in top position
- 1:12-0:00= 72 seconds in neutral
After That, Math:
- Time in neutral = 54 + 45 + 97 + 72 = 268 Seconds
- Time on top = 49 + 32 + 48 = 129 seconds
- Time on bottom = 23 seconds
- Neutral time (268) + Top (129 ) + Bottom (23) = 420 Seconds
Lastly, formula data entry. To determine his TD/60sec from neutral:
- (60/neutral time)* number of takedowns
- (60/268)*2 = approximately 0.45 TD/60 sec of neutral
The data for this individual match would be entered into a spreadsheet to be calculated with every other one of Zain’s NCAA matches to get his TD/60sec from neutral for entire the tournament.
As you might expect, isolating the neutral seconds can be a lengthy process for a wrestler like Jason Nolf where changes in action come quickly and often.
Appendix B: Sample Data Mining Using Flo Arena
To mine similar data from Flo Arena, find your sought-after bout result that looks like this:
When browsing their brackets, clicking the area indicated by the white arrow will pull up a bout score sheet that looks like this:
Then apply the same note-taking, math and formula operations as described in Appendix A.