Indoor Track Indexing

[Bruce Kritzler]

Glad to see NCAA is indexing indoor tracks. Question for the mathematicians on the forum. Was surprised to see the conversion factor is 4 significant figures (.000_) while timing is only 2 significant figures (.0_). Does this make sense?

[guru]

Considering it's the NCAA, I'm surprised they're not indexing the 60...

[j-a-m]

Glad to see NCAA is indexing indoor tracks. Question for the mathematicians on the forum. Was surprised to see the conversion factor is 4 significant figures (.000_) while timing is only 2 significant figures (.0_). Does this make sense?

Not a mathematician, but agree that it doesn't seem to make sense.

[MJR]

another great decision to further isolate the sport from the small fan base we already have. bravo to the ncaa and the coaches association for another idiotic decision that does nothing to improve the sport. maybe someone can create an app to do the conversions for coaches and fan so they can figure out if anyone has qualified for nationals. that should create loads of excitement for spectators and coaches at every meet.

[decafan]

another great decision to further isolate the sport from the small fan base we already have. bravo to the ncaa and the coaches association for another idiotic decision that does nothing to improve the sport. maybe someone can create an app to do the conversions for coaches and fan so they can figure out if anyone has qualified for nationals. that should create loads of excitement for spectators and coaches at every meet.

Hold on MJR. This indexing addresses a very specific issue with indoor track and field. As long as we have different sized indoor tracks, we have to make every effort to fairly index them. I would love to hear how you would address this issue while simultaneously improving the sport and NOT putting either 200m flat tracks, 200m banked tracks, undersized tracks or oversized tracks out of business.

[gh]

Yeah, MJR got it completely backwards.

This is a major win for those who have "disadvantaged" tracks (and which are still the majority of the undercover sites?). Levels the playing field quite nicely.

The thought that there are any conversions required for the fans is completely silly of course. The Q-time is what it is. That's what's listed ahead of time, what's printed in the program (if there is such a thing in this day and age), and what the meet announcer and and writers shoudl be saying during/after. It's irrelevant if somebody needs (making numbers up here) to run a 4:01.0 mile on one of the oversized speedways in Seattle or Notre Dame. If you're on a smaller setup, your target time is 4:03.0 and that's all that anybody needs to know.

Completely seamless.

Of course, I would also argue that if fans have to have an exact fix on whether or not somebody is Qing for the Nationals for them to enjoy racing, the sport is irreparably broken anyway. But that's a rant for another day.

[Conor Dary]

To quote Norbert Wiener, I am a mathematician, though obviously not at his level, and I agree with gh. Sounds a fine idea to level the playing field. The difference between Harvard's banked 220 oval and some 160 flat track is significant, and it is about time the NCAA recognized it.

[26mi235]

Glad to see NCAA is indexing indoor tracks. Question for the mathematicians on the forum. Was surprised to see the conversion factor is 4 significant figures (.000_) while timing is only 2 significant figures (.0_). Does this make sense?

Not a mathematician, but agree that it doesn't seem to make sense.

Yes, it does make sense. You have the notion of significant digits mixed up with the number of decimal places listed in the number. We could change the number of decimal places very easily by using different measuring units. For instance, the are a lot of decimal places in the time if we measure it in days (now we have added six decimal places to the time). When you measure a marathon course, you have the same accuracy if you are within one tenth of one percent of 42.195 km (42 meters) or 42,195 m (42 meters).

There is a second, more subtle aspect of what is the best estimate given the level of precision; I will not address that here.

[Bruce Kritzler]

Now I hear the indexing were based on individual athletes who ran on multiple size tracks within the same season? This is one of the dumbest things I've heard of. Athletes times are going to vary so much from early to late season, and from low key to championship meets.
A better method would be to take all times from a track, and maybe base the indexing on the 10th best time on that track.

[26mi235]

It is not track-specific, it is configuration-specific.

The process can be self-correcting over time. Few quality races are run on flat 200 tracks because of the disadvantages. Now, if the correction is too large, they will get a lot of faster marks on these tracks and they can adjust the factors. One season ought to allow much of the convergence, since NCAA Indoor is a list-driven qualification.

[RichC]

I guess I'll weigh in on this topic. Way back in 2005 or so when I was a member of the NCAA Committee I had access to a large pool of data on NCAA qualification. I started to review it since at that time the conference I was associated with had mostly 200 meter flat tracks. Obviously, it was almost impossible to qualify on those facilities. So I took all the NCAA qualifying marks for a two year period and assigned them to size track achieved on. I found that in the US 68% of all tracks are flat, 200, yet only 2.5% of NCAA qualifying performances came from flat 200 tracks. While not a rocket scientist I said......ahh hah. So I created a report that lead to adjustments on flat 200 tracks based on analyzing 50,000 performances. Someone said that there is great variance between early and late season. Precisely. Performances were only accepted for the data points from a 4 week, mid-season competition. So early and late season performances were eliminated.

From that research study another larger study was undertaken given with TFRRS one could gather hundreds of thousands of pieces of data and analyze it. The NCAA Committee assigned four individuals with outstanding statistical expertise (and they are coaches) to analyze and come up with solutions to the 200 flat, 200 banked, and oversized issue. The work they did took almost one full year. They did not use early or late season performances in the data analysis. Is there conclusions perfect? Certainly not, but what we have now is so much fairer than before and that was the simple intent. I would estimate that after a 3-5 year period, with much more data another review could take place. One of the aspects that I personally like about the new conversions is that it makes DI-DII-DIII the same. In the past a 4:09 miler in DI would have a different conversion than a 4:09 miler in DII or DIII. Obviously that was flawed.

As a coach I appreciate the new conversions and what they are attempting to do. One of the greatest pieces of information to come out of the data analysis is that there is no appreciable difference between a banked 200 track and an oversized 300 meter track.

Cheers,
Rich

[gh]

.....
As a coach I appreciate the new conversions and what they are attempting to do. One of the greatest pieces of information to come out of the data analysis is that there is no appreciable difference between a banked 200 track and an oversized 300 meter track.

For an 800 up, I believe you.

For 200s and 400s I don't believe it for a second

[RichC]

gh: I hear you. But the problem is that given the vast amount of data it just doesn't show up. So the issue for folks making the decisions (NCAA) is what do they do? Accept the data, or make a decision that is not supported by the data. But I do get your point.

[gh]

Allow me to posit that for the 200/400 data to work, it needs to have lane-draw factored in. I'm totally guessing on some of this here, but my experience tells me that while the "average" banked 200 time may well be equal to the 300 flat, I guess that lane 1 is actually at a disadvantage, while in lane 6, the advantage swings wildly to the banked track, not only because of the wider curve (which also happens on the flat 300) but also because of the drop in elevation, which is significant.

And 5 is also in favor, to a lesser degree... where the crossover point comes I don't know.

But the key is this: those putting up actual qualifying marks are "always" coming from the outer lanes (moreso on a banked track than a flat). And those are the numbers that matter. So the data set is rendered almost useless if the inner lanes are considered, even if the numbers seem to tell you something. GIGO.

I will admit, of course, that it's possible that somebody factored this in. I get the feeling no.

[RichC]

gh: Very perceptive as usual. The mandate of the committee was to generally compare flat, banked, and oversized tracks. That is what they used the data for and based on that mandate they did a great job. And I totally agree with your statement about qualifying marks coming from lanes 3-4-5-6, and more likely lanes 5 and 6. Rarely do you see athletes competing in lanes one and two in the 200-400. I wonder though if comparing data from lane 6 of the oversized to lane 6 of banked, lane 5 to 5 and so on is worthwhile. At some point we could have conversion factors for a litany of issues; temperature variance, lanes, radius, etc, etc. And where do we stop? We even have altitude conversions that take into consideration the size of the athlete but don't use them. So even though the committee could take the next step and compare lanes versus lanes there is a part of me that says....stop. Thanks for bringing up this interesting issue.

RJC

[26mi235]

Rich:

You need to let the data tell the story. You have reduced the variance around the estimates with your modeling. However, I suspect that if you go further doing essentially what gh suggests you will find that you will get a substantial reduction in the residual variance. To not test that step once the data are there seems to be to decide what the answer is in a manner that is not consistent with what you have done so far.

Above I suggested that the revised incentives will alter the competitive landscape and provide new, less biased data. Specifically, people did not gear up to the top effort for flat tracks except for some conference championships, and there the sprint-heavy regions had banked ovals, so it was 'thin' at the top. Using top athletes is an important element because they tend to be more consistent with things like starts and even the starting process (e.g., the starter, etc). Thus, I expect to see top-level qualifying efforts on tracks that would not have gotten such efforts before this. This will take the preparation bias out of the data to an extent. I expect that the revised data will, in general, reduce the adjustments for the flat 200 tracks.

Thus, as the new data come in, which should be less biased, you should also bring in other factors --- i.e., control for lane assignment! --- and you will have cleaner model and results.

[gh]

For me, the 500-pound gorilla in the room is this: anything that encourages major-meet competition on an oversized track should be discouraged, and saying that a 300-plus flat track isn't any faster than a 200 banked one is crazy.

Why discouraged? Because the sport lives and dies by its statistics, and if you encourage people to compete on tracks that aren't viable for all-time list or record consideration,what are you doing for the sport's credibility. Would anybody think it logical to allow people to throw a 15-pound shot as a nationals qualifier, then add x% to the mark? I think not.

The monster speedways may make for great training venues for large numbers of people, and that's great, but they make a mockery of the competitive sport. Particularly when they're used as non-competitive time-trial venues. Is that what our sport is really all about?

[kuha]

On the other hand, what overall importance does indoor competition have in the great scheme of things? Everyone knows that times don't truly matter that much indoors. Frankly, I think the absolute standardization of the sport is rather deadly--way too much of the same thing over and over. The variety of X-C competition is a perfectly good thing and I'd like to think that the variety of indoor competition might be an equally good thing. I'm just fine with the statistics being made complicated. Aren't they already, really?

[RichC]

GH said, "For me, the 500-pound gorilla in the room is this: anything that encourages major-meet competition on an oversized track should be discouraged, and saying that a 300-plus flat track isn't any faster than a 200 banked one is crazy. Why discouraged? Because the sport lives and dies by its statistics, and if you encourage people to compete on tracks that aren't viable for all-time list or record consideration,what are you doing for the sport's credibility. "

First off you have zero data to either support or reject your belief that 300 flat tracks and 200 banked tracks are the same. And I am not crazy. You used in an earlier post words like, "Possible", "guessing", "my experience". And that is exactly why we have a screwed up system. For the last 20 years I have been imploring the NCAA Track & Field Committee and people that make decisions to move toward a data-driven discussion. In other words have some actual facts before making important decisions and not rely on "beliefs". So many of the decisions that track & field has made over the last 30 years have been because of someones "experience" and what they "think" is best. Well thats not worth a hill of beans. We finally have intellectually gifted groups that are collecting mountains of data to try to come to informed decisions that are logical, rationale, and move the sport forward. Then there are the people that sit in the cheap seats who have nothing more to offer than "my experience". Well thats not good enough and has led us down a checkered path.

When my daughter was little she used to sit there and close her eyes and think everything around her would disappear because she couldn't see it. I will 100% agree with you that they are a statisticians nightmare. But maybe you haven't noticed that they are being built all over the country. So what do we do? Just close our eyes and make believe like they are not there? Sorry, they are there and we have to find a way to manage that in a positive manner. I loved my typewriter but when computers came I had to suck it up and move with the technology. 300 meter tracks are here to stay and more institutions will build them no matter whether you or I like them. We can't just ignore them so we have to find a way to manage them.

Just one guys opinion.

[26mi235]

Potential hidden advantage of the new system -- more dual/tri/quad meets.

Why? Because before there was a real incentive to save the big efforts for the favorable tracks, whereas most dual meets would be held on flat 200 ovals. Now it is easier to get a Q from a flat 200 and coaches might be more willing to schedule them and not travel to a few big invites.

Of course, the competition is not there so much for top athletes in dual meets (partly the scholarship limitations).

[gh]

.....

First off you have zero data to either support or reject your belief that 300 flat tracks and 200 banked tracks are the same. .....

I don't remotely believe that, even though the new standards say they are. Which is why I'm aghast that they're cast in the same category.

And the real problem is that different categories of events have different results (IMHO).

I think that people running in lanes 5 and 6 (for sure, and maybe even lane 4) on a 200 banked can run faster than anybody can on a flat 300 in any lane. And these lanes are where nationals qualifiers come from.

But I also believe that for races of a mile and above (not sure about the 800), a flat 300's marks will kick the crap out of a banked 200's. Gentler turns, and fewer of them, plus—as often as not—a much wider expanse of territory in which to run freely on the straightaways.

Bottom line: banked 200s and OT flat tracks do not remotely belong in the same category.

[Mighty Favog]

I'm with Rich on this one. If the data says something contrary to conventional wisdom, then the conventional wisdom is wrong. See: Bill James vs baseball traditionalists, Nate Silver vs political reporters.

I can appreciate the argument that the data set was not rich enough by not including lane draw, although to do so might prove that indoor long sprints are largely a farce--which I think they are, but I don't have the data to support that. But to say that the data is wrong because you don't like the conclusions it leads us to is...not the kind of thinking that should be associated with university activities.

[j-a-m]

The monster speedways may make for great training venues for large numbers of people, and that's great, but they make a mockery of the competitive sport. Particularly when they're used as non-competitive time-trial venues. Is that what our sport is really all about?

At the same time, given that outdoors is more important than indoors, isn't that an advantage that oversized indoor tracks more closely resemble outdoor tracks?

[gh]

I'm with Rich on this one. If the data says something contrary to conventional wisdom, then the conventional wisdom is wrong. See: Bill James vs baseball traditionalists, Nate Silver vs political reporters.

I can appreciate the argument that the data set was not rich enough by not including lane draw, although to do so might prove that indoor long sprints are largely a farce--which I think they are, but I don't have the data to support that. But to say that the data is wrong because you don't like the conclusions it leads us to is...not the kind of thinking that should be associated with university activities.

University activities also shouldn't be associated with a lack of critical thinking that allows one unable to tell the difference between lies, damned lies and statistics.

If this had been a black & white set of numbers it would be one thing, but because of all the variables involved, the compilers had to massage the data. I don't fault them for doing that; I fault them because it seems strongly to me that they chose a flawed methodology.

My understanding of the situation could be wrong, but if what I'm told is correct, it's easy to see how data could be manipulated to tell us that there's no difference between a 200 banked an an oversized unbanked. How?

Because the data (so I've been told) makes no distinction between oversized, and realllly oversized. So the data from Notre Dame's 352y monster and Washington's 307m giant, for example, are mixed in with results from Edinboro's 230m, Maine's 215, Swarthmore's 215, Tennessee State 209, etc.

Meanwhile, what two sites are already providing a disproportionate number of nationals qualifiers (which is what this is all about, not equalizing bottom-end performances)? Washington and Notre Dame.

I realize there can't be an endless number of permutations, but I do think they need to rework the data matching unbanked 300-plus against the 200 banked.

[26mi235]

One of the difficulties with the current batch of statistics is that it is hard to control for all of the factors. This adjustment will make it so that a lot of qualifying marks will now come from good flat 200 tracks, so the peaking/effort used will be equilibrated to a substantial degree. When that happens you will also get statistics that make the differences in the qualifying between 300m and Banked 200s more apparent. I expect to see three things:

1) and increase in Qs from Flat 200s

2) many 200s etc coming from top banked 200s (especially those that run 'downhill from lanes 5 and 6) relative to 300m tracks.

3) More distance marks from 300m tracks than from banked 200s.

These statistics will be cleaner and provide data for a revised set of factors. I love it, an incentive structure that helps reveal the advantages of the different configurations. gh, think of the long run, not what happens this year alone.

[gh]

Now posted to the front page: links to the two articles we ran on the subject in the December and January editions.

[gh]

also posted now: response from an NCAA coach on the committee.

[tandfman]

Also on the front page now, hepstrack.com weighing in on the subject (and calling the NCAA's indexing "junk science").

[j-a-m]

Doesn't the way this indexing has been done also devalue conference championships?

The argument would go like this: Conference championship are on a banked 200 track, many good 200m runners end up on inside lanes, therefore run slower than their potential. So the next week they go to a last chance meet where they run either on the outside lane of a banked track or on an oversized track. As a result, last chance meets become more important than conference championships.

[decafan]

Doesn't the way this indexing has been done also devalue conference championships?

The argument would go like this: Conference championship are on a banked 200 track, many good 200m runners end up on inside lanes, therefore run slower than their potential. So the next week they go to a last chance meet where they run either on the outside lane of a banked track or on an oversized track. As a result, last chance meets become more important than conference championships.

You bring up two very good points. 1. The 200m is never a fair fight on a 200m track, regardless of indexing. Lane draw is everything. The 200m should not be a championship event IMO for that reason. 2. The fact that so many Indoor NCAA Championship competitors are determined in last chance meets tells us there was a real need to make an indexing adjustment. Yes, I understand there are many, many variables that contribute to this point, but the fact that so few athletes qualify in the track events from a 200m flat track, when most schools have 200m flat tracks is a sizable elephant in the room to ignore. If it were up to me, you would ONLY be able to qualify for track events from 200m tracks. What is wrong with having a clear distinction between indoor and outdoor track?

[26mi235]

What is wrong is that racing on 200m track is injurious to some runners, even distance runners.

[decafan]

What is wrong is that racing on 200m track is injurious to some runners, even distance runners.

Really? That is quite a claim. Please submit one shred of empirical data that supports your statement and we can then debate the merits of ending indoor track and field altogether.

[26mi235]

What is wrong is that racing on 200m track is injurious to some runners, even distance runners.

Really? That is quite a claim. Please submit one shred of empirical data that supports your statement and we can then debate the merits of ending indoor track and field altogether.

Stuart Eagon (Wisconsin)

Actually, it is not uncommon for the coach to hold a number of the distance runners back in the indoor season because of the high stress of racing indoors (they cannot always run on Arkansas' 200B or Washington's 300+.

Given that I know a number of runners from the modest set I am familiar with, I am almost stunned that you think that there is not a shred of evidence on this.

I am not sure that those 200 and 400 guys are particularly thrilled about running a bunch of races on Flat 200s.

[decafan]

What is wrong is that racing on 200m track is injurious to some runners, even distance runners.

Really? That is quite a claim. Please submit one shred of empirical data that supports your statement and we can then debate the merits of ending indoor track and field altogether.

Stuart Eagon (Wisconsin)

Actually, it is not uncommon for the coach to hold a number of the distance runners back in the indoor season because of the high stress of racing indoors (they cannot always run on Arkansas' 200B or Washington's 300+.

Given that I know a number of runners from the modest set I am familiar with, I am almost stunned that you think that there is not a shred of evidence on this.

I am not sure that those 200 and 400 guys are particularly thrilled about running a bunch of races on Flat 200s.

OK, so no empirical evidence... Anecdotal evidence is a pretty slippery slope, my friend.

[26mi235]

Wrong reference point - you are not looking at an 'experiment' and cannot use that reference point for evaluation. In fact, you almost have to flip the Ho/evidence for alternative in this sort of case. Maybe you ought to talk with a lot of coaches.