Quals position often not a perfect predictor of race success

BY CHARLES SCUDDER | @cscudder

The name of this site may be 33to1, but there’s a lot more that goes into a team’s odds of winning the Little 500 on race day than meets the eye. Team experience, speed in exchanges, track conditions, unexpected crashes and so much else plays into whether or not a team will be successful at Bill Armstrong Stadium.

The spring series events help test those different variables. ITTs find the fastest individual rider. Team Pursuit lets us see the teams perform as a group. Miss ‘N Out combines both speed and strategy.

Throughout the spring series, I’ll be looking at the odds each team has in raising the Borg-Warner Trophy on race day. I’ll be trying to predict probabilities based off historical numbers charting past team’s successes and failures.

Disclaimer: I’m not a math guy. I don’t pretend to be one. I’m kind of making this up as I go, so if you have a problem with my math, or are a statistician who finds error in my methods, email me at cscudder@indiana.edu. I’ll try to be as open and honest about my methodology as I update these probabilities. No model can be perfect, but my calculations will hopefully provide an idea of what the field will look like when the checkered flag drops.

In the past 20 years of the race, only three men’s winners and four women’s winners came from the pole position. (Last year, Beta Theta Pi broke the mold by winning the men’s race from the pole.) It’s the “Curse of the Green Jersey,” and it means that — based exclusively on qualifying position — pole winners Phi Delta Theta and Alpha Chi Omega have a 15 percent chance and 20 percent chance of winning, respectively.

For the women, it’s second place in qualifications that earns the most wins with 25 percent. The men’s results are split evenly between first, second and fourth qualifying position. The lowest starting position during that period of time to come back and win was 1997, when Cutters made a comeback from 19th to first in the 200 laps.

Whether the curse will crop up again this year has yet to be seen, and qualifying position isn’t the best predictor of success.

The percentages below are based solely on qualifying position of winning teams for the past 20 years. (IUSF records for spring series events are only available from 1994 to the present.) These odds will be updated as the spring series events continue.

MEN
1. Phi Delta Theta — 15%
2. Black Key Bulls — 15%
3. Sigma Phi Epsilon — 10%
4. Sigma Alpha Epsilon — 15%
5. Phi Gamma Delta — 10%
6. CSF — 5%
7. Phi Kappa Sigma — 5%
8. Beta Theta Pi — 5%
9. Phi Kappa Psi — 0%
10. Cutters — 10%
11. Sigma Chi — 0%
12. Delta Tau Delta — 0%
13. Evans Scholars — 5%
14. Sigma Alpha Mu — 0%
15. Sigma Pi — 0%
16. Dodd’s House — 0%
17. Forest — 0%
18. Wright Cycling — 0%
19. Lambda Chi Alpha — 5%
20. Delta Sigma Pi — 0%
21. Gray Goat — 0%
22. Delta Chi — 0%
23. Pi Kappa Phi — 0%
24. Alpha Epsilon Pi — 0%
25. Alpha Tau Omega — 0%
26. Delta Upsilon — 0%
27. Pi Kappa Alpha — 0%
28. Collins — 0%
29. Northern Indiana Cycling — 0%
30. Phi Kappa Tau — 0%
31. Kappa Delta Rho — 0%
32. Sigma Nu — 0%
33. Phi Sigma Kappa — 0%

WOMEN
1. Alpha Chi Omega — 20%
2. Melanzana — 25%
3. Kappa Alpha Theta — 5%
4. Cru — 20%
5. Alpha Gamma Delta — 0%
6. Phi Mu — 20%
7. Ski — 0%
8. Wing It — 5%
9. CSF — 0%
10. Teter — 0%
11. Delta Gamma — 5%
12. Theta Phi Alpha — 0%
13. Army — 0%
14. Alpha Omicron Pi — 0%
15. Gamma Phi Beta — 0%
16. Kappa Kappa Gamma — 0%
17. Collins Cycling — 0%
18. Air Force — 0%
19. Kappa Delta — 0%
20. Chi Omega — 0%
21. Delta Sigma Pi — 0%
22. Ride On — 0%
23. Mezcla — 0%
24. Zeta Tau Alpha — 0%
25. Alpha Xi Delta — 0%
26. Alpha Delta Pi — 0%
27. IU Nursing — 0%
28. Sigma Delta Tau — 0%
29. Alpha Sigma Alpha — 0%
30. Delta Zeta — 0%
31. Delta Delta Delta — 0%
32. Delta Phi Epsilon — 0%
33. Alpha Phi — 0%

Charles Scudder is a senior studying journalism at IU. He put off taking a required statistics course until the last semester of his senior year, so go easy on him if his math is screwy.

Advertisements

23 thoughts on “Quals position often not a perfect predictor of race success

  1. This site is great! I hope people start paying attention and I’ll spread the word!

    Every once in a while a fault at quals on a first run will put good teams in bad spots but they can get to the front in under 2-3 laps in about the first 21 positions.

  2. Chi Omega faulted on their 3rd attempt clearly and should not have been allowed to take the spot they did. Tom S. loves to complain when he thinks other teams get an advantage. He should have called it on himself! Riding in the gutter with a video of the infraction should have been an automatic 4th attempt!!!

    • Wow riding in the gutter and not having it called is a big deal. I heard she just trod onto the gutter, which can be more difficult to spot but riding in the gutter must be called!
      Any chance that Pi Phi start legal action against IUSF and Tom surely would have done if the situation were reversed?

      • You “heard” you actually “heard”? If you are going to spread a lie, go big or go home. Why not just say that the entire attempt was run on the grass, and that he tackled the judges to keep them from making that call. And why stop there, certainly you must have “heard” that he is responsible for global warming, the disappearance of Flight 370, and the collapse of Bitcoin.

        If you have a beef against him, go to him or IUSF with proof, instead of posting what you “heard” here.

  3. Love the new site!

    Here’s my two cents on quals:

    There are certain perks to qualing well, such as a spot near the front to help avoid early crashes.
    But there are teams that put ALL their energy and practice time into quals who don’t have a chance on race day.

    A good example of this is SNU a few years ago, who had the pole and were a complete non-factor on race day. CSF has also had similar results that did not translate on race day. I predict SAE to be this year’s version of those teams

    I feel it is more important to have experienced riders who can eat up laps but also lay down quick, short sets towards the end of the race

    • Agree with your thoughts about quals not being a particularly good predictor of final placings. Look over the results from the last 25 to 30 years. (The race has not fundamentally changed for decades.)
      I think qual position is a slightly better predictor now that the race prohibits/discourages Cat 1 and 2 cyclists from participating. There were riders in the 80s who could ITT and sprint as fast as any rider today (except maybe Young) but were not very good at getting on the bike and getting the bike up to speed – they were less hybrid athlete cyclists and more pure cyclists.
      Teams used to routinely place top ten and even win (Cinzano) using only 3 or mostly 2 riders.

  4. Quals indicate very little in terms of how teams will fare on race day. Look to the other spring series events, especially team pursuit to see the strongest and deepest teams

  5. Quals mean little in predicting the race. It is Team Pursuit that will give everyone the best idea. Team Pursuit will give you your top 5 teams, top 10 teams, etc. Only crashes will disrupt those results.

  6. Team Pursuit and count of top 20 riders will give you the information you need to predict the race.

    Also the when the ITT time winner, is on the final lap, the win about 95% of the time. Last year was an exception but all the pasy 8 winners were winner of ITT.

  7. There was a good streak of ITT winners winning but, the % over 20 years is much lower than that. 2013: no, 2012: Yes, 2011: Yes, 2010: yes, 2009: yes, 2008: no 2007: No (teammate won though) 2006: yes, 2005: no 2004: yes, 2003: yes, 2002: yes, 2001: yes, 2000: yes, 1999: no, 1998: no, 1997: no, 1996: no (teammate won), 1994: no. 8 out of 19 for men or 10 of 19 teams. I don’t have 1995 results for Fiji.
    For women: 13:no, 12: no, 11: yes, 10, yes 09: no, 08, no, 07: no, 06: yes, 05: no, 04:no, 03:no, 02:yes, 01: no, 00:yes, 99:no, 98: no. So only 5 out of 16 women won Itt’s and the race in the same year.

    • Women’s race is a whole different race and strategy.

      Do you think the men’s race is the same as it is in the 1990’s? For some reason I feel the race has changed. Technology, now teams can watch all the previous races. Scouting. New information. I think the race is done differently than back in the day.

      Plus the data on the IU website starts to get pretty bad after 2005. Not good data.

      11/14 is pretty good, 75%. Yes, I count the teammate. Pretty telling. When I said 95% meant last 15 years BTW, as a ball park. Not the entire history. The sample size I spoke of was last 14 years. Not last 20.

      Since 2006 the top 4 finishers of each race has primarily.. (I projected it out in excel)

      finished top 5 in team pursuit. The exceptions were due to crash, or poor weather conditions.
      Had a a top 10 ITT rider.
      Winner primarily had #1.
      Had at least 2 riders in top 20, Many had 3.
      Had 3rd rider primarily finished top 30.

      Quals do not mean that much in grand scheme of things, as long as your pit is at the end of the straight away.

  8. The race has not changed much for the men since the 90’s. Their was more scouting in the 70’s then there is now. The women’s race is half the distance so you can use more energy on your sets and more burnouts are chased out but other than that, it’s not that much different. Technology? Are you 12? The race videos have been on videotape for a long, long time. Watching race tape has been going on for 3+ decades. The data on the site is correct. All you have to know is the winner of Itt’s and the winner of the race. A teammate winning the sprint could mean the ITT’er had a less chance of winning and that’s why he’s there. There is obviously a link to a great TT’er having a high win %, but too many other factors are involved and guys like Young that win 3 in a row are exceptions to the rule as he was overly dominant. It’s definitely a better predictor of quals but most of what you said was wrong. The 90’s weren’t that long ago and only seem that way if you’re too young to remember them. There was more notable cycling stars then than now. Young being one of the only stars of today that did anything outside of Little 5 worth noting other than a few local pros. Fun Fact. Eric Young holds the world record for a 500 meter sprint on rollers beating Olympians Taylor Phinney (BMC Racing) and Timmy Duggan
    Read more at http://velonews.competitor.com/2012/08/news/eric-young-sets-two-world-records-in-roller-sprint-fund-raiser_234626#PKYuj7cj5zwsGFoR.99

  9. Yes, I am too young to even know about those races. My knowledge is 2000’s to present. Sure those races were on VHS back than using one camera. Took weeks for the VHS’s to be produced. Well you seem knowledgeable.

    Question or would you rather rather scenario – Having Ideal stettings and it’s a 4 team sprint for lap 180 on. What do you think is the best exchange rate strategy.Starting Lap 180 on, all teams are in full bust mode.

    A. Cutters Exchange on Laps 180,184,188,192,196,198. (4 Lap intervals and end on 2)
    Winner for years
    B. Phi Delta Theta exchange on laps 180,182,184,186,187, etc. (Joke that they exchange every 2 laps).
    Second Place
    C. Beta Theta Pi. 179, 185,190, 194, 197, 199 (6-5-4-3-2-1).
    Won Last Year
    D. Delts 180-184-188- Finish
    Won 2012
    E. Other… maybe a breakaway early from the peleton….

  10. Qualifying position is malarkey in my idea. I think those percentage is a bunch of hot garbage data. Just a miss use of data. Rather than pole qualifying position. Think pit location is a better allocation of that data. You want your pit on the long back stretch and right after a corner. So you can get a better lead out for exchanges. FIGI probably hasn’t won for years cause they always take that same bad pit location, where all the accidents happen late in the race. High traffic area.

    Pit location would be a more effective use of that data in regards to qualification and overall finish.

    Your data is better off using team pursuit rankings rather than Qualification ranks. Telling you, it’s better for predicting top 4. Qualification data is bad data to use.

  11. Data Analytics,

    In 1996 and 1999 they did the exchange fire drill near the end of the race and it worked. usually the do 1 rotation of 2 laps each with really fast exchanges. If you run them through a 2nd round they will be extremely tired and will have a much bigger chance of dropping an exchange. A 8 lap, 4 lap, 4 lap, 2 lap, 2 lap might work. Exchanges take time and the more you do, it’s a disadvantage. You can usually get through 1 round ok, but then it’s diminishing returns. It’s also very risky and it’s usually a team that can’t win the traditional way of setting up a sprinter for the last 5-10 laps.

  12. Pingback: Team Pursuit gives better idea of race winner, final probability | 33to1

Join the Discussion

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s