Debate Sabermetrics: A Statistical Analysis of the Season's Top Individual Speakers

Ever wonder who has been the best speaker in the country over the course of the season? While the National Debate Coaches’ Association Baker Award recognizes the year long excellence of teams, no such award exists to recognize the excellence of individual debaters. The following is an attempt to fill that gap with a statistical analysis of the speaker points and speaker awards earned by the nation’s top policy debaters during the 2009-2010 season.

Methodology

The performance of individual debaters is assessed in each round through the assignment of speaker points and at each tournament through the tabulation of speaker awards. A statistical analysis of the season’s top individual speakers can therefore be conducted based on either or both of these criteria: total points and speaker awards. In order to determine this season’s best speakers, these criteria were analyzed separately; it is left up to individual interpretation how much weight to give to one as compared to the other.

Approach #1—Speaker Points

Speaker points were collected for the eight octafinals bid tournaments—Greenhill, St. Mark’s, Glenbrooks, Blake, MBA, Emory, Harvard, and Berkeley. Each debater’s total number of speaker points at each tournament was used for these calculations, not their adjusted (“high-low”) totals.

Points were calculated for all debaters that finished in the top ten at one or more octafinals bid tournaments. The total number of speaker points accumulated by each of these debaters was divided by the number of rounds in which they competed. The maximum number of rounds that any one debater could have participated in was 45; Andrew Markoff of Bronx Science and Christian Steckler of Bishop Guertin were the only two debaters to achieve that total (Laura Johnson of Saint Paul Central competed in 44).

Points for the St. Mark’s Heart of Texas Invitational were normalized from the 100 point scale to the traditional 30 point scale. This was more difficult than first thought because the 100 point scale is not just the 30 point scale multiplied by 3.333; the numbers had to be manually translated using the “RKS” scale so that 85 = 28.5, 90 = 29.0, etc. The normalized numbers were calculated to the second decimal place (for example, Ellis Allen’s 554 (92.33 points per debate) was translated to 29.23).

Because data for the St. Mark’s tournament may be unreliable, a separate ranking was calculated based only on the other seven tournaments.

Highest Average Speaker Points—Minimum 12 Rounds

  1. Andrew Markoff—Bronx Science (28.87733333, 45 rounds)
  2. Layne Kirshon—Kinkaid (28.86216216, 37 rounds)
  3. David Mullins—Westlake (28.85625, 32 rounds)
  4. Ellis Allen—Westminster (28.83025641, 39 rounds)
  5. Will Thibeau—Glenbrook South (28.75263158, 38 rounds)
  6. Andrew Arsht—Rowland Hall-St. Mark’s (28.73684211, 19 rounds)
  7. Thomas Hodgman—Pembroke Hill (28.73538462, 13 rounds)
  8. Matthew Pesce—Woodward (28.69210526, 38 rounds)
  9. Daniel Taylor—Westminster (28.69179487, 39 rounds)
  10. Alex Miles—St. Mark’s (28.66666667, 18 rounds)

Highest Average Speaker Points—Minimum 24 Rounds

  1. Andrew Markoff—Bronx Science (28.87733333, 45 rounds)
  2. Layne Kirshon—Kinkaid (28.86216216, 37 rounds)
  3. David Mullins—Westlake (28.85625, 32 rounds)
  4. Ellis Allen—Westminster (28.83025641, 39 rounds)
  5. Will Thibeau—Glenbrook South (28.75263158, 38 rounds)
  6. Matthew Pesce—Woodward (28.69210526, 38 rounds)
  7. Daniel Taylor—Westminster (28.69179487, 39 rounds)
  8. Reid Ehrlich-Quinn—Damien (28.664375, 32 rounds)
  9. John Baker—Westlake (28.65875, 32 rounds)
  10. Becca Rothfeld—Georgetown Day (28.62121212, 33 rounds)

Highest Average Speaker Points Not Including St. Mark’s—Minimum 12 Rounds

  1. Andrew Markoff—Bronx Science (28.84615385, 39 rounds)
  2. Layne Kirshon—Kinkaid (28.80645161, 31 rounds)
  3. David Mullins—Westlake (28.78846154, 26 rounds)
  4. Ellis Allen—Westminster (28.75757576, 33 rounds)
  5. Andrew Arsht—Rowland Hall-St. Mark’s (28.73684211, 19 rounds)
  6. Will Thibeau—Glenbrook South (28.734375, 32 rounds)
  7. Alex Miles—St. Mark’s (28.66666667, 18 rounds)
  8. Danny Bernick—Henry Sibley (28.64285714, 14 rounds)
  9. Matthew Pesce—Woodward (28.625, 32 rounds)
  10. (tie) Becca Rothfeld—Georgetown Day (28.62121212, 33 rounds)
  11. (tie) Daniel Taylor—Westminster (28.62121212, 33 rounds)

Highest Average Points At A Single Tournament

  1. 29.23: Ellis Allen—Westminster (St. Mark’s)
  2. 29.16: Andrew Arsht—Rowland Hall-St. Mark’s (Berkeley)
  3. 29.15: (tie) David Mullins—Westlake (St. Mark’s)
  4. 29.15: (tie) Layne Kirshon—Kinkaid (St. Mark’s)
  5. 29.08: (tie) Andrew Markoff—Bronx Science (St. Mark’s)
  6. 29.08: (tie) Daniel Taylor—Westminster (St. Mark’s)
  7. 29.05: Matthew Pesce—Woodward (St. Mark’s)
  8. 29.01: Thomas Hodgman—Pembroke Hill (St. Mark’s)
  9. 29.00: (tie) Andrew Markoff—Bronx Science (Blake, Emory)
  10. 29.00: (tie) Ellis Allen—Westminster (Blake)
  11. 29.00: (tie) Layne Kirshon—Kinkaid (Berkeley)

Highest Average Points At A Single Tournament—Not Including St. Mark’s

  1. 29.16: Andrew Arsht—Rowland Hall-St. Mark’s (Berkeley)
  2. 29.00: (tie) Andrew Markoff—Bronx Science (Blake, Emory)
  3. 29.00: (tie) Ellis Allen—Westminster (Blake)
  4. 29.00: (tie) Layne Kirshon—Kinkaid (Berkeley)
  5. 28.92: (tie) David Mullins—Westlake (Glenbrooks)
  6. 28.92: (tie) Andrew Markoff—Bronx Science (Glenbrooks)
  7. 28.92: (tie) Becca Rothfeld—Georgetown Day (Glenbrooks)
  8. 28.91: (tie) Layne Kirshon—Kinkaid (Emory)
  9. 28.91: (tie) Matthew Pesce—Woodward (Emory)
  10. 28.91: (tie) Campbell Haynes—Montgomery Bell (Berkeley)
  11. 28.91: (tie) Mario Feola—Rowland Hall-St. Mark’s (Berkeley)
  12. 28.85: (tie) Will Thibeau—Glenbrook South (Blake)
  13. 28.85: (tie) Andrew Markoff—Bronx Science (Harvard)
  14. 28.83: (tie) David Mullins—Westlake (Greenhill)
  15. 28.83: (tie) Alex Miles—St. Mark’s (Emory)

Approach #2—Speaker Awards

While an assessment of a debater’s speaker points over the course of the season is undoubtedly one of the most effective measures of their performance, the resulting data does not account for differences in relative point scales between tournaments. If the judging pool at a given tournament is more generous than average, the points that a debater accumulates at that tournament may be artificially inflated. (Adjusting individual speaker points based on the mean or median speaker point distributions at each tournament might be an effective way to correct this but it is beyond the scope of this author’s mathematical skills.)

Instead of measuring each individual’s total speaker points, this alternative approach measures the value of the speaker awards they earned at each tournament compared with their peers. Once again, data was collected from the eight octafinals bid tournaments—in total, 37 debaters earned at least one top ten speaker award and 20 earned more than one.

Values were assigned to each speaker award on a scale from 1 to 10: the top speaker received 10 points, the second speaker received 9 points, and so on. The total “share” of the speaker award total was then calculated both as a gross value (total speaker award share) and as a net value (total speaker award share divided by the number of tournaments the debater attended). Winning the top speaker award at seven tournaments (the maximum that a single debater could attend) would earn an individual 70 points.

Most Top Ten Speaker Awards

  1. Andrew Markoff—Bronx Science (7)
  2. Layne Kirshon—Kinkaid (6)
  3. (tie) David Mullins—Westlake (5)
  4. (tie) Ellis Allen—Westminster (5)
  5. (tie) Will Thibeau—Glenbrook South (5)
  6. (tie) Ben Chang—Edgemont (3)
  7. (tie) Daniel Taylor—Westminster (3)
  8. (tie) Ira Slomski-Pritz—New Trier (3)
  9. (tie) Matthew Pesce—Woodward (3)
  10. (tie) Misael Gonzalez—Whitney Young (3)

Highest Speaker Award Share—Total

  1. Andrew Markoff—Bronx Science (51)
  2. Layne Kirshon—Kinkaid (48)
  3. David Mullins—Westlake (32)
  4. Ellis Allen—Westminster (31)
  5. Will Thibeau—Glenbrook South (26)
  6. Matthew Pesce—Woodward (20)
  7. Andrew Arsht—Rowland Hall-St. Mark’s (19)
  8. (tie) Daniel Taylor—Westminster (17)
  9. (tie) Alex Miles—St. Mark’s (17)
  10. Kevin Hirn—Whitney Young (16)

Highest Speaker Award Share—Average Per Tournament Attended

  1. Layne Kirshon—Kinkaid (8.00 average shares, 6 tournaments)
  2. Andrew Markoff—Bronx Science (7.29 average shares, 7 tournaments)
  3. David Mullins—Westlake (6.40 average shares, 5 tournaments)
  4. Andrew Arsht—Rowland Hall-St. Mark’s (6.33 average shares, 3 tournaments)
  5. Alex Miles—St. Mark’s (5.67 average shares, 3 tournaments)
  6. Ellis Allen—Westminster (5.17 average shares, 6 tournaments)
  7. Will Thibeau—Glenbrook South (4.33 average shares, 6 tournaments)
  8. Matthew Pesce—Woodward (3.33 average shares, 6 tournaments)
  9. Becca Rothfeld—Georgetown Day (3.00 average shares, 5 tournaments)
  10. Daniel Taylor—Westminster (2.83 average shares, 6 tournaments)

Conclusions

Regardless of one’s preferred approach, Bronx Science’s Andrew Markoff and Kinkaid’s Layne Kirshon have been the season’s top individual speakers. Markoff has earned the highest average speaker points, the most speaker awards, the highest gross share of speaker awards, and three of the top ten highest totals at a single tournament (four if St. Mark’s is excluded). Kirshon, on the other hand, has earned the highest net speaker award share (averaging third speaker at 6 tournaments) and is second in highest average speaker points, total speaker awards, and gross speaker award share while posting two of the top ten highest totals at a single tournament.

So who wins? After the results of the NDCA Championships and the Tournament of Champions are included in the data, we’ll find out.

Please report any errors to the author either via email or by posting a comment.

29 thoughts on “Debate Sabermetrics: A Statistical Analysis of the Season's Top Individual Speakers

  1. Bill Batterman Post author

    If there is sufficient demand and if I can get all of the data, I might add quarterfinals bid tournaments to the calculations. For now, however, just keep in mind that this analysis doesn't include some pretty competitive tournaments—this isn't a list of THE BEST SPEAKERS IN THE COUNTRY PERIOD, it's just a list of the best speakers based on this specific dataset (a meaningful one, I think, but nonetheless an incomplete one).

    Also, I think this earns me first ballot membership in the Debate Geek Hall of Fame (should such an institution ever be created, and assuming I haven't already put together a first-ballot-worthy career). Pathetic? Guilty as charged. 🙂

  2. Bill Batterman Post author

    @Layne Kirshon

    Just sample size—12 rounds is only two tournaments so the data is less reliable. The 24 round threshold was chosen because it meant that a debater had attended 4 tournaments, something most of those who placed in the top ten at one or more octafinals bid tournaments had done. It's up to the individual to decide which cutoff makes more sense to them… the top five are the same either way.

  3. Bill Batterman Post author

    @Layne Kirshon

    That's next on the list… I'm getting close and will post when I'm finished. I'm also working on some "Baker+" stats (head-to-head, winning percentages, etc.). Getting all of the data has proven difficult, but it'll get done eventually.

  4. Roy Levkovitz Post author

    Points are absurdly high. Look at what the top 15 speakers are averaging. Its nuts. Top 20 speakers basically get 1 of 3 points, 29.5, 29, 28.5.

    While the St. Marks system's average was absurdly high (due to first time use, people not understanding what high 90s meant) I think if we keep using that we have the best chance of both stabilizing and differentiating between speakers.

    In the short term what do people think about the no 2 speakers can get the same points thing they did at St. Marks. It would effectively prevent judges from dropping block 28.5s and actually differentiate debaters. I think it would effectively lower points in the time we wait to transition to the 100 point system.

    Even if points are high in the system, the ability to differentiate means that you can noticeably see a difference between debater X and Y, where now with the top 20 is like 2 points and some high low random number nonsense

  5. Scott Phillips Post author

    Roy,

    How does your system deal with debates where all 4 participants were perfect? Am I correct in interpreting that you would only allow 1 30 and 1 29.5?? Giving the 4th speaker a 28.5 for a perfect performance seems crazy to me.

  6. Layne Kirshon

    @ bill's roy joke: haha wonderful

    although, roy, i don't really agree – if you look at the pure averages markoff is averaging around a 28.8. if a 29 performance is fantastic and reserved for someone who u think should be a top 5 speaker it seems ok that the best speaker in the country averages just below that.

  7. Bill Batterman Post author

    I thought it would be interesting to look at point inflation over the last few years; I used data from the 2003, 2004, 2005, 2006, 2007, 2008, and 2009 Greenhill tournaments. While this is only one tournament, it is probably a representative example because of its consistent field and judging quality. So what does the data show?

    The Top Speaker has posted (adjusted, high-low) totals (in chronological order) of 115.5, 115.0, 115.5, 116.0, 115.5, 115.5, and 115.5.

    The 5th speaker has posted totals of 114.5, 114.5, 114.5, 114.7, 115.0, 114.5, and 114.5.

    The 10th speaker has posted totals of 114.5, 114.5, 114.0, 114.5, 114.0, 114.0, and 114.5.

    The 15th speaker has posted totals of 114.0, 113.5, 114.0, 113.5, 114.0, 113.5, and 114.0.

    The 20th speaker has posted totals of 113.5, 113.5, 113.5, 113.5, 113.5, 113.0, and 113.5.

    Speaker points have remained remarkably consistent. The average points for a Top Twenty speaker have been 114.3, 114.2, 114.2, 114.3, 114.3, 114.1, and 114.3.

    (This obviously does not address the average points assigned *overall*, only the points assigned to the top 20 (roughly 15%) debaters in the field.)

  8. Abhik

    What about Alex Resar from McDonough – He's done pretty well at a host of tournaments including top speaker at the Harvard RR and the Harvard tournament. Would you be able to calculate the quarters tournaments soon seeing as many teams are unable to attend these big octo's tournaments yet there are still some very competitive debaters out there?

  9. Anonymous

    Great deal of self-interest involved here, but would you consider doing the same for top 20 speakers, or is it too much work?

  10. Layne Kirshon

    Abhik :
    What about Alex Resar from McDonough – He’s done pretty well at a host of tournaments including top speaker at the Harvard RR and the Harvard tournament. Would you be able to calculate the quarters tournaments soon seeing as many teams are unable to attend these big octo’s tournaments yet there are still some very competitive debaters out there?

    ya the problem is round robbins aren’t factored in

    bill, what do you think about adding round robbins associated w/octas bid tournaments – ghill Northwestern CPS and Harvard?

  11. John Haggerty

    It's interesting who is not listed here at all, yet is currently 1 or 2 in the Baker…..Richard Day. A top-notch debater in the country, he proves how points in many instances are largely derived from asthetic speaking ability. Obviously if you're good at debate, you'll be up here, I am not arguing that if you are a wonderful speaker but a horrible debater you'll be up here, but rather the converse(inverse?), that if you're not a wonderful speaker even if you are a wonderful debater it is difficult to be up here. It would be interesting to see what most judges feel is the largest determinent of speaker points. Also, just curious, but are there any trends between debaters who are stronger than their partners, in terms of speaker points and ranks, verses partnerships where the points are more even? It seems like two really strong partners can hurt each other in terms of speaker awards, not necessarily points, by out ranking each other in aff and neg rounds. For example, amazing debater x who is the 2N/1A who debates with good but young sophomore y will probably outrank debater y even in aff rounds, but two equally amazing debaters will mostly outrank the other in whatever debates they're the 2's in.

  12. Rishee Batra

    Yeah, that's not true. Miles (amazing debater x) usually outspeaks me (sophomore y) even in aff rounds

  13. Will Thibeau

    @John Haggery LOL. Points are indeed visually and auditorily aesthetically based.

  14. Bill Batterman Post author

    Two things:

    1. Richard is on both lists, just not in the top ten; he's:

    * 13th in Total Speaker Award Shares (tied with Campbell Haynes of MBA)
    * 16th in Net Speaker Award Shares per Tourney (with 1.83)
    * 17th in Average Points Per Round (28.61052632)
    * 15th in Average Points Per Round Not Counting St. Mark's (28.64285714)

    2. It is definitely true that 2N/1A speakers are overrepresented at the top of these lists compared with 2A/1N speakers. If 2A/2N speakers are removed, here's a list of the top Average Points Per Round for 2A/1Ns (and if I'm wrong about any of these debaters' speaker positions, let me know):

    1. Thomas Hodgman—Pembroke Hill (28.73538462)
    2. Daniel Taylor—Westminster (28.69179487)
    3. John Baker—Westlake (28.65875)
    4. Kaavya Ramesh—Chattahoochee (28.611875)
    5. Richard Day—Glenbrook South (28.61052632)
    6. Pablo Gannon—Damien (28.55)
    7. Flynn Makuch—Glenbrook North (28.54324324)
    8. Misael Gonzalez—Whitney Young (28.50410256)
    9. Mario Feola—Rowland Hall-St. Mark's (28.5)
    10. Vinay Sridharan—Glenbrook North (28.48894737)

    And here's a list of the top Total Speaker Award Shares for 2A/1Ns:

    1. Daniel Taylor—Westminster (17)
    2. Misael Gonzalez—Whitney Young (15)
    3. Richard Day—Glenbrook South (11)
    4. Ira Slomski-Pritz—New Trier (10)
    5. Mario Feola—Rowland Hall-St. Mark's (8)
    6. John Baker—Westlake (8)
    7. Kaavya Ramesh—Chattahoochee (6)

  15. Bill Batterman Post author

    Also, Glenbrook South DT has the second highest aff prelim winning percentage at octafinals bid tournaments—they are 17-1 (94.44%) and trail only Westminster AT (20-1, 95.24%). Westlake BM is third (14-1, 93.33%), Woodward PP is fourth (17-2, 89.47%), and Bronx Science EM is fifth (20-3, 86.96%).

    Damien EG is first in negative winning percentage—they finished the season 16-0. Glenbrook North MP also finished undefeated at 12-0, but their partnership switches for Greenhill and MBA complicate that data.

    Westminster finished with the highest total prelim winning percentage (37-2, 94.87%) with Glenbrook South (35-2, 94.74%) and Westlake (30-2, 93.75%) right behind.

  16. Bill Batterman Post author

    MP and SS are the normal partnerships, but Vinay debated with Flynn at Greenhill and MBA and Alexis debated with Pappas at Greenhill. As a result, the data for MP and SS excludes rounds that the individual debaters competed in, just not as a team.

    EDIT: So when Stratford MR defeated GBN at MBA, y'all defeated Flynn and Vinay, not Flynn and Pappas (and therefore that round does not count as a loss for GBN Flynn/Pappas).

  17. Kevin Hirn

    <blockquote cite="#commentbody-6221">
    Bill Batterman :
    Damien EG is first in negative winning percentage—they finished the season 16-0. Glenbrook North MP also finished undefeated at 12-0, but their partnership switches for Greenhill and MBA complicate that data.

    Not to call them out at all, because Reid and Pablo are really sweet, but I think there's a slight flaw in your methodology. We beat Damien when they were neg against us in quarters of Greenhill and Harvard (they did beat us in another round in which they were neg in the quarters of Glenbrooks, though).

  18. Bill Batterman Post author

    Jon Voss's reading comprehension skills are rubbing off on you, Kevin: *prelim* winning percentage. 🙂

    (I don't have sides for some of the elims so I haven't included that yet).

  19. Will Thibeau

    But, I agree, this is by no means a list of the best debaters in the country because of the sole fact Richard isn’t on any of the lists.

  20. Kevin Hirn

    @Bill Batterman

    …I had thought that it was possibly referring to prelim winning percentage, reread it, and then for some reason convinced myself that it wasn't. Maybe you were right about Jon Voss and our Districts tournament 😛 (just kidding of course).

Comments are closed.