Richard Crawford

Bravo, Rafał! There are a lot of interesting ideas in this article. Here are a couple of random thoughts:

In terms of the Principles (see my article from July this year if you are not familiar with them), this article presents very strong support for the first principle. It also suggests that meet organizers should not deviate from the defined order of events. [I'm thinking specifically of officials organizing combined event meetings that include a women's decathlon: Do they exchange the women's field events between the two days (day one having 100m, DT, PV, JT, 400m; day two having 100mH, LJ, SP, HJ, 1500m), or do they keep the same order that the men use? Can we usefully compare marks achieved the two different ways?] Further, it suggests that it is not useful for combined event statisticians to compile pseudo-decathlon scores based solely on the all-time best single-event marks for given athletes. [Perhaps these wouldn't be so problematic if the marks are limited to those achieved during decathlons.]

Decathlons entered to test the athlete's fitness (a common practice here in the USA in the early season) are not uncommon -- the athlete, for example, plans to contest only the first-day events or maybe only the field events. The question for the statistician becomes: Should these be treated the same as "full" decathlon? And how does one distinguish between an attempt at a full decathlon but truncated by injury from an intentionally truncated decathlon? And what if that affects his all-time best mark? Consider a case in which the athlete, say, sets his all-time best discus mark, no-heights in the pole vault, but then successfully completes the javelin throw and 1500m run with a total of 7500 points? Should that best mark be excluded from a summary of his career?

It makes no more sense to consider single-event performances in a decathlon individually than it does to consider the hop, step, and jump phases of a triple jump performance individually: Necessary for training and coaching, but little else.

There are, of course, athletes who have excelled at both decathlon and individual events simultaneously -- Simon Ehammer in the long jump currently, Tim Bright in the pole vault in the 1980's. But they are rare indeed.

Enough for now.

Again, good work, Rafał!

Richard Crawford

The 2001 edition of "IAAF Scoring Tables for Combined Events" includes the nine "guiding principles" that were used (or at least inspired) the 1984 tables currently in use. One of these is directly addressed by Mr. Snoch's articles and also by my own article from 2018 ("Are the Decathlon Tables Fair?") This is #2: "Results in various events should ... yield about the same number of points if the results are comparable as to quality and difficulty." Snoch also addresses #8: "It is desirable ... that the total scores using the new tables for the top world class athletes should remain approximately the same." (I think that his is a result of his methodology rather than an explicit goal.) But there is one principle which has been ignored -- #4: "It must be possible to use the scoring tables for beginners, juniors and top athletes." (How to apply this is quite unclear. It does not suggest, for example, how to accommodate the different hurdle heights or race lengths between juniors and seniors, much less implement weights.) It seems to me that basing the results exclusively on the upper levels of achievement skews them substantially. My article attempts to look at how well #2 is met at all performance levels, although it explicitly excludes data from non-senior performances. It asks whether an athlete who improves from one percentile to another is awarded with the same additional number of points regardless of the event. I have the advantage (perhaps dubious) of a database of about 100,000 usable (that is, with ten non-zero event scores) decathlon performances to work with, which allows me to subdivide them into various groups without losing too much statistical significance. This subdivision allows me to compare an athlete with others of comparable levels of achievement. But I made no attempt to propose modifications to the tables in use, much new tables altogether.