From bobg@Radix.Net Mon Jul 12 07:33:54 EDT 1999
Having been tempted by Mike Tennant (it's all his fault!) I
went back to looking at the progression of world records (olympic times
rather) to see what sorts of times might be conceivable, and which might
well be beyond the realm of possiblity.
Challenging 'romantic' times for the next 80 years, in order of 'ease':
2:00:00 Marathon (5.7%)
1:40 800m -- the century (1.11%)
12:25.7 5k -- 4 minute mile pace (1:39.3 800) (1.9%)
3:20 1500m -- double century (1:38.5 800) (3%)
25:00 10k (1:36.0 or 1:38.5 800) (5.5%)
3:35 Mile (1:36.4 800) (3.8%)
You'll note that I'm looking toward pretty large improvements at 10k and
marathon, less at 1500m/mile, and very little at 800 and 5k. The latter
is a special case, more later.
Times I'll call 'impossible' (barring drugs or unexpected revolutions)
Outrageously impossible My going out on a limb:
40 400 42
1:35 800 1:37.9
3:10 1500 3:18.7
3:25 Mile 3:33.1
12:00 5k 12:15.2
24:00 10k ---> Revolutions _expected_
1:50:00 Marathon ---> Revolutions _expected_
I took the olympic gold medal times from 1900 to 1996 from an almanac.
The first modern olympics were in 1896, but it is quite obvious that the
times from that year were not (aside from the marathon) at all
comparable to the rest of the olympics. For example, 1896 100m time was
12.0, 1900 was 11.0. 1896 800m time was 2:11, while 1900 was 2:01. If
anyone knows of the history of the olympics well enough, it'd be
interesting to hear why where were such drastic changes (10% or so at
most distances in the 4 years, while it took the next 80 to make that
much change again) from 1896 to 1900, it would be interesting.
The 5k and 10k were not run in the olympics until 1912, so
I took that as my reference year. All percentage change in times are
relative to the 1912 times (100m to marathon):
10.8, 21.7, 48.2, 1:51.9, 3:56.8, 14:36.6, 31:20.8, 2:36:54.8
Current Olympic records:
9.84, 19.32, 43.49, 1:42.59, 3:32.53, 13:05.59, 27:07.34, 2:09:21.0
Percent improvement relative to 1912:
8.9, 11.0, 9.8, 8.3, 10.2, 10.4, 13.5, 17.6
The eagle-eyed, however, saw those Olympic records, and noticed that
although the sprint distances were at (100, 200) or near (43.49 was
only 0.2 off) world record pace for the time, the longer distances,
especially 1500 and 5k, weren't particularly close. If we compare the
current world records to the 1912 olympics (secondary assumption being
that 1912 olympic times were close to the world records at the time),
then the improvements are:
9.4, 11.0, 10.2, 9.8, 13.4, 13.9, 16.3, 19.6
Which looks like fair reason to split 100-800 to one group and 1500-
marathon to the other.
The 800m:
The 800 has an interesting history, though, and probably useful on the
guessing of the future of the 1500/mile record. 1908/12 turns out to
have been a transition olympics for the 800. Before that time, the 800
was markedly slower than the 400. And since then, the ratio has been
relatively fixed. Secondarily, the 1908 olympics 800 marks the time
that the 800 left the realm of 'high school meet' (modern school) and
to 'pretty good at large scale modern meet'. -- At least I think a
1:52-53 would do pretty well these days even at a large open meet. We
can also see that the 800 world record now is _still_ slower than double
the olympic time of 100 years ago at 400 m.
Year 400 800 Ratio
1900: 49.4 2:01.2 2.45
1904: 49.2 1:56.0 2.36
1908: 50.0 1:52.8 2.26
1912: 48.2 1:51.9 2.32
xx
1968: 43.86 1:44.3 2.38 (Mexico City)
1984: 44.27 1:43.00 2.33
1996: 43.49 1:42.59 2.36
WR 43.29 1:41.11 2.34
800m Speculation:
Over the last 30 years, the world record at 400 has scarcely moved (1.3%
of its 1968 value), and althought the 800 record has changed, it is more
a matter of dropping back to its historic 2.33 or so ratio to the 400m
record. The 800m record itself has proven quite resistent to change,
having lasted from 1981 to last year. The 400m WR is over 10 years old
as well. Given these, and some guessing: The 40 second 400m is probably
not going to happen in at least the next 80 years, and to go out on the
limb a bit, 42 seconds either. For 800m at historic ratio to 400m, this
suggests 1:33.2 is ruled out firmly, and 1:37.9 is unlikely. 1:40 (100
second) would take about a 42.9 400m world record concurrently. That
looks like a pretty healthy challenge, but not impossible in the next 80
years.
Now let's try the 800 and 1500:
Year 800 1500 Ratio
1900 2:01.2 4:06.2 2.03
1904 1:56.0 4:05.4 2.12
1908 1:52.8 4:03.4 2.16
1912 1:51.9 3:56.8 2.12
xxx
1924 1:52.4 3:53.6 2.08
1932 1:49.7 3:51.2 2.11
xxx
1952 1:49.2 3:45.1 2.06
1956 1:47.7 3:41.2 2.05
1960 1:46.3 3:35.6 2.03
xxx
1976 1:43.50 3:39.17 2.12
1984 1:43.00 3:32.53 2.06
1996 1:42.59 3:35.78 2.10
WR 1:41.11 3:26.00 2.04
Take another look at 1960. The 1500 time (Herb Elliot) would have won
the last olympics (Morcelli in 1996) and several in between, while the
800 time (Peter Snell) would have been well back of the winner in '96
(Vebjoern Rodal). This accords with the earlier observation that the
olympic times at 1500 are further off the WR than the times at 800m.
Reasons for the discrepancy are welcome. I take it that at the shorter
distances, in order to win you have to run pretty close to your best.
1500 and up can be won with times well off your best, particularly 1500
and 5k. Be that as it may, we can see that early in the century, the
1500 was about 2.12-2.16 times the time in 800m, while it is now (when
run full up) 2.03-2.06 times the 800m time. Something important was
learned between the 1900-1932 span and 1952 or so about how to run the
1500. Since 1952 or thereabouts, times at 1500 have dropped because
people run the 800m faster. The 800, if I had better records, I
suspect went through its corresponding change relative to the 400
between the 1880's and 1910.
Note that on speed, the current best ratio of, call it, 2.03 is 1.08
times slower, versus 800 to 400 which is 1.17 times slower at the
longer distance. The early-century 2.12 ratio is 1.13 times slower.
Roughly, the gap between 1500 and 800 used to be almost as large as
between 800 and 400.
1500m Speculation:
The ratio between 1500 and 800m times is probably (at 2.03) as low as it
can go barring a revolution in running such as occurred between 1932 and
1952. With 1:40 as an agressive goal for the 800, 3:23 would be equally
agressive for 1500. I'm taking 1:35 as unreachable, so 3:13 would also
be. (Being over precise). Romantic goal for 1500: 3:20 (double
century) Impossible 3:10. Finally, for the mile, to take even pace from
the 1500 at 3:23 gives 3:37.8 for the mile. No romance there, or
anywhere close. For a highly optimistic ultimate goal, 3:30 on the
mile. Forget 3:20. Even 3:35 might ultimately be too fast. At 3
seconds/decade, the 3:35 'wall' would come up in 30 years.
Looking at the 5k:
This one was very strange. Having seen the 800 and 1500 go through
periods of revolution, and we'll see the 10k and marathon also had
revolutionary periods, the 5k has no revolutionary span. Times on the
5k have always (since 1912 olympics) been about 3.70 times the time at
1500. The only marked exceptions were 1968 (Mexico city) when the 1500
was won in a new olympic record time, and the 5k was disastrously slow
(as was the 10k), and 1976 when the 5k was won in a good time (new
olympic record) while the 1500 was won in a very slow (for olympics!)
time. The current world record is 3.69 times the time of the 1500 WR.
This surprised me, after all our news and talk of the fantastically fast
times now being run in the 5k. All that has been happening is that the
5k times have been dropping back to their historic relation to the 1500.
13:00, apparently, was a barrier that could have been crossed any time
after the 1500 record got below 3:30(.8). 4 minute miles for 5k
(recalling that the 2 mile record is just barely under 8:00 at the
moment) is 12:25.7, which corresponds to 1500 record time of 3:21.5,
then 800 of 1:39.3. Both look reachable, though it won't be easy or
soon. But romantic goal for 5k of 4 minute miles average. Impossible:
12:00 (3:15 1500). [Note: This one isn't as far off as some others I
labelled impossible. This one is more throwing down the gauntlet.] The
3.70 ratio corresponds to slowing down by a factor of 1.11, which is
quite high compared to the 800-1500 deceleration (1.08) or even the
400-800 (1.17).
Considering 10k:
At 10k, there is a fairly continuous progressive revolution from its
inception to about 1972. Early on, the 10k took about 7.9 times as long
as the 1500. This dropped to about 7.8 by 1924, where it held
relatively fixed (barring outliers like 1936, where the 1500 improved a
lot, and the 10k was slower than the previous olympics) until 1972,
where it dropped again to about 7.7, where it has remained. The current
world record at 10k is 7.68 times the 1500m record, 2.08 times the 5k
record. The slowdown between 5k and 10k (keeping in mind that it is
common for the same person to be running both, unlike 1500-5k, or even
any more the 800-1500) is only 1.04. This is amazing. Future of 10k: If
the revolution has halted, which there is no current sign for, then the
3:20 1500 should lead to a 25:40 10k. 26:00 should fall around the time
that 1500 reaches 3:22. To go out on the limb slightly, I'll call for a
25:00 10k. That would take either a 3:15 1500 (relatively unlikely) or
a continued slight reduction in the ratio between 10k and 1500 times,
down to 7.5 for a 3:20 1500. Since we've already seen reduction from
7.9 to 7.7, a continued drop to 7.5 doesn't seem outrageous. That would
secondarily (assuming the 5k ratio didn't change) drop the 10k to 5k
slowdown to only 1.015. So perhaps any time after 5k gets to 12:30
(from its current 12:39.36), the 25 minute 10k is an open target. Note
that this is just over 4 minute mile pace. Impossible: 24:00. The 10k
times have improved so much, I can't even bring myself to rule out 4
minute miles for 10k.
The Marathon:
Here is the most dramatic change in times. From 1900 to the current
olympic record, the marathon time has dropped by over 30%! From 2:59:45
to 2:09:21 ! The pattern is one of great lurches forward. 1900-1908
was won in about 3 hours (even 3:28 in 1904!). 1912, suddenly dropped
it to 2:36. Plus or minus about 5 minutes, this is where it remained
until 1952, where it dropped to 2:23 and began a steady decline to 2:10
in 1976, about which time it has remained the last 20 years. In terms
of the 1500, the ratio dropped from about 44 times to now 36.5 to 37
times. The ratio for the olympic records is 36.5, and 36.9 for the
world records. The distance ratio is 28.1, so the slow down is 1.31
w.r.t. the 1500, over half of which occurs between the 1500 and 5k
(taking 36.7 as the current proper ratio). Relative to 10k, the
slowdown is 1.13. (Taking 4.77 times as long as 10k). For a 26 minute
10k, that's a 2:04 marathon. The 2:00 marathon corresponds to about a
25:15 10k, or 3:16 1600, at current ratios. For modest continuations of
the drop in ratio, say to 36:1 between marathon and 1500, the 2:00
marathon would only need a 3:20 1500. For the current world record
1500, the 2:00 marathon requires a drop to 35:1. Given the history of
the marathon, even this doesn't seem unlikely in the relatively near
(10-20 year) future. It is hard to come up with an impossibility level
for the marathon since it has already shown such improvement.
Nevertheless, 1:45 as wholly impossible, and 1:50 I'll put up as nearly
so.
How to really destroy the predictions:
As already noted in the 10k and marathons guesses, if the race is run
faster relative to the shorter races, then you can see some rapid
changes in the records at this distance. I've left some leeway in my
guesses for those distances because of this. Still, the best place to
look for an improvement in training methods would be the distances which
show a large slowing relative to the shorter distance. The slowing
w.r.t. distance is:
1.17 400-800
1.13 10k-Mara
1.11 1500-5k
1.08 800-1500
1.04 5k-10k
The deceleration between 400 and 800 seems expectable. That's where
you're turning over from anerobic to aerobic running (50/50 at the 800).
The 10k to Marathon may well be reflecting the change in energy storage
and utilization. But 1500 to 5000m ... What's going on here? Perhaps
there isn't anything especially physical going on and it is simply a
matter (simple for us spectators) that the 5k isn't being trained as
well as it could be or we aren't recruiting the best possible people to
it. If the 5k slowed down only by 8%, rather than the current 11%,
relative to the 1500, then ratio would be 3.6 instead of 3.7, and the WR
would immediately drop to 12:22. Another 18 seconds may be there for the
taking (incidentally easily passing 4 minute mile pace) 'today'. If it
were a difference in training, rather than recruitment, then today's
13:00 5k runners could have been running 12:40's.
I'm not saying that this kind of improvement actually _is_ possible.
There is every sign that it isn't. But if my listed 'impossibilities'
were to be broken, then I think 5k is the distance most likely to do it.
An alternate route to vast improvements is psychological. It is routine
for the olympic (or world championships) to run 1500 and 5k at times
well off the world record, and even more common/further off WR at other
meets. Part of this is a necessary part of training. But at least at
world or olympics, it is not training. These are the meets you were training
_for_. A possible change would be for these runners to race more often full
out. (And the olympics may need some structural change to make it possible
for runners to _be_ full out at the finals.) Some of the strategy/tactics
aimed at winning may be detracting from running fast.
--
Robert Grumbine http://www.radix.net/~bobg/ Science faqs and amateur activities notes and links.
Sagredo (Galileo Galilei) "You present these recondite matters with too much
evidence and ease; this great facility makes them less appreciated than they
would be had they been presented in a more abstruse manner." Two New Sciences
Return to main running page
Return bobg main page
Send message to Robert Grumbine