As someone started to do regressions, he's my suggestion how to approach it:
1. Establish the "human physiology" curve shape by regressing 1500m, 3000m, 5000m and 10000m world records in the aerobic range. You will derive the best theoretically possible 4000m time (4000max) from that non-linear curve
2. Map 3K and 5K PRs as % of best time for each runner. Say, Komen is 100.0% at 3000m and 100.6% at 5000m (the difference between his 12:39 and Cheptegei's WR). Then choose the middle point 100.3% of 4000max as his 4000m speed.
3. This method assumes that everyone's "ideal distance" is either 3K or 5K, but not 4K. If Komen, say, is uniquely built for 4K vs 3K or 5K, he could go higher than 100.3 midpoint.
So, I quickly mapped those world records and the curve type slightly deviates from linear towards exponential (as opposed to logarithmic).
What this means in practice - someone is supposed to run 4K 1.5-2 sec faster than the average of his 3K and 5K times.
Komen's 3-5K average is 10:00.2, so he was likely capable of high 9:58s
Bekele's average is 10:01.1. so he was likely capable of mid 9:59s
Haile's average is 10:02.2, so he was NOT likely to go under 10 mins
And of course the super spikes make this slightly easier than before. Hicham could likely do it in the super spikes. Hicham ran 4:44 for 2 and 7:23 for 3 and now with the super spikes he could likely run 9:59 for 4.
As someone started to do regressions, he's my suggestion how to approach it:
1. Establish the "human physiology" curve shape by regressing 1500m, 3000m, 5000m and 10000m world records in the aerobic range. You will derive the best theoretically possible 4000m time (4000max) from that non-linear curve
2. Map 3K and 5K PRs as % of best time for each runner. Say, Komen is 100.0% at 3000m and 100.6% at 5000m (the difference between his 12:39 and Cheptegei's WR). Then choose the middle point 100.3% of 4000max as his 4000m speed.
3. This method assumes that everyone's "ideal distance" is either 3K or 5K, but not 4K. If Komen, say, is uniquely built for 4K vs 3K or 5K, he could go higher than 100.3 midpoint.
So, I quickly mapped those world records and the curve type slightly deviates from linear towards exponential (as opposed to logarithmic).
What this means in practice - someone is supposed to run 4K 1.5-2 sec faster than the average of his 3K and 5K times.
Komen's 3-5K average is 10:00.2, so he was likely capable of high 9:58s
Bekele's average is 10:01.1. so he was likely capable of mid 9:59s
Haile's average is 10:02.2, so he was NOT likely to go under 10 mins
Cheptegi ran 2:02 first 800, then 10:03 last 4000. I’d have a hard time seeing him not able to break 10:00 in a one off.
World Record Progression, Last 100 Years: Paavo Nurmi - 1922 - 11:30.4 Paavo Nurmi - 1924 - 11:27.2 Paavo Nurmi - 1926 - 11:23.8 Janusz Kusocinski - 1932 - 11:18.2 Lauri Lehtinen - 1932 - 11:15.5 Gunnar Höckert - 1936 - 11:14.1 Taisto Mäki - 1939 - 11:09.0 Henry Kälarne - 1940 - 11:08.3 Gunder Hägg - 1942 - 10:59.7 Gunder Hägg - 1944 - 10:57.9 Chris Chataway - 1954 - 10:52.1 Vladimir Kuts - 1954 - 10:50.8 Sándor Iharos - 1955 - 10:50.6 Vladimir Kuts - 1955 - 10:47.2 Sándor Iharos - 1955 - 10:45.8 Gordon Pirie - 1956 - 10:42.0 Bob Schul - 1964 - 10:40.6 Ron Clarke - 1965 - 10:38.1 Ron Clarke - 1965 - 10:33.8 Michel Jazy - 1965 - 10:33.4 Kip Keino - 1965 - 10:30.7 Ron Clarke - 1966 - 10:28.3 Emil Puttemans - 1971 - 10:24.7 Lasse Virén - 1972 - 10:23.8 Emil Puttemans - 1972 - 10:22.0 Brendan Foster - 1974 - 10:21.6 Henry Rono - 1978 - 10:17.1 David Moorcroft - 1982 - 10:15.95 Said Aouita - 1985 - 10:14.14 Said Aouita - 1987 - 10:12.37 Moses Kiptanui - 1992 - 10:11.61 Haile Gebrselassie - 1994 - 10:09.78 Noureddine Morceli - 1994 - 10:09.28 Moses Kiptanui - 1995 - 10:09.26 Haile Gebrselassie - 1995 - 10:04.22 Daniel Komen - 1996 - 9:59.31 Kenenisa Bekele - 2004 - 9:58.79 Joshua Cheptegei - 2020 - 9:58.56
*Top-100 Performers, All-Time: 1. Kenenisa Bekele 9:58.79 2004 2. Daniel Komen 9:59.31 1996 3. Haile Gebrselassie 10:01.45 1998 4. Hicham El Guerrouj 10:01.76 2001 5. Mo Farah 10:04.60 2013 6. Selemon Barega 10:04.83 2018 7. Yomif Kejelcha 10:05.03 2018 8. Bernard Lagat 10:05.12 2005 9. Hagos Gebrhiwet 10:05.67 2018 10. Ali Saidi-Sief 10:06.13 2000 11. Dejen Gebremeskel 10:06.38 2012 12. Eliud Kipchoge 10:06.42 2004 13. Isaac Songok 10:06.49 2006 14. Saif Saaeed Shaheen 10:07.06 2003 15. Mohammed Mourhit 10:07.51 2000 16. Paul Tergat 10:07.62 1997 17. Yenew Alamirew 10:07.83 2012 18. Sileshi Sihine 10:07.91 2004 19. Isiah Koech 10:07.96 2012 20. Thomas Longosiwa 10:08.22 2012 21. Mohammed Ahmed 10:08.40 2020 22. Brahim Lahlafi 10:08.43 2000 23. Moses Kiptanui 10:08.58 1997 24. Tom Nyariki 10:08.64 1996 25. Salah Hissou 10:08.94 1999 26. Edwin Soi 10:09.12 2011 27. John Kibowen 10:09.27 1998 28. Noureddine Morceli 10:09.28 1994 29. Luke Kipkosgei 10:09.40 1998 30. John Kipkoech 10:09.63 2012 31. Moses Kipsiro 10:09.74 2007 32. Tariku Bekele 10:09.82 2008 33. Moses Masai 10:09.91 2008 34. James Kwalia C’Kurui 10:10.16 2004 35. Abdelaati Iguider 10:10.26 2015 36. Augustine Choge 10:10.47 2011 37. Daniel Kip. Komen 10:10.51 2007 38. Benjamin Limo 10:10.58 1999 39. Vincent Chepkok 10:10.61 2010 40. Caleb Ndiku 10:10.73 2014 41. Jakob Ingebrigtsen 10:11.11 2019 42. Dieter Baumann 10:11.24 1998 43. Paul Bitok 10:11.32 1996 44. Craig Mottram 10:11.40 2007 45. Muktar Edris 10:11.66 2017 46. Abderrahim Goumri 10:11.91 2005 47. Assefa Mezgebu 10:12.02 1998 48. Mulugeta Wendimu 10:12.03 2004 49. Ronald Kwemoi 10:12.19 2017 50. Albert Rop 10:12.23 2013 51. Saïd Aouita 10:12.37 1987 52. Ismail Sghyr 10:12.49 1995 53. Telahun Haile Bekele 10:12.57 2019 54. Khalid Boulami 10:12.60 1997 55. Paul Chelimo 10:12.64 2017 56. Galen Rupp 10:12.71 2014 57. Bob Kennedy 10:12.83 1996 58. Venuste Niyongabo 10:12.86 1997 59. Chris Solinsky 10:12.96 2010 60. Joseph Ebuya 10:12.98 2007 61. Sammy Kipketer 10:13.06 2001 62. Mark Carroll 10:13.08 1999 63. Richard Limo 10:13.15 2001 64. Abraham Chebii 10:13.34 2003 65. Shadrack Korir 10:13.52 2007 66. Bethwell Birgen 10:13.70 2013 67. Imane Merga 10:13.80 2010 68. Abraham Cherkos 10:13.84 2006 69. Lopez Lomong 10:13.85 2020 70. Lucas Rotich 10:13.87 2010 71. Hailu Mekonnen 10:13.96 2001 72. Matt Centrowitz, Jr. 10:14.01 2019 73. Evan Jager 10:14.09 2015 74. Gebre Gebremariam 10:14.13 2005 75. Joshua Cheptegei 10:14.22 2019 76. Million Wolde 10:14.37 2000 77. Birhanu Balew 10:14.49 2018 78. Philip Mosima 10:14.54 1996 79. Moses Mosop 10:14.55 2006 80. Paul Koech 10:14.66 1997 81. Matt Tegenkamp 10:14.68 2007 82. Ryan Hill 10:14.68 2016 83. Boniface Songok 10:14.71 2004 84. Mark Kiptoo 10:14.75 2010 85. Isaac Viciosa 10:14.89 1998 86. Hicham Bellani 10:14.92 2006 87. Dathan Ritzenhein 10:14.93 2009 88. Boniface Kiprop 10:14.99 2006 89. Leonard Komon 10:15.00 2009 90. Josphat Menjo 10:15.04 2010 91. Yusuf Biwott 10:15.05 2010 92. Abebe Dinkessa 10:15.11 2005 93. Dejene Birhanu 10:15.13 2004 94. Nicholas Kimeli 10:15.22 2019 95. Markos Geneti 10:15.24 2005 96. David Chelule 10:15.30 1999 97. Mark Bett 10:15.31 2001 98. Hicham Bellani 10:15.36 2006 99. Sydney Maree 10:15.37 1985 100. Ali Abdosh 10:15.47 2009
*This was the all-time hypothetical toplist as of July 30th, 2020. I’ll have to update it to reflect what’s hypothetically happened since then.
How exactly did you compile this list? Are these their best 4K from the 5K’s they ran? Also why is Cheptegei on the first list as sub-10 but not the second?
How exactly did you compile this list? Are these their best 4K from the 5K’s they ran? Also why is Cheptegei on the first list as sub-10 but not the second?
Sorry, Cheptegei is on the first list because he broke the imaginary 4k WR a couple weeks after the lists were first compiled, but I omitted him from the second list for now until I can fully update the top-100.
At the end of the day, these lists are the work of my imagine and could be dismissed as a load of crap. But please understand just how much time and thought I wasted trying to make them as realistic as possible. Basically I referred to the actual 3k/5k toplists, visiting every athlete’s IAAF profile and attempting to extrapolate/speculate what they’d have run in a world where 1k/4k were the popular distances instead of 800/1500/5k.
So take any athlete from the list, for example Tom Nyakiri at #24, 10:08.64 in 1996. At the time I first posted the list, Nyakiri was 13th all-time at 3k (7:27.75) and 59th all-time at 5k (12:55.94). He also won the 5k bronze at the ‘97 WCs, was twice the Kenyan national champion at 5k, won in Stockholm in his 5k PB race, and was twice top-4 in the World XC long course—so he had good endurance and it’s likely his 12:55.94 isn’t fully reflective of his ability at that distance. Anyway, since he ran 3k at 2:29.25/K, and 5k at 2:35.19/K, it seems reasonable that he would have run 4k at 2:32.16/K if that 1996 3k (in which he was 2nd to Komen’s 7:25.16) had been a 4k instead.
How exactly did you compile this list? Are these their best 4K from the 5K’s they ran? Also why is Cheptegei on the first list as sub-10 but not the second?
Also, I’ll point out that Cheptegei IS on the second list, only at #75 (10:14.22). At the time the list was first posted, he was 73rd all time at 5k (12:57.41) and 82nd all time at 3k (7:33.26). That’s 2:31.08/K for 3k, 2:35.48/K for 5k, and 2:33.55/K for the hypothetical 4k. Then a couple weeks later, he slashed 22.05 seconds from his 5k, or in the case of my psycho babbling, cut 15.66 seconds from his 4k to break Bekele’s WR by 0.23 seconds.
John Wesley Harding07/30/2020 6:40pm EDT5 years ago
Things are pretty much the same over here. The sky is blue. Water is wet. Tom Waits is still super cool. The one big difference I can find from your world is that in this one, 100m/500m/1,000m/4,000m/10,000m are the standard...
I’ll admit upfront, I’ve been questioning whether Rudisha would be so successful at 1k. I really don’t know—he was the best ever at 800, and this scenario forces him to focus on either 500m or 1000m, so it’s possible he’d rise to the top of one of those, or maybe he’d be a regular world finalist but not a world beater. In my hypothetical he holds the WR by a small margin, and won half the golds he won at 800 (‘11/‘12 but not ‘15/‘16).
So, I quickly mapped those world records and the curve type slightly deviates from linear towards exponential (as opposed to logarithmic).
What this means in practice - someone is supposed to run 4K 1.5-2 sec faster than the average of his 3K and 5K times.
Komen's 3-5K average is 10:00.2, so he was likely capable of high 9:58s
Bekele's average is 10:01.1. so he was likely capable of mid 9:59s
Haile's average is 10:02.2, so he was NOT likely to go under 10 mins
Cheptegi ran 2:02 first 800, then 10:03 last 4000. I’d have a hard time seeing him not able to break 10:00 in a one off.
Best and most straightforward argument on this thread (the one about Bekele’s splits was good too). I don’t think it’s 100% clear he could’ve dropped 4s from that split, but it seems slightly more likely than not
I think all 4 of them would’ve been 9:58 high to 10:02 in perfect conditions with wave light… can’t decide on order of likelihood but I think, incredibly, Geb is 4th no matter how you slice it…
Cheptegi ran 2:02 first 800, then 10:03 last 4000. I’d have a hard time seeing him not able to break 10:00 in a one off.
Best and most straightforward argument on this thread (the one about Bekele’s splits was good too). I don’t think it’s 100% clear he could’ve dropped 4s from that split, but it seems slightly more likely than not
I think all 4 of them would’ve been 9:58 high to 10:02 in perfect conditions with wave light… can’t decide on order of likelihood but I think, incredibly, Geb is 4th no matter how you slice it…
Best and most straightforward argument on this thread (the one about Bekele’s splits was good too). I don’t think it’s 100% clear he could’ve dropped 4s from that split, but it seems slightly more likely than not
I think all 4 of them would’ve been 9:58 high to 10:02 in perfect conditions with wave light… can’t decide on order of likelihood but I think, incredibly, Geb is 4th no matter how you slice it…
He goes through 1600 behind pacers in ~4:01 then wins the race in the next 200 meters.
El Guerrouj took only one crack at it, in which he was too slow through 1k, tried to make it up too quickly in the 2nd K, and hung on in the last 800 for that 7:23.09. A couple more attempts and he’s good for 7:21, for sure.
It’s not that 7:20.67 isn’t a strong record, or that there’s necessarily ever been somebody capable of beating it. But if everyone who’s been capable of 7:20.68-7:23.99 had run to that potential, then 7:20.67 would appear much less untouchable.