We would expect the fastest runners on the rubberized track surface to outdo the fastest on pavement. But perhaps there's more.
I plotted the top 20 person-times for the years 2004 to 2014 (this is a opposed to the top 20 times overall, which can include duplicate performances. Hopefully this better represents overall trends rather than a single person having a very good racing year). The data was pulled from IAAF's database here.There are no top 20 lists before 2004, though possibly one can obtain these from other sources. Nevertheless, we have ten years of good data.
The first thing you might notice, unsurprisingly, is the top mean running times are always better for track than for road. The mean difference from 2004 to 2014 is 32 seconds.
Supplied with new information, new patterns emerge for the track. For instance, never in the 2000s have the top 20 athletes collectively "peaked" during an Olympic year. Those waiting with baited breath for record-breaking performances will be quadrennially disappointed. Rather, the year to expect a breakthrough 10 kilometer track time is during odd-numbered years, in particular those just prior to an Olympic year such as 2007 and 2011. But the trend is imperfect, for in 2004 (Athens), better runs came after the Olympics (2005) than before (2003). Possibly this is due to missing best times in those 'earlier' years.
What I find surprising is that the oscillations for track times appear to be growing larger and larger. The slowest year since 2000 was 2014. Is this an accident of fate, deliberate, or better anti-drug policies, or something else? It could be a combination of several overlapping trends. Clearly athletes collectively put less attention on even-numbered year, regardless of why. What is in store for 2015? We should probably expect some faster times assuming the 2007 and 2011 years are of any indication.
The big surprise for me is seeing no discernible pattern for road races. There are good years and better years, yet scanning for some obvious reason behind them is tricky. Road race times are not exactly correlated -nor entirely anti-correlated- with track times. Only one true valley point exists at 2011, otherwise times move up and down seemingly at random. Perhaps the reason for no time trends is that road races are often for places, not times (unless you set a world record). Doing well in a road race means a podium finish and/or money, whereas track times mean qualifying for other things, not least of which is the Olympics or Diamond League races (whereafter comes money). Road races can also be subject to weather, hence a good course in one year might be a "bad" course the next.
I fibbed a little when saying there was no road data for 2002 or 2003 for road. True there are no top 20 times, but I have available top ten lists. Here is a similar plot as pervious, only now for the top ten best person-times.
The patterns is rather the same: valleys in 2005, 2007, and 2011, then peaks in 2006, 2010 and 2014. Again, Olympic years are never peak years for top track runners. The mean difference between road and track is again about the same at 35 seconds. [Aside: I noticed Athletics Canada set up a criteria for an upcoming race, the Chiba Ekiden, whereby the "top finishing declared athlete ... who has a sub 29:40 road or 29:20 track performance". With only 20 seconds difference, qualifying on the road may prove disadvantageous compared with the track. Ergo if after a specific time, best stick to the track].
Looking at the top-ten road athletes, again it seems there is no obvious pattern. I see hints of anti-correlation between them and the track (i.e. that a good year for one means a bad year for the other), but not a strong one. What's really striking is the slowing pace of top-10 track times from 2011 until 2014 so that times have been slowing for four years straight. Perhaps track athletes are migrating to road races, where more money can be won. Also very tempting to consider the anti-doping hypothesis, but without further proof it is hard to say. Certainly a more open conversation about drugs in endurance athletics is being tabled, or at least admitting such things are possible. Consider doping by Kenyans Muraga and Jeptoo, and Qatari Hamza Driouch. Would-be doping athletes are more aware they could be caught.
Looking ahead again, we likely know what to expect from 2015: faster times, but with a caveat. Road races may continue to lure even more runners away from tracks, or stronger doping controls are affecting times (not a certain link), or bad weather could cause an unexpected slowdown for a top race. No signs of speedy 2015 times yet yet, as the fastest 10k road and track times are still above the 28 minute mark. As always, we can only wait, watch, and speculate. Cheers,