there is lot of talk here about why boston is slow compared to a berlin/rotterdam even though it is a net downhill ( & ignoring wind )
in theory, the course can be run nearly as fast as the latter 2, providing the pacing down/up hills is judged perfectly ( like cheruyiot obviously did )
here is some working :
the formula for smooth uphill/downhill from basic physics is :
T1 = ( T*d^2 ) / ( d^2 - ( 2*g*h*T ) )
where T1 = corrected time, d = distance, T = actual time, h = elvation or drop
an example is we have say, a 2"06+ run on a flat course, which is ~ 15'00 for 5k thruout
now, lets imagine a hilly course with for example a hilly 10k section of that for analysis :
lets say, that the initial 5k is smooth uphill at 1% gradient or 50m & the next is smooth downhill at also 1% or 50m, for an overall 0 change over 10k
- on a flat course, the 10k is covered in 30'00
- on a hilly course, the energy expenditure to match above :
a) 5k with of 50m uphill ->
900 = ( T1 * 5000^2 ) / ( 5000^2 + ( 2 * 9.81 * 50 * T1 ))
-> T1 = 15'32.95
so, YOU MUST run the uphill in 15'32 in order to expend same energy as running 15'00 on the flat - if you go too fast up the hill, say, 15'20, you will be screwed for the downhill
b) ->14'29.30
the downhill must be run in 14'29 in order to expend same energy as running 15'00 on the flat
what happens in boston appears to be that the initial downhills are run too fast - in this example, if you run it in 14'20, you have run it nearly 10s too fast & you will be screwed for the following uphill in order to maintain energy expenditure of 15'00/5k overall
the key is to run the hills much more conservatively than previously imagined & YOU WILL BE REWARDED
if you add the 2 times together, you get
30'02.25
which means, if you run the hills ideally, you will have capability of running an overall time very close to the pancake courses