It's time to decide a goal race time / pace for the Portland Marathon. If you have read this blog before you will know that my goals usually go beyond finishing. That said, a marathon is a long race so somewhere on the goal spectrum is finishing, hopefully without walking... or using a bathroom (my UC can make this a challenge).

To determine a race time I could get real simple and target the Boston Marathon qualifying time for my age group which is 3 hours and 5 minutes. This works out to 7 minute and 3 second per mile pace. I would really like to achieve this time, but I don't want to limit myself to this pace. Another easy time to target would be 2:59:59 (6:52 / mile). Breaking three hours would be awesome, but once again, I think I can do better.

So how do I calculate the best pace? Well, I think past performance and training are the best indicators of future results. This year I have ran a half marathon in 1:21 and a 10 mile race in just under 60 minutes. There are a plenty of theories and calculations that can be used to estimate pace based on recent results. Let's review some of the methodologies. I will use my last two races as a basis for these calculations.

A First the old school method: double the half marathon time and add 10 minutes. Essentially this adds 23 seconds to your half marathon pace. This method is crude. It is unlikely to be accurate for both a 4:45 / mile half marathoner and a 12:00 / mile half marathoner. 23 seconds is a lot of time to the first person but practically nothing to the second. Using my half marathon PR time from January gives me a marathon time of 2:53:10 (6:36) . Despite my skepticism about the methodology, I kinda like this number.

B-F The next group of predictions are based on my 10 mile race time from April, which establishes my current fitness level. The various calculators use this value to extrapolate to other race distances through formulas and normalization methods. Jack Daniels' uses VDOT as the fitness normalization method, while Age Grading measures your performance relative to the best at your age. These methods give four times surrounding 2:47:00 and a slight outlier time (relative to the rest of this group) of 2:49:22. These feel like very optimistic times and most of the calculations have disclaimers that they assume you are in optimal fitness for the new race distance. I would love to run this quickly but I am worried about starting too fast. I have had experience walking the last 10 miles of a marathon, and I do not want to repeat that nightmare.

G-H The next two calculations rely on two inputs to calculate the marathon pace: race time and training intensity. It needs a 10 km time from four weeks before the marathon and the average mileage in the buildup to the marathon. I estimated both values since I am still eight weeks out from my marathon. I calculated the 10 km time using my 10 mile PR with the McMillan conversion. I estimate my weekly mileage at 60 taking into account mileage covered and anticipated miles. This mileage gives a multiplier that ranges from 4.75 to 4.85 based on the hillrunner table. Combining these inputs gives a marathon estimate ranging from 2:49 on the low side to 2:53 on the high. If I get a chance to run a 10 km four weeks before the marathon I will redo this calculation.

I-J The last calculations use my estimated FTP based on historic performance extrapolated to marathon race day. This calculation takes into account my entire race history and therefore has the most data inputs. Unfortunately none of the prior races are marathons. It is unclear whether the marathon will be within the race baseline. Using this data I calculated two values. The slower time, 2:45, is based on a second order fit to the data. On a positive note this time is near the range of values calculated from the B-F group. Unfortunately it still does not fill me with confidence in its accuracy.The faster time, 2:39, is based on a linear extrapolation, and feels like nothing more than wishful thinking.

I have listed all the estimated race times below. The median of the dataset is 2:47:45 (6:24), and the mean is 2:47:51 (6:24).

Marathon Time (Mile Pace) Calculation Method
A 2:53:10 (6:36) old school calculation using half marathon PR
B 2:46:42 (6:21) Runworks (VDOT) using 10 mile PR
C 2:47:43 (6:24) McMillan Running using 10 mile PR
D 2:47:47 (6:24) Running For Fitness Age Grading using 10 mile PR
E 2:46:01 (6:20) Running For Fitness Riegel Formula using 10 mile PR
F 2:49:22 (6:28) Running For Fitness Cameron Formula using 10 mile PR
G 2:49:49 (6:29) faster hillrunner Jim2 using 10 mile PR
H 2:53:23 (6:37) slower hillrunner Jim2 using 10 mile PR
I 2:39:03 (6:04) Runworks (VDOT) using minimum FTP estimate
J 2:45:29 (6:19) Runworks (VDOT) using minimum FTP estimate

So there we have it. The list starts with a calculation based on a half marathon time and a crude computation, progresses to single input calculations with more complex conversions, on to two input calculations, and finally to extrapolation of my running record.

Which value will serve as the basis for my race pace? I think it is good to use as much data as possible which eliminates table items A-F. My FTP provides the most complete dataset, but it is unproven for converting to the marathon distance. Therefore, I am leaning to the next most rigorous methodology, using my race performance and training intensity to estimate a race time. This method was formulated for the sole purpose of marathon time prediction so I find it that much more trustworthy. It is more conservative than the other predictions, but I have found the other conversions too optimistic (fast) when converting up to longer races (in particular the half marathon from 10 km times).

Baring a change of heart or a warm up race that deviates from my historical performances I will be targeting my first mile in the 6:29 - 6:37 time zone. I will try to split the difference for the first 20 miles. I will plan to give'er hell for the final 10 km, and will celebrate no matter how it turns out!