Objective and Methodology

Objective:
Predict the cut-off time for the 2016 Boston Marathon.

Discussion:
The qualification and entry procedure for the 2015 Boston Marathon resulted in a cut-off of 1 minute and 2 seconds. That is to say, those who applied for entry were successful if they had a qualifying time of 1 minute and 2 seconds (or more) faster than the minimum BQ standard for their age and gender. For example, a 46 year old female would be required to have a finish time of 3:53:58 or better.

Since it has been introduced, the time cut-off has varied by year by year. It is influenced by several factors including: the number of general entry slots available in any given year; and, the individual motivation for runners to participate in the Boston Marathon.

This experiment is an attempt to predict the cut-off time for the 2016 Boston Marathon by comparing trends in marathon times from year 2016 qualifying races with those trends in marathon times from the previous year’s races.

If fewer people meet the minimum BQ cut-off time at the 2016 qualifying events than did at the same 2015 events, it stands to reason that there would be room for more entrants, and thus the BQ cut-off would be less than 1 minute and 2 seconds. The opposite holds true of course.

The qualifying period for the 2016 Boston Marathon event commenced September 13, 2014.

Boston Athletics Association (BAA) has posted a list of 25 marathons that are most frequently used to qualify for the Boston Marathon [http://www.baa.org/races/boston-marathon/participant-information/qualifying/top-qualifying-races.aspx]. These 25 events cumulatively accounted for 265,000 marathon finishers for the 2015 qualifying period.

#	Feeder Marathon	2015 BQ Date	2016 BQ Date	Finishers
1	Erie Marathon	15-Sep-13	14-Sep-14	978
2	Berlin Marathon	29-Sep-13	28-Sep-14	36,122
3	St. George Marathon	5-Oct-13	4-Oct-14	5,818
4	Portland Marathon	6-Oct-13	5-Oct-14	6,945
5	Twin Cities Marathon	6-Oct-13	5-Oct-14	8,855
6	Chicago Marathon	13-Oct-13	12-Oct-14	38,854
7	Steamtown Marathon	13-Oct-13	12-Oct-14	2,165
8	Baystate Marathon	20-Oct-13	19-Oct-14	1,304
9	Columbus Marathon	20-Oct-13	19-Oct-14	5,523
10	Marine Corp Marathon	27-Oct-13	26-Oct-14	23,385
11	Indianapolis Monumental Marathon	2-Nov-13	1-Nov-14	3,521
12	New York City Marathon	3-Nov-13	2-Nov-14	50,134
13	Richmond Marathon	16-Nov-13	15-Nov-14	4,842
14	Philadelphia Marathon	17-Nov-13	23-Nov-14	10,914
15	California International Marathon	8-Dec-13	7-Dec-14	6,238
16	Houston Marathon	19-Jan-14	18-Jan-15	6,945
17	Boston Marathon	21-Apr-14	20-Apr-15	31,210
18	Eugene Marathon	27-Jul-14	10-May-14	1,361
19	Bayshore Marathon	24-May-14	23-May-15	2,065
20	Mountains 2 Beach Marathon	25-May-14	24-May-15	1,622
21	Ottawa Marathon	25-May-14	24-May-15	5,295
22	Vermont City Marathon	25-May-14	24-May-15	2,434
23	Grandma's Marathon	21-Jun-14	20-Jun-15	6,212
24	Santa Rosa Marathon	24-Aug-14	23-Aug-15	1,235
25	Lehigh Valley Marathon	7-Sep-14	6-Sep-15	1,065
Total				265,042

Methodology:

1. Obtain the race results for all finishers from the 25 feeder marathons identified by BAA, for the 2015 qualifying year.
2. Obtain the race results for all finishers from the same 25 feeder marathons identified by BAA, for the 2016 qualifying year.
3. As each 2016 qualifying year feeder marathon is run, perform an analysis to determine how many people from each gender and age group met the 2015 BQ cut-off for the 2015 event.
4. Perform an iterative analysis to determine what an equivalent cut-off time would be to limit the qualifiers from the 2016 qualifying events to the same number as those from the 2015 events.

Assumptions:

1. Motivation to apply for the Boston Marathon remains the same for the 2015 and 2016 races.
2. The proportion of applicants from the feeder races remains the same for the 2015 and 2016 races.
3. The feeder races, combined, are an accurate representation of finisher times for most marathons.
4. The allocation of General Entry slots for the 2015 and 2016 Boston Marathon remains the same.

Limitations:

1. The list of 25 feeder marathons as posted by BAA is not an all-inclusive list; however, this list does account for over 250,000 data points. The 25 events analyzed are assumed to be a good surrogate indicator for marathon times and trends for the given qualifying period.
2. The age of participants is not always available in the published results, and it is not always published in a consistent format:
• Some feeder marathons allow runners under the age of 18 whereas the BAA does not. The finish times for runners under the age of 18 are lumped with those in the 18-24 age category. There are very few cases of this happening; however, this approach is applied for both the 2015 BQ analysis and 2016 BQ analysis so it should even out.
• The BQ Standards are based on a runner’s age on the date of the Boston Marathon. It is feasible that a runner could be 44 in September 2014, and 45 in April 2016. This causes the runner to be in a different age group for BQ purposes. That runner may not have met the minimum time standard for a 40-44 year old runner, but may have met the time standard for a 45-49 year old runner. The analysis would flag them as not meeting the BQ Standard; however, this approach is applied for both the 2015 BQ analysis and 2015 BQ analysis so it should even out.
3. Deferrals – In 2010, some participants were given the option to defer entry until 2011 because of a volcanic eruption restricting air travel from Europe to USA. In 2012, participants were given the option to defer their entry until the 2013 race because of unexpected heat. In 2013, participants who were unable to finish the race because of the bombings were invited back to 2014. If such a situation occurs at the 2015 event, it would likely limit the availability of general entry slots for 2016.