Jim2's Running Page

Marathon vs HM Comparison & Analysis

Home
About Me
Favorite Links
Contact Me

2/16/09

A runner in Israel, Hanan Joseph, sent to me a copy of two articles that he published on an Israeli website analyzing the relationship between the results of runners who completed both a half marathon and a marathon that are run four weeks apart (December and January) in Israel. The first article (Part I) is based on the results of the 2007-2008 races. The second article (Part II) reflects data from the 2008-2009 races. A translation of the articles, which were originally written and published in Hebrew, are posted here with Hanan’s permission.

Hanan’s articles complement and expand on two previous essays posted on my Running Page (Predicting A Marathon Time and Half Marathon Pace vs. Marathon Pace) that address actual marathon vs. shorter distance race performances compared to projections by popular running calculators. His interesting analyses of (1) the relationship between marathon and half marathon results of approximately 600 finishers grouped into seven marathon finishing time categories and (2) the quality of marathon pacing (within 4% of even splits) in each of the seven categories add additional dimensions to the subject.

Predicting Marathon result based on Half Marathon result. – Part I

Written By: Hanan Joseph. Edited by: Moshe Bretter

Introduction:

In Israel, there are two main running events. Beit-Shean Half marathon and 4 weeks later Tiberias marathon. Both races have similar courses and weather conditions. After the HM, most runners, before even taking a shower, will check their estimated marathon result based on McMilan's calculator.

McMillan's calculator evaluates the ratio between the HM result and the marathon result as 2.11, e.g. 1:45:00 HM result is predicted to be 3:41 at the Marathon.

I tried to validate this ratio, based on the results of 348 runners who participated both the Beit-Shean HM and Tiberias marathon in 2008.

Test Method:

 

  1. Calculating Net time
  2. Calculating the ratio between HM and Marathon results
  3. Calculating half way point at the marathon (21.1 km)
  4. Cleaning the data from unreasonable results, like ratio less than 2 (No connection between HM and Marathon) and ratio bigger the 2.5, runner had serious difficulty completing the marathon.
  5. Dividing Marathon results into time categories.

 

Ratio Results by Categories:

 

No. of Results

Marathon / HM Ratio

Categories by Marathon result

31

2.12

Sub 3:00

29

2.14

3:00 to 3:15

68

2.17

3:15 to 3:30

53

2.2

3:30 to 3:45

44

2.2

3:45 to 4:00

33

2.22

4:00 to 4:30

19

2.29

Over 4:30

 


Ratio Results by Categories - Graph:

Marathon_vs_HM_1.jpg

We can clearly see that the faster the runner is the ratio is smaller.

The next step in my experiment was to check how well the runner planned his run. I defined a metric which I called "Good Run".

"Good run" is a reasonable ratio between HM and marathon, and running even splits. 

  • Reasonable marathon ratio – The runner's ratio is around the category average, meaning not too conservative or too pretentious, e.g. for 3:15 to 3:30 runners the reasonable ratio is between 2.1 to 2.22 (AVG is 2.17). The lower end of the range is the McMillan prediction and the upper end is the average ratio plus 2%.
  • Even Split – The ratio between halfway mark and final result is between 1.96 and 2.04. Ratio equals 2.00 – perfect splits, bigger than 2.04- runner hit the "wall", less than 1.96 starting too slow.
  • Exceptions – I consider "good run" for those runners who had ratio less then lower limit but had "Even split".

"Good runs" analysis:

 

"Good Run" %

Even Split %

In Range %

Ratio Range

Marathon / HM Ratio

Categories by Marathon result

90%

90%

68%

2.09-2.15

2.12

Sub 3:00

69%

79%

62%

2.1-2.18

2.14

3:00 to 3:15

62%

69%

68%

2.1-2.22

2.17

3:15 to 3:30

53%

53%

77%

2.1-2.25

2.2

3:30 to 3:45

36%

41%

66%

2.1-2.25

2.2

3:45 to 4:00

24%

33%

67%

2.1–2.27

2.22

4:00 to 4:30

11%

11%

58%

2.1-2.35

2.29

Over 4:30

 

Graph:

Marathon_vs_HM_2.jpg

Results Analysis:

 

  1. I find McMillan basic assumption, that the ratio between HM and marathon results is a constant, has no validity. There is a changing ratio, depending on the runner's ability. In my opinion slower runners have insufficient aerobic fitness to handle the progress from HM to marathon.
  2. We can clearly see that slow runners have less "even splits". This means slow runners start too fast for their ability.
  3. There is a significant difference between runners with marathon result up to 3:30 and those over 3:30. It seems that running a marathon requires the ability to practice runs at a faster pace.
  4. Running the desired negative split is very difficult, and seems a little bit unrealistic for slower runners (over 3:30)
  5. We can assume that slower runners tend to be less experienced, therefore they make more pacing mistakes.

 

Recommendations:

 

  1. Based on the results it seems a runner should evaluate his marathon pace based on his desired category and not by McMillan's calculator evaluation.
  2. The less experience the runner is, he should be more conservative regarding his marathon pace.

 

-------------------------------------------------------------------------

Predicting Marathon result based on Half Marathon result. – Part II

Introduction:

 

Part I of the article checked the ratio between 2008 half marathon and Marathon results. I claimed the assumption that the ratio between HM and Marathon is a constant (according to different calculators) isn't valid.

I made an additional check based on 2009 Tiberias marathon and Beit-Shean HM (occurred 4 weeks earlier).

Test Method:

 

Based on Marathon result, marathon splits and HM result I calculated the following ratios:

  1. Actual Ratio – Marathon result divided by HM result
  2. "Half" Ratio – what suppose to be the runner's marathon result based on his half point split, divided by HM result
  3. 30 KM (18.75 mile) Ratio - what suppose to be the runner's marathon result based on his 30 km point split, divided by HM result
  4. Clearing data of ratio bigger than 2.5

 

Ratio Result by Categories:

 

Time loss per 7.45 mile [minutes]

Pace loss [sec/mile]

30 KM Ratio

"Half" Ratio

Actual Ratio

No. of Results

Categories by Marathon result

1:00

8

2.10

2.10

2.14

28

Sub 3:00

0:48

6.5

2.14

2.12

2.17

37

3:00 to 3:15

1:12

10

2.14

2.12

2.18

56

3:15 to 3:30

1:48

14.5

2.16

2.16

2.22

54

3:30 to 3:45

3:48

31

2.13

2.13

2.25

57

3:45 to 4:00

3:36

30

2.20

2.17

2.31

42

4:00 to 4:15

6:00

48

2.16

2.15

2.33

30

4:15 to 4:45

6:48

55

2.20

2.14

2.38

12

Over 4:45

 

Pace loss – evaluation of how many seconds per mile the runner has slowed down in the last 7.45 mile.

 

Time loss – how many minutes the runner lost due to slowing down

Marathon_vs_HM_3.jpg

Results Analysis:

 

  1. Once again, we can clearly see that the ratio between marathon result and HM result isn't a constant, but a variable according to the runner's level.
  2. Until the half distance point, most runners were able to stay near the McMillan's ratio (2.10 to 2.17)
  3. Some of the runners started slowing down from half distance point. The difference between yellow line and the pink line. Between 13.1 mile mark and 18 mile mark, runners slowed down a bit.
  4. Runners started to hit "the wall" from the 18 mile mark.
    1. 3:30 hours runners, lost about 8 sec/mile
    2. 3:45 to 4:15 runners lost about 30 sec/mile
    3. Over 4:15 runners lost about 50 sec/mile
  5. We can see that most runners can run at McMillan's predicted pace until 18 mile mark.
  6. The slower the runner is, the ratio gets bigger, meaning unwanted positive split.

 

Recommendations:

 

  1. The recommendations from part I are still valid, chose your category ratio and chose conservative pace.
  2. It's better to start slowly and keep the pace until 18 mile mark
Most runners eventually will hit "the wall", therefore be prepare for that during your marathon run and during training.