This is part of a trilogy of articles providing an in-depth statistical analysis of Comrades 2022. The articles on Elite Men and Elite Women are also available.
Getting to the Start
The graph below is a hurricane funnel answering the often-asked question, “What happens to all the entries?”
16479 enter, 1688 don’t bother to submit a qualifier, 661 don’t make it to registration, 916 collect their race pack but don’t start and 1503 fall on the road to Durban leaving us with 11711 Comrades 2022 finishers.
READ: A List of Valid Comrades Excuses
The next graph shows the same story by province. Limpopo entrants are the most dedicated (and they want their race packs – almost everyone who qualified got their T-shirt) followed by Eastern Cape runners.
Of interest is that 10% of the original entries don’t submit a qualifier and there are more people that collect their race pack and don’t start than those that qualify but don’t make it to registration.
Finisher Stats
Almost 9/10 starters earned their medal.
Within South Africa, Western Cape runners are the toughest with 92% of their starters finishing, whereas Limpopo runners are the softest with an 86.6% finisher rate.
However, European runners put everyone to shame with a 95.4% finisher rate and South Americans also make the most of their carbon footprint (95.2%). Someone definitely needs to make North American ultra running great again – they are bottom of the pile with 84.7% of their starters crossing the finish line in under 12 hours.
Gauteng and KwaZulu Natal runners make up almost two-thirds of the finishers with Gauteng alone comprising close to half of the field.
I’m really glad that sanity prevailed and Comrades is back to a June start date in 2023 – personally I couldn’t face training through a Highveld winter and I expect I am not alone. Pissing off half your client base is bad for business.
Finish Rate
The graph below shows how many runners crossed the finish line every 15 minutes.
Humans are goal (or medal) driven and it’s the normal Christmas tree, pagoda pattern for each medal cut-off with a ‘love handle’ on the last 15 minutes for each medal. The only exception is between 10h15 and 10h30. Interestingly, both male and female runners had the 10h15-10h30 anomaly – conjecture is that this is the timing of the safe bronze bus who came in just under 10h30.
The doughnut graphs are the same data set in a format that provides a different visual of the percentages by 15 minutes split (overall, male and female).
The final view of the finisher numbers by 15-minute splits provides a timeline comparison.
Of interest:
35% finish over the last hour and just under half the field (46%) over the last 75 mins.
Just 3.9% finish under 7h30.
If you got Bill Rowan (sub-9) or better you’re in the top 18.5%.
Historic Participation Trends
The graph below is the finisher percentage for all down runs this millennium showing total starters and max temperature at the finish in Durban. There’s a slight increase of 2.1% from 2018 to 2022.
The next graph shows the finisher percentage for all Comrades run since 2000. The big anomaly is 2013 when almost half the field did not finish due to the brutal weather (extreme heat and head winds). 2012 also had a low finish rate – there was wind but not much heat that year. My hypothesis is that wind has a bigger impact than wind on finisher rates – and when you get both you’re in for a really tough day. The peak finisher percentage period was 2003 to 2007 when the cut-off was extended by one hour to 12 hours (the 2000 event had a 12 hour cut-off as a ‘once-off’ dispensation). Looks like we’ve regressed back to the mean after getting used to the additional hour.
The third graph in this section shows the growth (and post-Covid shrinkage) of the Comrades Marathon. There’s steady growth through the 1960s and 1970s followed by an explosion of entries in the mid-1980s correlating with the Fordyce era and television coverage. This results in Comrades becoming by far the largest ultra marathon in the world. There are less than ten ultras worldwide with over 1,000 finishers (most of them are in South Africa) and only one with over 10,000 finishers (although Two Oceans did have just over 10,000 finishers once in 2019).
Next graph is the same data on a shorter timeline (from 1975 when the race was opened to black and female runners for the first time). The label numbers are finisher numbers.
The final graphs in this section show participation details since 2000 over all runs and down runs only. The label percentages are entries that start, starters that finish and percentage of finishers by gender (I do not have the number split of female / male starters for most years unfortunately). The percentage of female runners (20.3%) is still very low (by comparison Two Oceans is around 30%).
Start Line Stats
How long did it take to cross the start line at Comrades 2022? An average of 9 seconds in batch A and 5m36s if you were at the back of the field in H batch.
Below is a different view showing how many runners from each seeding batch cross the start line every 30 seconds.
Next is a graphic representation of how long it took Comrades runners in each batch to cross the start line courtesy of data scientist and Green #54898 Andrew Collier (highly recommend giving him a follow on Twitter @datawookie).
The graph also provides a comparison of between the 2022 and the previous down run in 2018. It’s quite incredible the impact of 6,000 additional runners – most of E-H batches were well on their way to Durban in 2022 whereas 2018 runners still had several minutes before starting their journey. H batch runners in 2022 probably saved about 5 minutes of gun time.
Medal Breakdown
Here’s the medal table by seeding batch.
46% of As finish under 7h30 but overambition leads to a higher attrition rate of 7.5%. There must be some interesting stories from the 1.5% of As who ended up with a Vic Clapham medal.
A silver is definitely possible from B (6.4%) but here the Bill Rowan medal (65%) dominates.
Cs favour the Robert Mtshali medal and Ds the Bronze, with Fs dividing the spoils between Bronze and Vic Clapham. It would be very interesting to know the story of the 6 Cs and 2 Ds that earned silver.
G and H are earning Clapham’s copper medal. Hs live life on the margins with a third falling on the road to Durban (93% of Hs earn Vic Clapham or DNF).
The Green Number Club had 772 representatives at Comrades 2022. Half of them made the most of the experience coming home in the final hour and most of the other half either earned a bronze or took an early shower. Less than 4% of Greens finish under 10 hours (but it should be noted that there would have been Greens in batches A-D if they submitted a faster qualifier).
As expected, average finish time correlates closely with seeding batch at Comrades. A seeds are an hour faster than Bs who are 51mins faster than Cs who are 44mins faster than Ds who are 39mins faster than Fs who are 22mins faster than Gs who are 13mins faster than Hs.
The table below is the source data for the above graphs which also includes the fastest male and female time per seeding batch at Comrades 2022.
I’ve also done graphs breaking down medals by seeding batch for the two traditional genders.
Only 3 A ladies DNF – sadly two of these were former champions Ann Ashworth and Charne Bosman and only one finished slower than 9 hours.
B-grade women dramatically outperform B-grade men (an incredible 83% of B ladies finish Bill Rowan or better). However the depth in the women’s field should be concerning – there are just 111 A and B female starters compared to almost 2,000 men.
The ladies continue to outperform the men in the medal standing percentages through C, D, F and G seedings.
Somewhat surprisingly (since women are normally more sensible ultra runners than men), the B, C, D, and F females have a higher DNF rate than the men. The natural order of things is however restored in batches H, CC and E.
I’ve included the source data tables for those that want the details. Positive / negative split stats are available later in the article. Of note is the HollywoodBets A batch runner who took over 5 hours longer on the second half of the race (he went through halfway under 3 hours and scraped through for a bronze). The best “genuine” positive split I’ve found is a D seeded KZN Striders runner who ran was 24 minutes faster over the second half.
Age
The graph below shows the finishers by age – almost half the field is in the 40-49 age group and 76% are under 50.
There are some differences between the sexes with a much higher percentage of ladies under 40.
I’ve also added a table showing the average age per batch. The only batch that is (just) below 40 is A. As expected, the average age gradually increases per batch with the real outlier being the golden oldie Green Numbers in E who, at an average age of 55.5 years, are almost a decade older than the next oldest batch (group H at 46).
Medals by Age & Gender
The medal breakdown graph highlights just how tough it is to break 7h30 and that the majority of the field finishes towards the end of the day (with the female side of the field generally getting more value for their R1200 entry fee).
The next graph provides a detailed breakdown of medals achieved by each age group for men and women. No surprises here – no one can escape Father Time and we get slower as we get older (although it looks like the ladies age more gracefully than the men).
The summary table that was used for the above graphs is included for reference below.
Fastest & Slowest by Age
The table below lists the fastest runner by age. Due to the size of the screen caps I’ve split into over and under 50 years old.
Of interest is the small number of runners under 30 (275 men and 91 ladies) who started the race. Many runners are of the opinion that you shouldn’t run Comrades before you’re 30 so perhaps youngsters are heeding this advice. However, I think it’s more likely that Comrades is failing to attract the younger demographics which is not good news for the future of Comrades as this might indicate a rapid drop in entry numbers in future years.
In the men’s field the oldest runner to get a silver is 60-year-old Shaun Meiklejohn. That’s his 31st finish and 21st silver to go with 10 golds (and one win). Quite incredible longevity – I think he makes a strong case for the greatest ever Comrades runner.
The oldest Bill Rowan is RAC’s Moses Motshabi (67). I used to give Moses a lift home from the RAC time trial (although he probably could have run home faster than I drove!). 70-year-old Theo Swanepoel claims a comfortable Robert Mtshali to be the oldest in this category whilst octogenarian, Maros Mosehla from Polokwane AC, is by some way the oldest finisher.
On the female side of the field, it’s noticeable that the fastest times of the 37-45 age group are markedly faster than their more youthful counterparts (with one or two exceptions). The oldest lady to earn silver is Phantane AC’s Jenni Kruse (48) whilst Julie Shadwell (60) is turns back the clock for the oldest Bill Rowan.
Two matriarchs of South African running, Blanche Moila (66) and Val Watson (67) round off the Robert Mtshalis whilst Linda Icely gives credence to her club’s name (Old Eds) as the oldest female finishing, coming home with 20 seconds to spare on the bronze cut-off.
I’ve included the slowest by age as well for completeness. No comments or judgement on this chart – anyone who finishes Comrades is a hero – but if you have a friend on this list they might want to know!
Here’s some more on how Comrades performance degrades with age
The graph below compares the average finish time and attrition rate by 5-year age buckets. Average finish time peaks in the 26-30 bracket and deteriorates from there. Anyone below 45 years old beats the “full field” average finish time and bailer rate but after that there’s a rapid decline in finish line fortunes.
A more detailed year on year view of the same data is in the next graph. The best ages by finish time are 28, 27, 33 and 32 (in that order). The number prefixing the average finish time is the finishers for that age. Other than the old and young fringes, the numbers for each age are large enough to be statistically relevant.
Next up is a great visual from data scientist Andrew Collier showing the gender-based finish rate by batch and age category compared to the overall average. When it comes to claiming your Comrades medal, B (below the age of 60), C and D seeded men are the strongest performers whilst A seeded men between 46 and 55 show the youngsters how it’s done.
For women, D seedings are the best bet for a finish with C and E also slightly above average. For females, the biggest positive deviation from the average are the older ladies in B, D and F.
A table with summarised source data is provided below.
“Split Happens” Champions
Here’s the slowest men through Drummond per medal earned. A true negative split is incredibly hard since the course is 2.12km longer from the Drummond mat and no gold or Wally Hayward medallists achieve this accolade. There’s some incredible pacing from Lindt de la Port who earns silver by 8 minutes from B batch. Vic Clapham is where it happens with Vusi Mavuso running almost a minute/km faster over the last 9 kilometres to make it home with 22 seconds to spare.
On the ladies’ side there were also some heroic runs and some big negative splits. RunZone’s Yasmeen Suleman is the only runner in the field to go through Drummond over 6 hours and still finish the race – she eventually coasts home with almost two minutes to spare.
Unless you win the race (which Tete Dijana does), the fastest time through Drummond by medal earned is probably not a list you want to be on. There are some spectacular positive splits here with Herbert Mokgala’s bronze being the standout performance – he takes over 5 hours longer to complete the second half. Full credit to these runners for hanging in there – lesser men would have got into the bailer bus.
The positive splits on the women’s table are a lot less spectacular than the men’s. Namibia’s Anna Amutoko is the only A seeded lady to finish slower than 9 hours. As with the men, the tenacity shown to stay on the road and earn a medal is admirable.
For completeness, I’ve also included the fastest and slowest time through each split by medal earned. There are plenty of repeat offenders on both lists. Gideon Smuts, who ran the last 9 kilometres from Sherwood in 5:39 pace must have some good stories to tell – dead men tell no tales but Comrades runners who finish with 45 seconds to spare certainly do!
On the ladies side, incredibly Janie Grundling (who finishes 8th) is the slowest gold medal winner through every split whilst Mitsie Van Der Westhuizen was the slowest from Lynnfield Park onwards for a perfectly paced Isavel Roche-Kelly medal (slowest of course being relative in both cases).
On the other side of the table, two former gold medal winners, Carla Molinaro and Belinda Waghorn are the most consistent entries. I chatted to both of them about this dubious honour…
Molinaro, whose approach to running ultra marathons can be described as “all or nothing Gung Ho-style” described her race as follows, “I tried to be sensible this year and not race like an idiot but I hadn’t been feeling great for the three days before the race and once I started I knew that something wasn’t quite right and it was a race of survival to get to the finish line. Took it steady and managed to get to 75km in just about one piece and then really began to suffer.” She was unable to hold off the challengers and dropped from 8th to unlucky 13th over the last 10 kilometres. The good news is that she’ll be back next year with a less cautious approach, “My goal next year is to go as fast as I can, full Gung Ho until the finish line!”
Check out Carla’s vibrant social media page: https://www.facebook.com/CarlaRuns
When life gives you lemons, make lemonade. When Comrades destroys your hip on the road to Durban, add a Vic Clapham medal to your collection. Belinda Waghorn is undoubtably the only female Comrades runner who has earned every medal available to women over her eight finishes. However, there’s one technicality – she’s earned a silver in 2013 before it was surrogated by the Isavel Roche-Kelly medal. Naturally Belinda’s plan is to return in 2023 and add the IRK to her collection. Interestingly, Belinda was previously “robbed” of a Vic Clapham on her debut run in 2000 when there was a “once-off” cut-off extension to 12 hours as part of millennium madness.
How accurate are the timing mats at Comrades?
There are an incredibly small number of runners – just 229 out of about 12,000 who miss a timing mat and “claim” a finish. Of those 229, the majority (207) miss one mat, 18 miss two mats, 3 miss three or four mats and one person misses all eight.
As an example of just how low the probability is of missing a mat, only 8 out of almost 12,000 runners missed Drummond which is a minuscule probability of 0.067%. However, the “missed eight mats” runner is a sad rather than malicious story. The number belongs to a deceased runner whose friends ran over the finish line with a banner of him and they attached his race number to the banner.
I did some detailed analysis of the runners that missed mats which you can read in the Comrades Cheats article.
Blinding You with Data Science
I roped in data scientist Andrew Collier again (he’s on Twitter @datawookie) to provide some additional insights into start, split and finish time distributions.
The graph below shows how the field spreads over each split with a comparison to 2018. Up to Drummond, 2022 is slightly faster than 2018 (probably attributable to a faster start with fewer runners), to Pinetown it’s even-stevens and then a bit faster again (likely the effect of the new 10-hour Robert Mtshali medal).
You would generally expect fairly even distribution (except for the finish), the large bumps (e.g. around 3 hours at Cato Ridge is most likely the effect of large busses, both runners choosing to be paced by a bus as well as runners get stuck in the heavy traffic).
Next is a great view of the finish time distribution per batch. Bs who finish slower than 9-hours have underachieved, in D the target is Robert Mtshali whilst the majority of G and H are running for copper Claphams.
The remaining graphs do a direct finish time tracking between the 2018 and 2022 fields. There’s very little difference at the front of the field but from 9-hours there is a marked decline for about 30 minutes. The impact of the new Robert Mtshali medal is clear with the largest deviation at the 10-hour mark. 2022 remains faster all the way to the final 12-hour gun. The large deviation at 11h15 is most likely the relative timing of large busses in both years.
The final graph in this section provides a slightly different perspective of this data. The data shows that there is a big shift in the middle of the field – a significant number of former bronze medallists respond to the incentive of a Mtshali medal by finishing under 10-hours. There also appears to be a very slight negative impact on potential Bill Rowan medallists who are happy to slow the pace and settle for a Mtshali.
Where your 2022 Comrades Fell
About 13000 of the 13200 starters are still in the game at halfway. The biggest attrition is usually over the next few splits before flattening again on final split. However in 2022, almost 40% of the 1500 bailers run 90% of Comrades but don’t earn a medal.
Did the extra 10 minutes at split cut-off result in many more finishers?
The short answer is “No”.
An additional 10 minutes was allowed at the 2022 event at all the cut-offs except Drummond (10 minutes had already been added there in 2018). However, there are only 12 instances (8 of which are at Pinetown) where a runner goes through the split within 10 minutes of the cut-off and finishes the race. As there are some runners who run a very tight split more than once, there were 9 runners who would have had an early shower in 2018 but benefitted from the more relaxed cut-offs in 2022.
Who are these 9 runners? They are acknowledged in the next section.
There were only 144 runners in total who crossed at least one cut-off within 20 minutes and finished the race. The fastest was Jeppe runner Warren Thomas (11:40:51), four others finished under 11:50 and the remaining 139 all claimed last 10-minute finishes.
Interestingly, there is not one person who runs the first split to Lynnfield Park within 20 minutes of cut-off and finishes the race and just one person within the 20 minutes cut-off radius at Cato Ridge that finishes. All other runners in the 20 minute window on the first two splits fail to finish (although just 17 people are physically cut-off in the first 30km.
What does this data on the first two splits suggest? Either that you have no chance of a medal if you struggle to make the first two splits or there are a lot of runners starting out way too fast (I strongly suspect the latter).
The drop in numbers (finishers and non-finishers) at the Winston Park cut-off supports Norrie Williamson’s view that this cut-off is too tight and should be extended.
A last insight is the massive swing in numbers at the finish line – Pinetown has the highest split total (92 eventual finishers) whereas an incredible 1800 people finish in the last 20 minutes.
And a final consolatory footnote for the 176 runners that crossed the line within 10 minutes after cut-off but had nothing but sore legs to show for their efforts.
Were these the 9 Greatest Athletes at Comrades 2022?
I generally find that some of the most interesting stats are from the back of the field. Below are the only 9 humans who made a split cut-off within 10 minutes and still earned a medal.
Their tenacity to refuse to throw in the towel is admirable. Of particular note are Gideon Smuts who had the closest shave (just 1m35s to spare at Sherwood) whilst Yasmeen Suleman was the only athlete to go through Drummond over 6 hours and finish the race.
Negative Splits
Just 83 runners achieve a negative split at Comrades 2022 (that’s 0.63% of the starters). This is made all the more difficult with the distance from the Drummond timing mat to the finish being 2.12km longer than from the start to Drummond.
Most of the negative splits come from D and F batches and ladies outperform the men percentage wise. There’s just 1 silver medallist (Lindt de la Port) with a negative split (and no golds or Wallys).
The tables below show all 83 negative split runners: Ignore the loudmouthed pretenders, these are the people you want to get advice from for Comrades 2023!
[Note: I have done some basic verification and removed – and reported – obvious cheating. All 83 of these negative split runners had a recorded time through every timing mat.]
Negative Pace
With the placement of the Drummond timing mat making the second ‘half’ 2.12km longer, negative pace is a good proxy for negative splits but there are still just 699 runners (5% of starters) who achieve this feat.
Unless you win the race, your best chance of running negative pace is as a female silver medallist (and the ladies are about 3 times more likely to run negative pace than the men).
Youthful legs beat wise old heads – not many people over 50 manage a negative split. One in ten of the H batch runners achieve a negative pace but none of the 11 Wally Haywards comes close.
The table below shows the Top 20 and all silver negative paced men. As with 2019 only the first two men run a negative pace. In 2019 you had to go down to 83rd place to get the next negative paced male but in 2022 it’s 122nd placed Fourways clubmate Dawie Meyer who is next on the list. There are very few sub-7h30 runners that run a negative pace and home town advantage may play a role – the list looks like it’s heavily skewed towards local KwaZulu Natal runners!
There were also just two gold ladies that managed a negative pace, Morozova who won the race and local KZN runner Janie Grundling. Unlike the men, there are several top 100 ladies running a negative pace but the list of women who run under 9 hours at a negative pace is still a short one.
Do Yellow Number Runners Have A Higher Finish Rate Than Other Runners?
Andre Maakal sent me the following question, “Does having a serious goal/reason to finish Comrades like a Green Number milestone increase your chance of success?”
I would have predicted that the hypotheses would prove true. We’ve seen how goal orientated runners are with the highest number of runners for every medal finishing in the last 15 minutes before the medal cut-off. However, the data says milestone runners (those running for Green, Double Green, Triple Green and Quadruple Green) perform no better than the average runner with an 88.8% versus 88.7% success rate respectively. The best performing group are runners between 1 and 8 finishes with a 90.5% success rate.
The graph below provides the same information without the categorisations. The lowest failure rate is those with 2 medals (8.1%) and generally the failure rate increases from 10 finishes onwards. Interesting that those running in Green for the first time have a much higher failure rate (15.7%) than the others around them. There are some tough old buggers at the top of the table – everyone with 37 or more finishes added another medal to their collection.
The tables below provide the detailed source information and include the split between men and women. Finishing rates (89.1% for men versus 87.2% for women) is probably the only area where men outperform the ladies at Comrades.
The final table in this section shows a split based on the starting batch – there’s lots of detail but I’ve run out of energy to make any insights here!
Is Comrades Really an Old Fogies Race?
The data says no. Despite what they tell you, it’s the youngsters that have the by far the best finishing rate. Your highest probability of earning a medal is in the 20-25 age bracket with the young ladies cracking a 95% success rate (albeit on a small sample size).
Life might begin at 40 but Comrades runners decline rapidly after 45 – although the septuagenarians show the 60 year-olds what longevity is all about.
Shoe Brands
Asics is by far the most popular shoe followed by Nike. Add New Balance and Adidas to the party and you’ve got 80% of the field covered.
Note: This is based on shoe selected when entering. Brands with fewer than 50 starters are grouped under “Smaller Brands” category.
Saucony wearers achieved the highest success rate with 91.6% of their starters earning a medal. Other brands where more than 9 out of 10 starters finished were Nike, Mizuno and Adidas. Some cursory analysis shows that Asics is favoured by the older runners (half of E Batch run in Asics) which might account for the lowest finish rate of the big brands (as previously highlighted age is the biggest factor in finisher rate).
The table below provides detailed information for all brands whilst the next two provide a breakdown of brands with more than 50 starters by seeding batch.
The next graph shows the percentage of market share each shoe had by medal. Nike absolutely dominates up front in the sub-7h30 medals. This is partially explained by their partnership with the Nedbank running club who have many top athletes in their ranks (7/11 Golds, 5/7 Wallys and 23/191 Nike wearing Silvers run for Nedbank). Maxed (which is a much cheaper option than rival brands) also outpaces the competition based on the number of runners who wear their shoes.
Many (but definitely far from all) of the top athletes are sponsored which will impact the shoes in the first few columns. Based on the total number of runners wearing their shoes, Asics and New Balance are grossly underrepresented in the faster medals with increasing growth towards the back of the field.
A second related graph provides a breakdown of medals for each brand (i.e. the percentage is based on the number of runners wearing the shoe rather than the total number of runners in the field). Maxed and Puma are favoured by (or sponsor) a higher proportion of faster runners than other brands.
Below are two tables providing full details of the values for both graphs. It’s worth noting the number of runners per brand as a smaller sample size will naturally result in more variability. In the graph, brands with fewer than 50 runners were grouped together under the “Smaller Brands” heading.
The next table shows the breakdown of shoes used by male and female runners. Brooks and New Balance are noticeably more popular amongst the ladies.
Finally, one last table with some more shoe data, this time cross referencing shoe worn, age category and finish rate. The table shows which shoes perform better (and worse) than average for finisher rate. I checked with Andrew whether there was anything of statistical significance here – there’s not. Finish rate is more dependent on the legs or the runner than the shoes on the feet.
However, it is interesting just how well Nike runners perform. If you’ve spent all that money increasing your carbon footprint you’d better make sure you cross the finish line!
Also of interest is how Asics are favoured by the older generation – 33.9% overall market share but this increases to over 40% for 50 years and older. This goes some way to explaining the lower success rate of Asics runners as age is the biggest factor in predicting Comrades success.
Repeat Runners
Almost 6000 and 7500 Comrades 2022 runners were also part of the 2018 and 2019 party respectively. The faster / slower seesaw was almost even for 2019 but those who ran in 2018 were generally much slower in 2022 – can’t blame Covid for this one!
In addition, 4622 people ran all three years and 357 people who finished both 2018/9 failed to add to their medal count in 2022.
A surprisingly large number of successful 2018/9 finishers bailed in 2022 (528 and 602) – that may well be attributable to the extra Covid-kilograms. Perhaps even more surprisingly, there are five people that ran precisely the same time in 2022 as they did in 2018/9. These masters of consistency are acknowledged in the table. Interestingly, they all earned Vic Claphams and four or the five started in G batch.
Wanderers runner Michael Fleshel improved his 2018 down run time by an impressive 1:45:39 in 2022. He asked me whether I could do some “PB stats comparing 2022 and 2018”. I was surprised to find he was only 155th on the biggest improvements list between the last two down runs.
So, thanks to Michael, here’s a more detailed comparison of the Comrades’ repeat offenders (those who ran in 2022 and 2018/9).
Whilst the graph highlights that most people run very close to their previous runs (whether up or down), there are still about 3000 runners who finish over an hour faster or slower in 2022 than in 2018/9 – including a few people who breach the 4-hour difference mark!
The tables below give you the Top 30 hall of famers on both sides of the equation. I suspect that many of the improvements are due to acquiring a good coach. For those that feature on the deterioration list, full credit for hanging in there for another medal (or perhaps with the increased entry fee, you just wanted to make sure you got value for money).
The final tables show the biggest improvement and deterioration in 2022 by medal earned for men and women.
Qualifiers
Below are the most popular qualifiers for Comrades 2022.
The most popular was World Marathon Majors aspirant, the Sanlam Cape Town Marathon, with 1048 entrants using the October race as their qualifier. The race was also one of the fastest average times and best finisher rates (more on this aspect later in the article).
The other events scoring over 500 entrants were all races held in the last few months of qualifying: Dolphin Coast (947), Benoni Northerns (895), Wally Hayward (710) and Edenvale (592).
Three ultras make the Top 10 – Total Sports Two Oceans Marathon (6th), Om Die Dam (8th) and Irene (9th) but no other ultra makes the Top 20. Two Oceans (92.8%) and Om Die Dam (93%) also have the best finish rates.
Unsurprisingly, the list is dominated by flat and downhill races but, what is somewhat surprising, is that Uniwisp Kaapsehoop is missing from the list (they had a Covid-reduced 2021 entry field). I’ve run most of the marathons on the list and, for my money, the Tony Viljoen Masters Marathon is the easiest in the country.
The second table in this section is a list of the Top 10 international marathons that were used as qualifiers. Neighbouring Gaborone claims top spot with 63 entrants but none of the majors feature in the Top 5 – which are Pune, Jaipur, Manchester and Victoria Falls respectively.
Qualifiers by Month compared to Finish Rate and Time
The graph below highlights that the overall trend is that average qualifier time increases from September 2021 to July 2022 and there is a correlation in the number of runners who start the race but fail to earn a medal. There are some months (especially September and January with very low sample sizes) but I was really surprised that October (10 months before race date) has the fastest average qualifier time of high-volume months and the second lowest bailer rate. The lowest bailer rate is in November (and this is also one of the fastest qualifying months). I guess the early bird gets their qualifier done and the medal!
I would have expected qualifying time and finish rate to peak between March and May as runners put everything into a fast qualifier to minimise the impact of tapering and recovery on their overall Comrades training. However, as depicted in the graph below, over half the field (56%) qualify in the final three months.
The third graph in this section shows that success rate generally declines and Comrades finish time increases the later you leave your qualifier. The lowest success rate (85%) are lastminute.com runners who qualify in July.
Only September has an average finish time below 10 hours. This is explained by a low sample size (just 25 runners) and that nearly all of these are international runners (Berlin and Lima Marathons) who generally have faster qualifying and finish times than local runners – if you’re spending all that money travelling to Durban you want to be sure you return with a medal!
The table below contains all the source data. With a June 2023 race date, I expect next year’s data will look very different.
Note: Ultra marathon qualifying times were adjusted a distance adjustment equation (see below) derived from Jan Louw’s Two Oceans / Marathon adjustment scale to predict Comrades times.
Qualifiers by Province
Whilst Gauteng houses over 43% of the entrants, only 31% of the 2022 Comrades entrants run their qualifiers in Gauteng. The big winners in attracting Vaalies to their start lines are Western Cape (9% of entrants vs. 18% of qualifiers), Mpumalanga (4% of entrants vs. 8% of qualifiers) and North West (2% of entrants vs. 4% of qualifiers).
Gauteng, KwaZulu Natal and Western Cape collectively account for 70% of the qualifiers submitted. On the other end of the scale, the Northern Cape has just 1% of the Comrades runners using their qualifier marathon and Free State qualifiers are also scarce at 2%.
I’ve run most of the Northern Cape’s marathons – they are flat and friendly but not so good for finishing Comrades. The data shows that almost 20% of Northern Cape qualifiers don’t earn a medal in 2022. Limpopo, another province that specialises almost exclusively in flat marathons, also has a high bailer rate (16.2%).
The best provinces to run your qualifier if you want to earn a medal are the North West and Western Cape (7.4% and 8.1% failure rate respectively). The Eastern Cape, Mpumalanga and Free State also have good success rates (about 90% of people using their marathons to qualify finish Comrades 2022).
However, the lowest failure rate are marathon qualifiers from other African countries (5%) and International runners (7.5%) are also good for a medal. International runners also have the lowest average marathon qualifier time (3:55:00) narrowly followed by the Eastern Cape and KwaZulu Natal.
There’s definitely some home town advantage with KwaZulu Natal qualifiers resulting in the fastest average Comrades finish time (10:06:44) but running marathons at altitude in Gauteng does nothing for your finish time – Gauteng qualifying runners have the second slowest average finish time (10:32:35) behind Northern Cape (10:35:11).
Distance and Completion Rate
The general trend is that the longer the race you qualify with, the faster your time and the better your chance of earning a medal.
There are however two anomalies:
- Ironmen (and women) have the second lowest finish success (spending all that time on the bike obviously makes you soft).
- The 52-54km category has the lowest finish rate but the fastest average finish time (9:59:55 – the only category below 10 hours). The low finish rate is largely impacted by two KwaZulu Natal 52km races, Arthur Creswell (82.4%) and Prince Mangosuthu (84.7%). Both these races are point-to-point with fairly small fields so it’s possible that some taxiing took place to cheat on qualification times.
There are 12 people who used a race of 60km or longer to qualify and this dirty dozen all make it safely to Durban.
However, picking a tough qualifying marathon has no impact on your medal earning probability. I marked off the six toughest marathons used as qualifiers and compared the results of these 114 masochists to the class average: There is very little difference in success rate and eventual finish time.
Does this mean anything? According to Data Scientist Andrew Collier, there is no statistical significance between qualifying distance and Comrades success rate but there IS a significant correlation between the month you qualify and success rate (as illustrated earlier in the article – the earlier you run your qualifier, the better your chance of finishing).
Qualifiers with the Best and Worst Finish Rates
The table below contains the races that had the best finish rate percentage. False Bay 50 was top of the pops with an amazing 96.6% success rate followed by Weskus and Kaapsehoop. Of note is that there are three other ultras on the list (Swasap, Om Die Dam and Two Oceans in 4th, 5th and 6th) and that there are only two late (June or July) qualifiers on the list.
Data Scientist Andrew Collier has confirmed that the month and distance combo is significant when it comes to predicting a Comrades finish – running an early qualifying ultra leads to a better chance of earning a medal.
The second table shows the other side of the coin – the qualifying races with the lowest Comrades success rate. Most of the races are in the last two qualifying months and there are just two ultras on this list.
Faheemah Limbada contacted me as she was curious who used 100 kilometre races to “qualify” for Comrades. The table of “Show offs” is below containing the nine international runners who used races of 60km or above to qualify. The most notable is at the top of the table where former winner and 6th placed 2022 finisher Camille Herron used her 100k split (on the way to breaking the 12 hour and 100 mile world record) at the Jackpot Ultra as her qualifier!
There were three South Africans who used 60km+ races as qualifiers – two used Kosmos 3-in-1 which is bending the rules a little (as the 42, 21 and 10 are three separate runs albeit on the same day) and one person used Comrades itself (not sure how they slipped that one through). However, all three of these runners finished.
Below are the Top 10 qualifiers for each batch. Cape Town is #1 for batches A-D and Benoni Northerns is #1 for batches E-H.
A seeds who used Two Oceans as their qualifying have by far the fastest average finish time (6:43) followed by AmaRavens Best of the Best (7:04). At the other end of the field, Two Oceans (88.2%) and Om Die Dam (82%) result in by far the highest successful finish rate for H batch runners.
Finish Time versus Qualifying Time Prediction
The popular way to predict your Comrades finish time is your qualifying marathon time * 2.5 (42.2km * 2.5 = 105.5km). According to Lindsey Parry, elite runners can more accurately predict their optimum Comrades time with an equation of qualifying marathon time * 2 plus 1 hour. So how did the class of 2022 do according to their potential?
The graph below shows that the *2.5 is a fairly good forecaster of finish time with 50% of the field finishing within 30 minutes of prediction. Just under 32% of the field beats the *2.5 predictor with women being much more successful than men at exceeding their potential.
However, the *2 plus 1h calculation is beyond the reach of most mortal runners – over 97% of the field lose to this predictor. Interestingly, 80% of Gold medal winners are within 15 minutes of their predicted time. For more analysis on this is, see the male and female elite articles.
The Best at Beating Predicted Time
The traditional way to predict the finish time of a recreational runner is marathon time * 2.5. The previous section showed how the field fared on this predictor with only about 32% managing to beat the equation.
It’s easy to blow on race day and lose to the equation by several hours but I was surprised at how many runners were able to beat their predicted time by an hour or more. The table shows the Top 50 runners who came in under par at Comrades 2022.
Some logical explanations would be that E and CC runners (who have a guaranteed seeding) just use the first marathon they do in the qualification period and don’t stress about updating the qualifier time later in the year when they are fitter, some runners only get the chance to run one qualifier (due to finances or illness) and have a shocker or runners like Colly Ronald Mahumani (4th on the list) who use an incredibly difficult marathon like the Tzaneen Tuffy as their qualifier. There is also the possibility that some banditting occurred (running in someone else’s number).
A couple of notable runners on the list are Fanuel Mapamura (he was the guy who had to tap Camille Herron on the shoulder in 2017 to tell her that the finish line was still a few hundred metres ahead around the corner) and well known trail runner Simon Tshabalala (who recently defended his 100k Sky Run title).
Anything Other Requests?
If there’s any other stats you’d like to see for the 2023 Comrades Marathon drop me a message and I’ll see what I can do.
Follow Running Mann:
This is exceptional work. Can we get a PDF document?
The blog format does not make it easy to convert to .pdf (other than doing a cut and paste) but if there is something specific you’re after I can send it to you.