SB

Should Mondo Duplantis do the decathlon?

Mondo Duplantis is easily one of the greatest athletes of his generation, if not ever. At 26 years old, the Swedish-American pole vaulter has won 2 Olympic gold medals, 3 world championships and holds 15 world records in pole vaulting. It is an understatement to say he dominates the event. What makes him remarkable is not just how high he jumps, but how consistent he is: the record he keeps breaking is his own.

That consistency is why I think he should try a second discipline. Much like Michael Jordan, who famously left basketball at his peak to play professional baseball. But where Jordan swapped sports entirely, Mondo would move into something that still rewards what he does best: the men’s decathlon. It already includes pole vault and with his all-round explosiveness, he would be competitive in most of the rest too. Let’s dig deeper into how competitive he could actually be.

The 10 Disciplines

The decathlon is a combined event consisting of 10 track and field events. It typically takes place across two days, where athletes compete for points based on their time, distance or height. The decathlon combines four runs, three jumps, and three throws:

Day 1:

  1. 100 metres (run)

  2. Long jump (jump)

  3. Shot put (throw)

  4. High jump (jump)

  5. 400 metres (run)

Day 2:

  1. 110 metres hurdles (run)

  2. Discus throw (throw)

  3. Pole vault (jump)

  4. Javelin throw (throw)

  5. 1500 metres (run)

After 2 days and 10 disciplines, the person with the most points gets declared the winner and can refer to themselves as the “World’s Greatest Athlete”. This title was introduced in 1912 after King Gustav V of Sweden told the then winner of the decathlon Jim Thorpe “Sir, you are the world's greatest athlete”. This title is why I believe Mondo Duplantis not only should, but will compete in the decathlon. Before we show that Duplantis can do that, let’s look in more detail at the points system.

The Points System

The points are calculated based on two formulas, one for time values (track events) and one for distance values (field events);

Points_track = floor(A * (B - P) ^ C)

Points_field = floor(A * (P - B) ^ C)

Below is the table of parameters:

Event

A

B

C

100 m

25.4347

18

1.81

Long jump

0.14354

220

1.4

Shot put

51.39

1.5

1.05

High jump

0.8465

75

1.42

400 m

1.53775

82

1.81

110 m hurdles

5.74352

28.5

1.92

Discus throw

12.91

4

1.1

Pole vault

0.2797

100

1.35

Javelin throw

10.14

7

1.08

1500 m

0.03768

480

1.85

This formula ensures that athletes get a similar amount of points for performing better/according to benchmarks in each event. For example if we compare Usain Bolt’s 100m WR with Mondo Duplantis pole vaulting record we get the following:

Points_Usain = floor(25.4347 * (18 - 9.58) ^ 1.81) = 1202

Points_Duplantis = floor(0.2797 * (630 - 100) ^ 1.35) = 1331

Both numbers are roughly in the same ballpark and equate two very different disciplines with each other. In other words, a world record in the 100m and a world record in pole vault score about the same, which is exactly what the system is designed to do. If we now use these formulas to look at how Mondo would perform, we will have to do some estimation and research.

Looking At Mondo Duplantis

The method

How can we assess how Mondo Duplantis would perform without just plain guessing? The answer is simple: Monte Carlo simulation. For each event, we assume a distribution with a mean and standard deviation (uncertainty) and sample a performance out of this distribution. After sampling, we calculate the scores and add them up. If we repeat this process many thousands of times, we can get an estimate of how well Mondo will perform. To keep things simple, we will assume that the events are uncorrelated, while they are obviously not. An athlete that performed extremely well in a previous event will be much more confident and likely perform better on the next event. To perform this simulation, we will need some data.

The data - what we know

In the case of Mondo Duplantis, we have few data points to look at:

  • For pole vaulting, we can use his competitive records and it is not unlikely he will jump 6m or more

  • For 100m, we can use the 10.37s time when he competed against Karsten Warholm in Zurich (video)

  • For the other disciplines, we will need to look at historic records and other similar athletes to get to an estimate

Unfortunately for us, Mondo only competed in 100m dash, 200m dash (which is not part of the decathlon) and long jump events during high school:

  • 100m dash: 10.57s

  • 200m dash: 23.03s

  • Long jump: 7.15m

Finding similar athletes - what we don’t know

To estimate his performance on the events where we don’t have a single data point, we will have to get creative. That’s where embeddings come in, an embedding is basically a mathematical way of representing a real-world object. They are typically used in natural language processing to compare words, for example man and king are closer to each other than man and queen. Rather than designing a complex new embeddings space, we can use the points system as a way to calculate the embeddings for every athlete in our database. Then we use Mondo’s 100m dash and long jump to estimate which athletes are the closest:

Nearest neighbour

distance in embeddings space

100 m

Long jump

Shot put

High jump

400 m

110 hurldes

Discus

Pole vaulting

Javelin throw

1500 m

Sven Roosen

0.26

10.49

7.48

15.03

1.91

46.64

14.18

43.88

4.70

64.04

4:20.77

Averyanov

0.27

10.55

7.34

14.44

1.94

49.72

14.39

39.88

4.80

54.51

4:31.02

Bryan Clay

0.43

10.39

7.39

15.17

2.08

48.41

13.75

52.74

5.00

70.55

4:50.97

Eduard Hämäläinen

0.43

10.50

7.26

16.05

2.11

47.63

13.82

49.70

4.90

60.32

4:35.09

Kip Janvrin

0.46

10.61

7.34

14.64

1.96

48.20

14.48

45.84

5.00

61.82

4:20.12

Tomas Järvinen

0.50

10.61

7.48

12.88

2.09

47.69

14.18

43.76

4.90

62.20

4:35.82

Maurice Smith

0.56

10.62

7.50

17.32

1.97

47.48

13.91

52.36

4.80

53.61

4:33.52

Kristjan Rahnu

0.58

10.52

7.58

15.51

1.99

48.60

14.04

50.81

4.95

60.71

4:52.18

Siegfried Stark

0.58

10.52

7.58

15.51

1.99

48.60

14.04

50.81

4.95

60.71

4:52.18

Paul Meier

0.60

10.57

7.57

15.45

2.15

47.73

14.63

45.72

4.60

61.22

4:32.05

Lindon Victor

0.60

10.60

7.55

15.94

2.02

48.05

14.47

54.97

4.80

68.05

4:39.67

Lev Lobodin

0.61

10.66

7.42

15.67

2.03

48.65

13.97

46.55

5.20

56.55

4:30.27

We will use these athletes to estimate Mondo’s performance where there is limited data available. The biggest areas of uncertainty are the throwing events and the longer distances, so that is where these comparisons matter most.

Setting the distributions

Now that we have the data, it is time to set the distributions we will sample from. We will assume skew normal distributions. This is a type of normal distribution that is slightly skewed towards one direction. This way, we stop Mondo from breaking world records in the simulation. It introduces one additional parameter alpha that indicates the skewness, positive alphas trend more to the right, while negative alphas trend more to the left. This gives us the tools to play with when estimating the parameters. We will discern 3 main categories:

Category

Description

Events

Strategy

Certain

Events where we have good estimators for Mondo

100m dash, pole vaulting, long jump

Use the data to set the mean, with relatively low variation and skewness

Good estimates

Events where we can reasonably assume based on the data

400m, high jump, 110m hurdles

Extrapolate based on similar athletes and set medium variation

Bad estimates

Events where there is no data we can base ourselves on

1500m, javelin throw, shot put, discus throw

Extrapolate based on the similar athletes and set extremely high variation, skewing to lower numbers

Based on this the following distributions are set:

Event

Mean

Standard deviation (uncertainty)

Alpha (skewness)

100 m

10.50 s

0.12 s

1.04

Long jump

7.39 m

0.22 m

-1.35

Shot put

14.97 m

1.80 m

-6.30

High jump

2.01 m

0.08 m

-1.81

400 m

48.21 s

1.10 s

2.38

110 m hurdles

14.26 s

0.45 s

3.23

Discus throw

46.83 m

6.50 m

-9.34

Pole vault

5.99 m

0.22 m

-1.65

Javelin throw

59.74 m

7.50 m

-9.34

1500 m

278.7 s

14.0 s

9.34

When we look at this visually for the shot put and the 1500m, we see a clear pattern emerging: there is a steep drop-off in performance above that of the world-class performers and a gradual decline for worse performance.

Two skewed distributions

There is one more thing to keep in mind before we start simulating. All the numbers we have for Mondo, like that 10.37s over 100m, come from a body trained purely for pole vault, run fresh on the day. The decathlon is different. Athletes run the 100m on the morning of day one and the 1500m at the end of day two, tired from nine events before it. On top of that, training for ten events means Mondo could not train pole vault the way he does now. So our estimates are likely a little optimistic: they show what his body could do in isolation, not what it would do across two exhausting days. Keep that in the back of your mind as we go.

Running the Monte Carlo simulation

Once we have set the distributions, we can run the Monte Carlo simulation. To get a smooth line, we will run the simulation 200,000 times and see what we get out.

The results give a clear picture: Mondo could indeed be a world-class decathlon athlete. To put it in context, Markus Rooth won gold at Paris 2024 with 8796 points, ahead of silver on 8748 and bronze on 8711. Mondo’s simulated mean came out at 8788, almost exactly on that podium. In other words, an average performance from him would already be fighting for a medal, and in some simulations he even exceeded the current world record.

But there is a big caveat. The whole analysis leans on one assumption: that Mondo could throw about as well as decathletes who share his speed and jumping profile. That is far from guaranteed. Being similarly fast does not make you similarly good at the shot put, discus or javelin, which are technical events built over years. His throws are clearly the weak point, and if he lands well below his comparable athletes there, the whole total drops with them. The optimistic scenarios assume he can close that gap. The pessimistic ones assume he never quite does.

Conclusion

So, should he do it? The numbers say yes. With his pole vault and his explosiveness off the ground, Mondo would already be a serious contender, and in some of the simulations he even cleared the current world record. The throws are his weak spot, and the two-day grind of the 1500m won't come naturally to a vaulter. But those are gaps you close with training, not talent he lacks. The real question isn't whether he can. It's whether he wants to. Chasing the decathlon means taking time away from the event he already owns, and possibly giving up a few centimetres on the pole vault he has spent his life perfecting. That's the trade: stay the greatest pole vaulter of all time, or risk some of that to chase the oldest title in the sport. My bet is that he does it. Not next season, maybe not for years. But the "World's Greatest Athlete" title has pulled in the best before him, and Mondo has never been someone who avoids a bar just because it's set high - rather he raises it, 1 centimeter at a time.

Disclaimer: the views posted on this website are my own and are not representative of Belfort or any other entity.