real biathlon
    • Athletes
    • Teams
    • Races
    • Seasons
    • Scores
    • Records
    • Blog(current)
    • More
      Patreon Content Videos Explanations Shortcuts
      Error Report
      Privacy Policy About
    •     
  • Forum
  • Patreon
  • X
  • YouTube
    Instagram
    Facebook

Recent Articles

  • Junior + U23 Development Index – Men
  • Junior + U23 Development Index – Women
  • The Path to GOAT Status
  • Most improved athletes this winter
  • New biathlon point system

Categories

  • Biathlon Media
  • Biathlon News
  • Long-term trends
  • Statistical analysis
  • Website updates

Archives

  • 2024
    • March
  • 2023
    • March
  • 2022
    • December
    • June
    • May
    • March
    • February
    • January
  • 2021
    • December
    • November
    • September
    • July
    • June
    • May
    • April
    • March
    • February
    • January
  • 2020
    • December
    • November
    • August
    • June
    • March
  • 2015
    • December
  • 2013
    • August
    • July
  • 2012
    • July

Search Articles

Recent Tweets

Tweets by realbiathlon

Author: biathlonanalytics

Proud dad&husband; analyst & visualization specialist (Tableau, SQL & R); creator of Biathlon Analytics; blog poster on realbiathlon.com; passionate about biathlon, cross country skiing and canoeing

A follow-up on going the distance.

Posted on 2021-05-06 | by biathlonanalytics | Leave a Comment on A follow-up on going the distance.

On March 14 I wrote an article called “Is the IBU going the distance” in which I found that biathlon races typically don’t go the exact distance set for race events. The following is a follow up on that article, using GPS data from athlete Baiba Bendika (follow her on Twitter or Instagram) of her two races in Nove Mesto and two races in Oberhof. I should note that the GPS data from her Strava account (she kindly exported and send them to me) comes from her watch, not sure which brand, but the accuracy will not be perfect. But then again, it’s the best I can get when it comes to tracking the course.

First I read her data in Tableau and manually divided her GPS measurements (points) into sections like lap 1, penalty lap 2, range 2, etc.

This also allowed me to see her race in a “flat line” based on distance raced and time elapsed:

We already know how many penalty laps Baiba had, so we can take out the time spent in the penalty loops. This leaves us with the following distances:

Even if we include the penalty loops and decrease the distances by 150 per loop, we get:

  • Nove Mesto on March 6: 7,307 – 600 = 6,707 (793 m under)
  • Nove Mesto on March 12: 6,948 – 150 = 6,798 (702 m under)
  • Oberhof on January 8: 8,186 – 300 = 7,886 (386 m over)
  • Oberhof on January 14: 8,281 – 450 = 7,831 (331 m over)

Now we know that her watches’ GPS is not perfect, and that some of the penalty loop section (that overlaps the course) should be included, but even then our previous conclusion that the IBU doesn’t take the distance requirements all too serious is still correct, going both under (Nove Mesto) and over (Oberhof).

The main reason why this is important is that a lot of stats like ski speed and range time are often based on the assumption that ski distances and range sections are equal from race to race and location to location. Clearly, they are not and actual distances should be accounted for when available.

(Thank you Baiba for sharing your GPS files with me!)

Posted in Statistical analysis

It’s all for the money. Or is it?

Posted on 2021-04-19 | by biathlonanalytics | 1 Comment on It’s all for the money. Or is it?

Introduction

Recently Real Biathlon posted data about World Cup Prize Money in biathlon on his website for Patreon members. Combined with all the other data sources he provides, if you like working with data and biathlon why not look into becoming a member as well and support him in keeping the site running? For the Prize Money data, you can easily find the major all-time earners, sort by gender and year, and use the data combined with other data sources he provides.

In the post below I have tried to look at the data from the less obvious perspectives, and see if we can get some interesting stories out of it.

Data source

The data contains all income from the IBU at the World Cup level, and excludes any income made from external sources like sponsorships, bonuses from national federations, etc. the data range starts with the 2003-2004 season and it should be kept in mind that some athletes were already active before this season so that data is not included. For team races, the data is based on the assumption that prize money is split entirely and equally among the participating athletes.

The Prize Money data source is in Euros as that’s the currency used in the IBU season guides, the source of this data for most (or all) seasons.

Equality

The first thing I wanted to explore is how the IBU deals with equality. And it’s fairly simple: prize money is based on race results, world cup scores, and bibs. And it is unrelated to the gender of the athlete; in other words, the prize money is equally handed out between man and women.

The small changes are likely happening as in some races more athletes get prize money, which then adds up for the season (say three athletes don’t finish a mass-start, the prize money is not handed out). The 2014 season seems to be the only year with a bigger difference, and quite frankly I don’t know why that is. But generally, we can see the pink line for the women pretty much overlaps the men’s blue line.

How is the money divided?

If we look at how the prize money is divided, we see that the majority of athletes do not make a lot from their sport (strictly looking at prize money). Of the 561 athletes in this dataset, only 27 (4.8%) have made over a million where 241 made €10,000 or less. And 428 athletes (76%) have made €100K or less since 2003-2004. It also appears the men are a bit more top-heavy with 26% making €100,000 or more, which is only 22% for the women.

Earning pace

I was curious to see how athletes compared when looking at how much they earned in their first, second, etc. year in the World Cup. I should remind you that if athletes were already active before the 2003-2004 season, their total earnings do not include the seasons before that.

We can see a number of interesting things here, for example that Bjoerndalen didn’t make money until his 12th season (due to the above-mentioned limit in available data), Martin Fourcade made the most in total but could be caught up with by JT Boe, and Jacquelin, Dale and Laegreid are also on pace to get to Fourcades total faster or go above it. Hanna Oeberg and Denise Herrmann are on track for the women to become the biggest prize money earners, although Herrmann started later in her career with biathlon.

Another way to look at this data is average money made per race. I filtered the chart to only show those athletes that have between 30 and 150 races (to highlight the “new kids on the block”) and have a average of at least €1,500 per race:

After Laegreid’s amazing first full season, it is no surprise to see him high up in the top-left corner. Dale, Jacquelin and Ponsiluoma are also in a strong position to make good money in the upcoming years if they can keep their rates going. For the women, Hanna and Elvira Oeberg, Herrmann, Tandrevold, Simon, Vilukhina and Davidova also have great ratios up to this point in their careers.

Riding the good team train

The last thing I wanted to look at are athletes that have made good money in Team races, but not so much in individual races. In other words, who has been riding the good-team train?

The chart below shows all athletes and their total prize money for team/relay races, and the percentage of earnings for team/relay races compared to their total earnings. To highlight the “train riders” (filled in circles) I selected those athletes who made more money with team/relay events than other ways (team/relay earnings > 50%) and who made more than €40,000 on team races:

Below are the highlighted athletes with more details. We can see that an athlete like Merkushyna made 81.5% of her earnings through team/relay events for a total of €51,750. Given that she has not made much money in individual races, I think she is one of the athletes who most benefitted from being on a good team. Sophie Bailley is another great example.

This Prize Money dataset is a great addition to the site, with plenty of interesting stories to tell!

Posted in Biathlon Media, Statistical analysis | Tagged Prize Money

Canadian biathletes finding balance

Posted on 2021-04-06 | by biathlonanalytics | Leave a Comment on Canadian biathletes finding balance

With three top 10s, sixteen top 20s and twenty-five top 30s in individual events, the 2020-2021 season was the best Canadian season in the last five years. After a dip in total World cup points in the 2019-2020 season, last season continued on the five-year trend of continuous increase in total world cup points:

So what did Canada do well to get here, and where is there still room for improvement? That is what the following review of the Canadian individual performances in the 2020-2021 season will analyse! In this analysis, I will only show some examples of Canadian performances, but a supportive dashboard with all used data for this analysis can be found on my Tableau Public page. It shows all stats for the Canadian team, both men and women, and has a second tab where a Canadian athlete can be selected to show his or her specific stats.

Skiing time

The skiing is still the area where the Canadian team overall can gain the most time. In the specific example below I compare the average ski times of the Canadian athletes to the field average, the average of the top 30 athletes, and the average of the top 10 athletes:

Although we can see Canada is closing up the gap a little compared to last season, we are still behind the overall average, and far behind the top 10 athletes. But the positive news is that we made great progress in the 2020-2021 season, indicated by the steep drop towards the averages.

Shooting time

This is where Canada has the least to gain (and the most to loose) as we continue to be very strong in fast shooting, leading many of the true biathlon nations:

Shooting accuracy

In previous seasons the fast shooting times by the Canadians often lead to less desired shooting percentages, but it appears the team and coaches are starting to find a balance between shooting fast and shooting accurate, demonstrated by both increased accuracy in Prone and Standing:

Shooting speed -vs- Accuracy

This balance between shooting fast and clean is of course the golden grail of shooting in biathlon (I wrote about this in more depth on this website in December last year). The chart below shows how this balance has changed for the different team, starting in 2016-2017 (thin line) and how teams have moved from shooting more or less accurate, and faster or slower. We see that in 2018-2019 team Canada put a lot of focus on shooting faster, which worked, but at the expense of less accuracy. Now we see that we have lost some speed (but are still very fast) but have gained accuracy. Hopefully, this trend can continue upwards towards more accuracy while remaining fast.

Individuals

The second part of the evaluation is looking at the individual athletes. All Canadian athletes can be viewed on the interactive dashboard, but here I’m letting Emma Lunder be the example here. We can see that here Accuracy -vs- Speed trend is similar to the Team’s and that she is gaining accuracy while giving up a little speed. But we can also see that here range, shooting and prep time (prep = range – shooting) is still very good and well below the field’s average:

Her total percentage of shooting accuracy (T) has increased mostly based on a strong increase in her prone shooting (P) and she well above the field’s average Total shooting percentage (T avg.). Lastly her skiing speed seems to have stabilized and is on par with the field’s average. But she is also closing the gap with the Top 30 and Top 10 averages.

Conclusion

As mentioned team Canada had a great season, as did Emma Lunder. The above will hopefully give a good explanation of how to read the interactive dashboard to see both men’s and women’s team performances for individual races, and the individual dashboards for any of the Canadian athletes. If the current trends continue and with some of the younger athletes coming through the pipeline, hopefully, we can some Canadian podiums in the next season! Go Canada!

Posted in Biathlon Media, Long-term trends, Statistical analysis

Wierer in pursuit… of my mind

Posted on 2021-03-28 | by biathlonanalytics | Leave a Comment on Wierer in pursuit… of my mind

In early March I wrote a piece about Wierer’s Pursuit efforts and results, and although I still stand by what I wrote it somehow felt incomplete. In the latest Extra Runde podcast, they brought up again that Wierer seems to do better when starting with a lower bib number in the Pursuit. It feels right, but I couldn’t find the data to support that for Wierer specifically. I did find that the later you start, the more places you can, and typically will, make up, but that applied to all athletes.

Then it struck me (yes, I’m a bit slow sometimes…) there was another way to measure performance in the Pursuit that could be helpful. Look at the Isolated Pursuit time, or in other words the “race time” – “seconds behind at start”. You can also call it the actual race time, and it’s a simple calculation: Total Time – Start Info. The latter has the time an athlete started behind Bib nr. one, the winner of the prior Spring race. Having this Isolated Pursuit time and the ranking of this time would show me her true performance of the day. Plotting that for the last three seasons gives me the following chart:

Now we can see if her start bib is higher her Isolated Race results vary between very good and very bad. Her races with a bib number higher than 15 though are all good to very good for the Isolated Race performance.

Values show change in ranking of isolated race result and start bib (positive = improvement)

If we only take her 8 Pursuit races in the 2020-2021 season we can see she had her best performances when starting lower than 5th and two of her three worst performances when starting 5th or higher. I think it’s fair to say, although based on a small sample size, that the guys from Extra Runde were correct in their assumption. Starting later in the Pursuit races brings up the better performances in Dorothea Wierer. Now they have some data to prove it.

Data from RealBiathlon.com, Feature image from Manzoni/IBU

Posted in Statistical analysis

Stina, look at what you made me do!

Posted on 2021-03-28 | by biathlonanalytics | Leave a Comment on Stina, look at what you made me do!

After watching the final season races in Oestersund, and one moment in one race specifically, I wanted to do further analysis into the time that athletes prepare themselves for shooting and for skiing after the shooting. And see if this would be even possible. But first, let me show you the moment that triggered this:

Stina Nilsson takes FOREVER to put her poles back on and get back to fully functioning skiing after her prone shoot in the Women’s Sprint race in Oestersund. It made me wonder if there is a way to analyze how fast or slow athletes are outside of the shooting while not skiing on the course. Taking poles and rifle off, getting in position, getting the rifle and poles back on and getting skiing again.

Data

The data I hope we can use for this is the Range time and the Shooting Time. As I’m interested in the time on the Range but while not shooting, it’s a simple subtraction: Range time – Shooting time = Prep time. So Prep time is the time spent in the Range while not shooting. That would be the time described above, getting off the poles and rifle, getting ready, and then getting moving again.

The only problem is that there are athletes like Stina Nilsson that take so long to get the poles back on that they already have left the range. Now I’m sure this happens more often (especially for those athletes shooting in the lane closest to the penalty zone at end of the range) but I don’t recall ever seeing an athlete going past the time recording and still having to start putting on the second pole!

Can we use this data then?

Ironically when I look at Stina’s race data, her Range time rank is 57th (103.1 sec.) and her Shooting time rank is 70th (62.3 sec.). But she does have the fastest Prep time of the whole race. We know that a) her prep goes well beyond the range, and b) there is logic in that the more time you spend shooting while in the range, the less time you have for Prep while in the Range. Considering this, the Range and Shooting time are not able to answer who is faster and slower at “prepping”, and I don’t think there is data available that can actually answer this question.

Alternative insights perhaps?

What is still interesting though is to look at how much of the Range time is used for shooting. At least that will tell something about the time spent not shooting while in the Range. In this specific race, it varied from 33% to 53% of Range time, which tells you there is still a lot of time to be gained by either skiing / gliding a bit faster in the range before the shooting, getting in position faster, and getting up and going after the shooting.

Phases

Another fact we cannot get from the data is that the phases before and after the shooting are very different. The phase before shooting is focused on slowing down, reducing the heart rate and focussing on the shooting. In other words, this tends to be slow. The phase after the shooting is fast and all focussed on getting ready and on to the course to ski.

Depending on what lane the athlete shoots in, the distance spent in these two phases while inside the Range area differs significantly. This also means we cannot look at these data for one race and draw any conclusions. For all races in three seasons, however, this should even out to some degree, although I would expect that the top athletes who shoot in the last lanes most of the time spend more time in the slow phase while in the Range area.

Alternative insights perhaps? – Part II

Sticking with the women’s field but including all races since the 2018-2019 season, we now see the average percentage of Range time spent shooting varies from 45% to 70%. Or 30% to 55% spent not shooting. Let’s check that assumption that higher-ranked athletes would be slower because they relatively spend more time in the slow phase due to their shooting lane. first I look at the average rank and average time spent shooting per athlete:

This only shows us that as the Ranks get higher there is more variability between the athletes. How about the same but per athlete and race:

The trend line is barely statistically significant (p ~ 0.05%) but the R-square value tells me not much variation is explained by the model.

What ever it does and does not tell me, my question if higher ranked athletes are spending more time in the slow phase and thus slower in the time spent not shooting is not answered. On top of that, higher ranked athletes are generally speaking better shooters, both accuracy and speed, so that is another variable in the mix.

Conclusion

My sad conclusion is that based on the available data we cannot make any conclusions about how fast or slow athletes are prepping for shooting and then for skiing. If we would have data that starts from the moment the athletes prepares for shooting and continues until an athlete is in full skiing position, I don’t see how I can come to any conclusion on this topic. Perhaps I should rename this piece “Ramblings about Preparation time.” Nah. Hopefully, these ramblings will trigger something with readers and who knows where that may lead?

Posted in Statistical analysis

Posts navigation

Older posts
Newer posts

Recent Articles

  • Junior + U23 Development Index – Men
  • Junior + U23 Development Index – Women
  • The Path to GOAT Status
  • Most improved athletes this winter
  • New biathlon point system

Categories

  • Biathlon Media
  • Biathlon News
  • Long-term trends
  • Statistical analysis
  • Website updates

Archives by Month

  • 2024: J F M A M J J A S O N D
  • 2023: J F M A M J J A S O N D
  • 2022: J F M A M J J A S O N D
  • 2021: J F M A M J J A S O N D
  • 2020: J F M A M J J A S O N D
  • 2015: J F M A M J J A S O N D
  • 2013: J F M A M J J A S O N D
  • 2012: J F M A M J J A S O N D

Search Articles