Some autorenewing subscriptions have failed to automatically renew. If you find you can't read subscriber articles, or think this applies to you please read this.

Crunching The Numbers: Do Crash Statistics Show That Sprint Races Made MotoGP More Dangerous?

By David Emmett |

Fabio Quartararo and Miguel Oliveira colllide at the start of the race at Jerez

It certainly felt like there were a lot of crashes during the 2023 MotoGP season. That impression was reinforced by the fact that 2023 did not see a single MotoGP race on Sunday with a full grid. At all 20 of the Sunday grand prix races, there was at least one, and usually multiple riders missing, with replacement riders getting a lot of extra work.

Impressions are one thing, but do the numbers really reflect that? On Tuesday, we received a neatly formatted report on all of the crashes which took place in the three grand prix classes in 2023, and comparison charts with previous seasons. The 2023 Final Falls Report, as it is called, seems to bear out that, yes, there were more crashes in the MotoGP class last year than usual. And the rise appears to be confined to the MotoGP class.

Table showing the total falls across all 3 grand prix classes between 2010 and 2023

Here are the numbers of crashes per class going back to 2010. (Technically, these are 'falls', which is defined as 'the bike physically touching the ground and remaining on the ground' by Dorna. That can be due to someone running wide and into the gravel, tipping over, then getting back on and carrying on, but it also includes incidents such as the major first corner pile up at Barcelona, with multiple riders involved and some sustaining injuries.)

To make seasons comparable, Dorna has also averaged the number of crashes per event for each class and each year. As you might expect in a packed and dramatic class full of youngsters, the numbers for the Moto3 class vary wildly per season, from a high of 23 crashes per event in 2016 to a low of 15.4 per event in 2023. The same is true for the Moto2 class, though the variance is less, between 24.1 per event in 2017 to 17.2 last year.

Overall, the numbers for the MotoGP class are much lower. This reflects both the lower number of entries in the class – usually between 22 and 24 riders, instead of between 28 and 32 – but also a greater divergence in the relative performance between machines. And there is an experience and skill gap as well: the point of MotoGP is that it should contain as many of the best, most talented, and most experienced riders in the world as possible.

Yet there are clear signs here too that things got worse in the MotoGP class in 2023 compared to previous years. If we convert the crashes per event into percentages compared to other classes, then the data is a lot clearer.

  Crashes per event Percentage compared to MotoGP
Year MotoGP Moto2 Moto3 Moto2/MotoGP Moto3/MotoGP
2023 17.9 17.2 15.4 96% 86%
2022 16.7 21.4 17.5 128% 105%
2021 15.4 20.4 16.9 132% 110%
2020 12.8 19.4 16.6 152% 130%
2019 11.6 18.3 21.3 158% 184%
2018 15.9 22 18.7 138% 118%
2017 17.4 24.1 21 139% 121%
2016 16 20 23 125% 144%
2015 11.9 19.5 22.7 164% 191%
2014 11.4 22.6 20.3 198% 178%
2013 11.4 20.2 18.5 177% 162%
2012 10.3 21.8 20.5 212% 199%
2011 8.7 24.8 18.6 285% 214%
2010 7.4 22.9 17.7 309% 239%

What jumps out at you immediately is the fact that 2023 was the first year since the data was provided where there were more crashes in MotoGP than in the other classes. In Moto2, the number of crashes per event was 96% of the number in MotoGP, despite their being 31 riders instead of 22. In Moto3, there were just 86% of the crashes, with 28 riders vs 22 in MotoGP.

In comparison, in the seven previous seasons, the difference was 139% in Moto2 and 130% in Moto3. Those correlate very closely to field sizes, with the Moto2 class having around 140% of the number of riders of MotoGP, and Moto3 about 127% of the number riders. It seems a reasonable assumption that in normal seasons, the number of crashes per event per rider is fairly equal across all three classes. But MotoGP had a much higher crash rate per event per rider in 2023 than Moto2 or Moto3.

This is reflected in the historical data provided by Dorna. Below are the crash totals for all three classes combined, and for Moto3 and Moto2.

Graph showing the total falls in grand prix racing from 2011 to 2023
Crashes for all classes

Table showing total falls in the Moto3 class from 2011 to 2023
Crashes for Moto3

Graph showing the total falls in Moto2 from 2011 to 2023
Crashes for Moto2

Now, compare those charts with the one below for MotoGP:

Graph showing the total falls in MotoGP from 2011 to 2023
Crashes for MotoGP

The number of crashes in MotoGP rises more or less in line with the number of events per season. But 2023, with the same number of events as 2022, is clearly an outlier. (The bump in total crashes around 2016 is almost certainly due to the introduction of Michelin tires and spec electronics, which took some adaptation.)

What is causing all of these extra crashes in MotoGP? An exact cause is hard to pin down, but what we can do is look at when those crashes are happening and extrapolate. Below is a table with the number of crashes for each session. (The data provided by Dorna was not tallied entirely correctly, due to the renaming of sessions which happened mid season. Friday afternoon's FP2 became Practice, and Saturday morning's Free Practice became FP2.) Given the different structure to the weekend for the three classes, the practice sessions are split by the time of day they are held, rather than by name.

  Moto3 Moto2 MotoGP Total per session
Friday morning practice 45 56 42 143
Friday afternoon practice 71 59 86 216
Saturday morning practice 43 54 35 132
Qualifying Practice 1 17 18 16 51
Qualifying Practice 2 30 34 32 96
Sprint     49 49
Warm Up* 1 6 12 19
Race 100 117 86 303
Total per class 307 344 358 1009

* Moto2 and Moto3 only had warm up in Australia due to weather conditions

The data is pretty clear. The Friday afternoon sessions, which determine who goes through to Q2 and who doesn't, are absolutely crucial. And MotoGP riders are crashing at twice the rate on Friday afternoon than they are in the Friday and Saturday morning sessions, which do not count toward Q2 (at least since Silverstone, when the schedule changed).

The importance of qualifying is also clear from the stats across all three classes. Crash rates in Q1 are roughly half that of Q2, as the number of riders capable of getting through from Q1 to Q2 is limited. But in Q2, riders are taking a lot more risk to try to secure the best starting position possible.

Finally, there's the addition of the sprint race. That race alone saw 49 crashes in 2023, a rate of 2.6 crashes per race. But given that the sprint races are half the distance of the full race, that is higher than the crash rate in the Sunday races (86 crashes / 20 events / 50% = is equivalent to 2.15 crashes over the same distance). This is logical, as the start of the race and the opening laps are the most dangerous part of a race, when the bikes are closest together and riders are fighting hardest for positions.

Are the sprint races distorting the figures, though? Take the crash stats for sprint races out, and the data suggests they are not. If we remove the crashes in sprint races, as well as the crashes in warm up (Moto2 and Moto3 didn't have a warm up session in 2023 except at Phillip Island), then you get a picture which is much easier to compare and draw conclusions from.

Without the warm up and sprint races, the crash total in MotoGP drops from 358 to 297, with minor adjustments to Moto2 and Moto3. But weight the crash totals by rider numbers and it is clear that the crash risk in MotoGP is still much higher than in other classes.

  Moto3 Moto2 MotoGP
Friday morning practice 45 56 42
Friday afternoon practice 71 59 86
Saturday morning practice 43 54 35
Qualifying Practice 1 17 18 16
Qualifying Practice 2 30 34 32
Race 100 117 86
Total per class 306 338 297
# Riders 28 31 22
Crashes per rider 10.9 10.9 13.5

What the above table shows is that riders in the MotoGP class have a nearly 25% greater chance of crashing than Moto2 and Moto3. As we saw above, that has changed compared to previous years, where MotoGP had a lower crash rate per rider than the support classes.

Though the above analysis is far from being statistically rigorous, and does not account for all confounding factors, I think it is fair to say that the addition of sprint races has made MotoGP more dangerous. And the danger is not necessarily coming from the sprint races themselves, but from the increased pressure on qualifying, especially in the Friday afternoon session of timed practice that decides passage to Q2.

There is now a lot more pressure on the riders and teams from the very first day of the weekend, with less time for slip ups and mistakes. The riders have to take more risks to get through to Q2, more risks in Q2 to qualify, and have twice the number of starts and opening laps, which are the most dangerous part of the race.

Add all this in to the fact that the bikes are closer in performance than ever, and the increased importance of aerodynamics and ride-height devices have made overtaking much more difficult than in the past, and you have a recipe for the need to take more risk. More risk means more crashes, and more crashes mean more danger. And that is reflected on the health of the grid. The fact that we didn't see a single Sunday race start with a full grid of official riders bears that out. That has to be a bad thing for MotoGP in the long run.

The sprint races have demonstrably improved attendance at MotoGP events. But they have also demonstrably made the weekends more dangerous for riders, and increased the risk of injury. What we should be asking is whether that is a worthwhile trade off.

Alex Marquez, Fabio Di Giannantonio, and Johann Zarco in the gravel at Turn 1 after the start of the Barcelona MotoGP race

If you enjoyed this article, please consider supporting MotoMatters.com. You can help by either taking out a subscription, supporting us on Patreon, by making a donation, or contributing via our GoFundMe page. You can find out more about subscribing to MotoMatters.com here.

2023
MotoGP
Valencia, Spain
CormacGP
Dorna
↑Back to top

Comments

Lots of work there. Your conclusion certainly supports the general feeling about the sprint races that many have expressed without that kind of analysis. 

 

Nice to see your IT background showing a bit ... :-)

As a data science guy I usually cringe a bit when seeing a statistical analysis in the general media. Sometimes there's an obvious bias in what data is chosen and how it's presented, more often it's clear that the person writing the article couldn't tell you the difference between a mean and a median if their life depended on it. :)

But this is very nicely done. Intelligent analysis and a clear presentation.

As a partial mitigation, might I suggest the Q1/Q2 circus be done away with? The reason for it initially was to make Saturdays more compelling, which it did. But now with the sprint races on Saturday that seems like less of a need, and we could go back to the old way. Q1 and Q2 together take up about 40 minutes, just make that P4, with both P2 and P4 times counting towards qualifying. 

I agree, it was good to read a statistical analysis of the past season as opposed to poorly written, ill informed op-ed clickbait based entirely on a nebulous emotional stance: “aERo IS rUIniNG thE RaCINg!” Cool, that’s interesting now show me the research, evidence and analysis to support that view and no, an interview with Aleix, lousy with hyperbole and overstatement, doesn’t count as evidence.

I would have liked to see each season normalised for grid size instead of counting totals.
Very quickly (might be incorrect) I got 0.814 crashes per event per rider in ‘23, 0.698 in ‘22 and 0.438 in’10. Ideally I would like per Km and perhaps with a chip standing or points per event weighting to simulate pressure on the rider’s performance. For a sport where technical expertise is fundamental to its very existence it’s disappointing that the level of analysis available to the fans is so primitive.

Might there be a further discussion on how the level of competition has driven the increase too? More riders are closer than ever to achieving the thing they’ve dedicated their lives to and so the incentive to take greater risks is more prevalent. This appears to be born out by the fp2 and q1 stats. 
I think of Dovi “cruising” round on the Ducati waiting for the bike to improve for 2 or 3 seasons and the late 800 era when the race was a strung out time trial for 5th place with a small grid size. There was less incentive or opportunity to roll the dice on a risky pass or chase an opponent down for glory.

Now we have the Saturday race can we please have 1 hour long FP1, FP2, FP3, and QP? The arbitrary “competition” imposed on different sections of the weekend is getting in the way, let’s simplify and purify and leave F1 to have asterisks on their caveats, regardless of how thrilling the last 10 minutes of fp2 can be.

Not so sure.

You're right in that total number of crashes doesn't take into acount the number of riders. 2010 being a good example, a year of small grids....also 900's, lots of bike gizmos, fuel management races etc. A different class of bike, a different class of race. 

David mentions that the total number of crashes in MotoGP '23 is an outlier which I guess is a tag that could be given when comparing the total against Moto3 and Moto2. How much value is gained from this comparison is not clear. If we start with the idea that sprint races equate to more crashes then 2022 was a more 'crashy' year than 2023 for MotoGP. Half the number of races but only a fraction fewer crashes. At a rough guess from the chart, 20 more (checks table)....23 more crashes in 2023 than 2022. On average only 1.2 more crashes per event. Far from an outlier when compared to the previous years. More progression of a trend. 

I think the 'bump' in the years after 2016 is, as David writes, due to the spec electronics and Michelin tyres. 2019-20 see an apparent return to the previous norm but 2020 can also be seen as the start of an upward trend. Hate to say it but 2020 was the year aero got serious. Wings no more, a whole bike approach begins. On the Ducati it can be seen in 2019. Compare the 2018/2019 Ducati aero, 2018 is funny. The form of the current Ducati begins in 2020. Looks like the same bike, just more innocent. From 2020, not all at once, every bike starts to go in that direction. From aero being a helpful add-on to bikes built around aero. We could exclude Yamaha from that and the results are known. 

I'm not sure, that aero causes more crashes but more crashes with than without. New tyres, new electronics...I can see that causing an increase in the number of crashes as teams/riders adjust. Maybe this increase is similar, teams/riders adjusting to the 'new' and given the rapid pace of aero development in recent years the bikes are always 'new'. There's also the possibility that the dreaded front tyre issues are causing this increase but in 2023 there were more crashes before the minimum pressure rule was introduced than after. Looking at the number of crashes split across session, the importance of grid position seems to be a biggie. Bad grid slot, bad race, ergo big risks taken for grid slot. I wonder why.

David your rigorous data leads to a pretty clear conclusion: no more racing, only free practice. More racing means more crashes.

Also, what would the numbers look like without Honda?

Dave, I disagree with part of you analysis, specifically "The number of crashes in MotoGP rises more or less in line with the number of events per season." Not true at all. In 2010 there were 18 races in 2022 there were 20, correct? Yet 2022 saw over 2X as many falls as 2010. 2021 had the same number of races as 2010 and 2X the number of falls. Can do this for any year and there is no direct correlation between number of races and falls, not even "more or less" kinda sorta. Sorry.

Why is MGP crashing more? The "they push to limit" argument is clearly nonsense. MGP riders have always pushed their machinery to limit unless someone wants to argue that Doohan, Roberts, Rainey, Lawson, Rossi, Stoner, Lorenzo were slackers. So if there's no difference  in rider mentality, the bikes must be the cause. What the data suggests (crashes trending up over time - 2X+ more crashes in 2022, before the sprints, than 2010) is that as the bikes have become more sophisticated - electronics, aero, etc. - they have become less rideable at the limit even for elite riders. The exact opposite of what technology is suposed to deliver. Should certainly inform MGP  considering rules for 2027 and beyond. Fans want to see good racing with riders staying in the saddle vs sitting the couch at home recovering from a fall. Echoing recent comments by Stoner, seems past time to put control back into the hand of the riders where the data indicates they were safer, than keep it in the hands of the electronics wizards threatening their lives and livelihood.  

....500's were clearly harder to ride but the difference between the factory Honda/Yamahas and satellite bikes was huge. So theres one reason why riders are having to push so hard.  Everyone now gets the same tyres, back then you had A & B tyres.   Not forgetting FP1 & 2 for years being where you had a chance to get your race set-up sorted.  Essentially what I'm saying is this is the closest grand prix has ever been in its history, in terms of bike quality front to back & tyre fairness.  Add in the importance of qualifying well, which in my opinion has taken on too much importance in a similar vein to F1.  

I have been watching this game long enough to not need to read a piece on stats to tell me riders are under more pressure than ever before.  Also Dornas policy of copying F1, 'you want a race, here have one' is ridiculous. In my opinion there are too many races, when does fatigue come into play regarding the amount of crashes?

Yes the racing is close in MGP, but closer than M3 where you can throw a blanket over the top 20 from beginning to end and guys are flying five wide into corners bouncing off each other? Compared to the chaos of every M3 race, MGP is not "close racing." Even if it were similar, the best riders in the world should be better able to manage it than a bunch of kids. Close racing "causes" more crashes in MGP, nonsense.

I would wan't to disqualify wet sessions as a variable. Appreciate that standardising a data point around 'wetness' is bit problematic but I would want to know whether a preponderance of wet set sessions, all other variables being equal, would lead to more falls... . 

The two common sense inferences that more points to race for means more falls, or that idea that qualifying is more important now than ever because it is hard to overtake, would seem like reasonable behavioural explanations for DE's conclusion.

But as St Stephen asks, might not ANY conclusion be buggered up by Honda's record falling (and injury?) total for the year. Not only impossible to ride safely but a statistical nightmare too Honda?

The absence of detailed data from each off (who has time for this?) leaves lots of variables but I suspect the energy in the premier class (vehicle mass and velocity) has something to do with it in contacts, which we seem to have a lot more of due to closer racing. Also, performance data is much more granular and accessible in the garage which means riders are pressured to seek marginal improvements. And then cross the line between grip and gone. Perhaps tyres (oh no!) are less predictable, and this is compounded by aero (the nemesis) which exerts tyre and suspension loads not previously experienced. Probably too many unknowns to know.

What does all this cost? Hodgson has touched upon even light offs costing £30k in parts. Do teams need more staff to repair the damage quickly?

is that the teams don't even bother trying to fix the bike if time is short, it gets stuck on the work stand under a cover to sort out later.  A mechanic (perhaps Alex Briggs) mentioned this in an interview not too long ago, that the bikes have become so complex that quick repairs are not possible anymore.  So many electronic parts to assess for damage and replace, and now the (fully mechanical, hence more sensitive to perfect components and assembly/setup) ride height devices too.  I presume we can't flog the dead aero horse in this respect at least, since it all changes with the fairings with much the same ease as ever.

I'll contradict myself though since I was lucky enough (read: paid a month's salary) to be in the Gresini box at Catalunya when both their bikes were involved in the T1 crash.  I stood transfixed as the team frantically went to work.  TBH, it was such a sensory overload that I don't even recall if they used the same bikes for the restart or not.  Probably they set the #2 undamaged bikes up for the restart while the crashed bikes were worked on as fast as possible to be the hot spare in case of a problem or a flag to flag.

The cost to riders is the most significant factor. The pain is part of the job, I guess, and the small offs seem not to affect them. The more significant injuries now attract specialist surgical interventions and cryogenic and other treatments, plus physio/taping etc. Again, what is the cost of all this in terms of people, time, cost, and lingering effects. My very limited personal experience of fractures and bruises was very significant on normal life, never mind trying to swing a leg over a bike again.

And while this data may put credence to the fact that the current MotoGP weekend is more dangerous for the riders with the inclusion of the sprint races, I don't really see how Dorna 'steps back' from this new format.  Sprint races at only select tracks?  Change the sprint format so that finishing position dictates Sunday's grid -- to maybe partially alleviate some stress on Q1 / Q2?  Both of those feel like very negligible band-aids.

Perhaps the new reality in MotoGP is that the riders just get that much closer to being actual gladiators...

Whoever said there are too many race weekends is right IMO. Ditto on race starts. And a big part of the issue with bikes is not enough testing.

IMO Dorna should go back to 15-16 race weekends per year, with no sprints. Then do 5-6 test sessions throughout the year, and host a sprint race at each of those. Choose venues to hit every market obviously, but also focus on the venues that produce the best spectacle and product. There should only be two races in the same country each season. More testing will allow Michelin to finally fix the front tire and obviously allow lagging factories to catch up. As well as give the whole paddock necessary time to implement technical changes (i.e. scrapping aero/ride height devices). I'd like to see the US race move out of CoTA as well (see point about venues that produce the best spectacle).

As is Dorna is operating penny wise pound foolish, literally diluting the value of their product at the expense of the teams. It's stupid and shortsighted IMO. They have made some great moves over the last decade but they are making some bad ones now.

I haven't really followed any stick and ball sports for a while, but I know that the people in charge there also like to have a bit of monkey with the rules from time to time to try to increase drama or interest or whatever. But in those cases, they probably don't need to worry about the impact on athlete safety as much.

Or in the case of American football, the owners actively seek to hide athlete safety issues.

But this data set is a good illustration of how, for motorcycle racing, changes made on paper to the schedule, are putting riders in the hospital. MotoGP is not tiddlywinks. While I do sometimes think the Ezpeletas are morally bankrupt vultures, I believe they're not the worst offenders in professional sport.