Jump to content


Photo

Got 15 Seconds?


  • Please log in to reply
11 replies to this topic

#1 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 18 September 2020 - 10:16

Anyone for a bit of F1 statistics anorakism?

 

Recently, I entered the individual lap times of the 1995 German GP into my specially prepared spread sheet, and was then amazed to see that all the finishing times were exactly 15 seconds off from all the period sources I could find! :confused: That is, for example, Schumacher's winning time is given everywhere as 1:22'56.043", while the sum of his individual lap times is 1:23'11.043"!! :confused: :confused:

 

I took a very hard look at my data source, but it looked all very reasonable, and so I searched out a YT video of the race, to check the figures given during the boradcast. Sadly, by 1995 the onscreen data no longer showed overall race times, but the gaps between drivers at various laps during the race were all consistent with my data, so how did this happen? At the end of the race, the final results were given, and sure enough, they're the same as the ones in Autocourse, Autosport, Motorsport aktuell etc., i.e. 1:22'56.043" for Schumacher and so on. Wtf is going on here??? :confused: :confused: :confused:

 

Not easily defeated, me, so I continued looking for clues, and my calculator worked overtime - I was absolutely sure that my lap times and final race times made sense, much more sense at least as those on screen and in the magazines, but how can I prove it? There! The broadcast ended with an official list of results, listing all the classified finishers and their (fishy?) race times, but also all the retirements and their average speed up to the point of their retirement, and there it was at the very end of the list: Damon Hill with an absolutely impossible average speed of over 248 kph for his one single lap from a standing start! Out with the calculator again, and his time for that lap would have to have been (and here comes!) exactly fifteen seconds faster than the one I had in my data set, and the one he was credited with as his fastest race lap! Ha! Double Ha! Ha!

 

So, what happened? I can only guess, but my bet is that, for one reason or another, they simply switched the clock on roughly fifteen seconds late, and thus the race went on with fishy times for the first lap, but by the end of the race (45 laps) that error was diluted and nobody noticed. Nobody, that is, until they published the list of individual lap times, and then they just added those 15 seconds to the times for the first lap. But, by then, of course, all the magazines and newspapers were already going ahead and printing the "fishy" overall race times.

 

Oh, what fun! :)


Edited by Michael Ferner, 18 September 2020 - 10:22.


Advertisement

#2 Tim Murray

Tim Murray
  • Moderator

  • 24,592 posts
  • Joined: May 02

Posted 18 September 2020 - 11:57

Congratulations Michael on a great piece of detective work. So even automatic timing systems can apparently suffer from human error. I wonder how many other similar cock-ups have happened over the years. :well:

#3 Imperial

Imperial
  • Member

  • 4,820 posts
  • Joined: February 02

Posted 18 September 2020 - 13:31

Further proof that Benetton were cheating!  :p



#4 Izzyeviel

Izzyeviel
  • Member

  • 3,172 posts
  • Joined: March 16

Posted 18 September 2020 - 13:33

great find.

 

Maybe the timing beam wasn't on the start line or maybe someone stood in front of it at the start? 



#5 ReWind

ReWind
  • Member

  • 3,394 posts
  • Joined: October 03

Posted 18 September 2020 - 17:29

Wie kann ein Rennen, das um 14 h 03 min 40 sec gestartet wird und bei dem der Sieger um 15 h 26 min 50 sec über die Ziellinie fährt, offiziell bereits nach 1 h 22 min 56 sec beendet sein? Zu diesem Zeitpunkt befand sich Michael Schumacher noch nicht einmal in der Opelkurve.
Gemäß der Fernsehübertragung dauerte das Rennen eindeutig 1 h 23 min 10 sec (vielleicht auch 9 oder 11 sec). Gibt es eine Erklärung dafür, dass das Rennen offiziell ca. 14 sec kürzer war?

 
That is the reader’s letter I sent to „MOTORSPORT aktuell“ on 5 August 1995. It was not published and I got no answer.
On my self-designed data sheet I had noted the time the race started and the time it ended. So I quickly became aware of the difference between the elapsed time and the official duration. I suspected then that the TV coverage on RTL was belated after a commercial break. But 25 years later I learn that in fact it was a time-keeping problem.



#6 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 19 September 2020 - 15:46

Congratulations Michael on a great piece of detective work. So even automatic timing systems can apparently suffer from human error. I wonder how many other similar cock-ups have happened over the years. :well:

 

 

Electronic timekeeping is a huge improvement over manual timing, but it is not perfect - at the very least, it wasn't in the beginning! For example, during the very first race of the new Olivetti/Longines partnership in 1982 (although there were a few test runs in '81 already), it is noticeable that there were no gaps between cars of less than one tenth of a second, other than about a handful of exactly 0.001", and over fifty occasions when both cars were given the exact same time to the one-hundredth of a second - normally, you would expect that to happen maybe once or twice during one race! It is obvious that the timekeeping/processing system was not capable of distinguishing between two signals that were no more than 0.100" apart. That means, of course, that a car following another one over the line within one tenth of a second was given the same time stamp as the one ahead, in other words it was credited with a lap time that was up to one tenth of a second better than the one it actually achieved - which does make you wonder about the accuracy of the qualifying times, considering that there were, for example, only 0.004" between 13th and 15th on the grid!

 

In the same race, Chico Serra was credited by the electronic timing device with a fastest race lap of 1'12.970" on lap 4, during which he suddenly jumped two places in the standings, the same two places which he lost again on lap 5 - no manually recorded lap chart of the race that I've seen has recorded those changes of position, and Serra's next best lap was nearly three seconds slower - in fact, his qualifying time was almost half a second slower! I'm fairly certain that's a timing error, too. Or, how about Derek Warwick during the 1983 European Grand Prix at Brands Hatch, who was recorded by Olivetti/Longines to have lost six positions on lap 3, and to have regained them all again the following lap, overtaking the likes of Keke Rosberg, Mauro Baldi, Bruno Giacomelli, John Watson , Niki Lauda and Andrea de Cesaris, all within 75 seconds - I'm sure we would have read about that feat in the comics if it really had happened!

 

The timing of all the official practice sessions and of the races was carried out in the best conditions, with no errors and no "technical hitches" to be regretted. As far as the sport is concerned, this represents a considerable attainment, because the electronics associated with the tradition of timing have meant that a 100 % reliability rate has been reached and that all risks of errors and therefore of sporting injustice have been effectively removed.

 

(excerpt taken from the foreword to the Formula One 1983 Championship Information Service, jointly signed by The President of FISA, Jean-Marie Balestre, and The President of FOCA, Bernard Ecclestone)



#7 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 19 September 2020 - 21:54

Another example of an electronic timekeeping error - French GP of 1984:

 

 

Wind to 35'20" to see Patrick Tambay leading Niki Lauda past a couple of backmarkers (Mauro Baldi and Piercarlo Ghinzani) at the end of lap 39. The Olivetti/Longines timing gives the following time stamps:

 

 :45'03.521" (Ghinzani)

 :45'03.651" (Lauda)

 :45'03.656" (Tambay)

 :45'04.101" (Baldi)

 

But you can clearly see that, although Lauda tries very hard, he can't even draw level with Tambay, yet the timing has him 0,005" ahead! On the official Olivetti/Longines lap chart, the error was corrected manually, showing Tambay still ahead, but the individual lap times add up to the time stamps as shown.



#8 dgs

dgs
  • Member

  • 249 posts
  • Joined: November 12

Posted 21 September 2020 - 13:13

Back in 1995 Getting hold of the FIA Race Classification sheets on Grand Prix's was not anywhere as easy as it is today

I was lucky enough to have a contact at Goodyear Tyres who would send me copies of practice/race results.on each GP. Some races more sheets than others.

 

I received the 1995 German Grand Prix Race Classification after 45 laps. This was stamped 'Provisional' as were some others I obtained over the years. It showed all drivers positions or retirements, no of laps, race time and drivers fastest lap and on what lap it happened on.

 

Result:  1st, M.Schumacher 45 laps, 1:22'56.043". However when I obtained 'Autospor't' magazine I noticed a number of drivers fastest lap were incorrect (ie shown a different time on different lap).. But the driver that stood out was Damon Hill who only completed one lap. It read D.Hill, 1 lap, 1'48.989", fastest lap, lap 1,  1'49.989". Obviously one of these times must be incorrect.

 

I wrote to 'Formula One Promotions & Administration' who kindly supplied me with the official FIA time sheet, which had exactly the same race times for all drivers, but correct fastest lap and on what lap it happened on for the 8 drivers (except D.Hill). 

 

The exception was Damon Hill's race/lap time which showed both were incorrect on provisional sheet. Correct time was 1'53.989".

 

At that stage (remember 25 years ago) I thought I had correct times until I saw Michael comments on missing 15"

 

I do not have FIA individual race laps, but a friend sent me a printout of individual race laps (not FIA but 'mirror' times) These all add up to same total times as FIA sheet.

 

Since reading Michael's topic I have added up a number manually and they all add up to exactly 15" longer, with exception of Damon Hill who only completed 1 lap.

 

Another source is 'Forix' results which as well as having Official Race times, has a  'Classification @45 lap sheet, which adds up the laps times that shows all drivers (except D.Hill) times with the 15" added, even Pedro Diniz who only completed 8 laps, but has FIA time 17'49.491", Forix Classification @45 laps sheet 18'04.491".

 

 

I am completely at a loss to explain this difference

 

 



#9 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 29 September 2020 - 21:42

The prize for the biggest balls-up in the history of electronic timing must go to the 1984 South African GP, I guess. Some of you may be aware of the Boutsen snafu, as explained in Autocourse and Mike Lang's Grand Prix: he actually finished 9th with 71 laps completed, but the provisional results handed out after the race placed him 13th, with only 70 laps, which is what all the magazines and newspapers published. However, this error was corrected when the Formula One - 1984 Championship Information Service was published, and I have 71 individual lap times for him, although I'm not really sure if they are in the right order – as it is, the difference between his overall time in the provisional results and the final, official ones is given as the time for his 71st lap, alas there’s evidence that the lap that was “lost” by the timekeepers was an earlier one, maybe his 51st. But in any case, this is only small fry when compared to some of the other timekeeping errors!

When entering the individual lap times in my specially prepared spread sheet, I already noticed some “weird data” – e.g. normally, a pit stop can be easily identified by a couple of lap times that are slower than usual, i.e. a long in-lap/short out-lap or short in-lap/long out-lap couple, depending on the location of the pits relative to the timing line. A spin or other driving error can cause similar lap times, but they rarely happen on consecutive laps, in addition to which I can always consult race reports to make sure I get it right. At Kyalami, now, the pits are located behind the timing line, so it’s short-in and long-out, but on a few occasions this pattern was not clearly visible, or not happening at the reported moments, so I decided to cross check my data with the TV feed. Luckily, I found a YouTube video of the Brazilian Globo (I think) broadcast with a continuous feed, only interrupted by three or four commercials of 30 or 45 seconds, which I could clearly identify even though all but one of them were cut from the video. Thus, I was able to check a good number of time stamps from my spread sheet against the running clock, and whilst not a very precise procedure, it still enabled me to see that the vast majority of the times checked that way were at the very least “reasonable”, i.e. within a margin of error of less than one second or so.

Some, however, were not, and concerned the time stamps (and thus, the individual lap times) of only three drivers: Patrick Tambay, Stefan Bellof and, to a lesser degree, Alain Prost, but since the fastest lap of the race was credited to Tambay, and Prost finished second besides recording the second fastest race lap time, those errors are anything but trifle! Beginning with the relative minor issue of the running order on the first lap, the video and a couple of photographs published in Motorsport aktuell at the time, show clearly that Bellof was running in 17th position through the kink on the main straight and arriving in 18th at Crowthorne, having been passed by Hesnault, while Prost was in 25th position coming out of Leeukop, and two places further up at Crowthorne, having passed both Arrows, yet the Olivetti/Longines timing data shows them in 23rd (Bellof) and 26th (Prost) positions, respectively. As a result, Bellof is credited with a Chico-Serra-type of personal best time on lap 2, almost two seconds faster than any other he recorded that afternoon, and less than half a second slower than his qualifying time. Granted, not an earth-shaking error, but stay tuned – it gets worse, much worse in fact!

A mistake like that is likely the product of a “hiccup” in the timing and/or data processing, and apparently not that rare – cf. the Chico Serra and Derek Warwick incidents mentioned in my earlier post. Furthermore, it is quite likely that a number of such “hiccups” will never be detected, e.g. if a car is not running in the close proximity of other cars, and outside of view of the TV cameras. It’s impact during a race will normally be limited to the accuracy of the lap chart generated from the timing data, and perhaps the fastest race lap of a competitor, however in practice and qualifying such an error could conceivably be much more “vicious”, in that it could result in shifts of the starting order, or even decide whether a competitor qualifies to start a race or not! Incidentally, I believe that an error like the one in the Tambay/Lauda case from my previous post may be the result of one car running closer to the receiver station than the other – if you have watched the video, you will have noticed that Lauda was forced onto the grass strip between the pit wall and the track, which may have triggered his transponder signal a fraction of a second earlier – at that speed, one thousandth of a second is equivalent to three inches on the track, so it doesn’t take much to create an error in that range!



To get back to the South African GP, here’s a YouTube video from the CBC broadcast with BBC commentary, which is of better quality than the Globo broadcast, but incomplete. If you wind forward to about 20’15”, you can watch the group fighting over second place close in on the lapped Tyrrell of Bellof – up to this point of the race (lap 31), I can’t find any further obvious errors, so here’s the time stamps from the Olivetti/Longines data of the previous lap (position, #, driver, car, lap, overall time):

16 4 Bellof Tyrrell/Ford 29 :36'08.768"
2 6 Rosberg Williams/Honda 30 :36'11.825"
3 7 Prost McLaren/TAG 30 :36'12.303"
4 16 Warwick Renault 30 :36'12.884"
5 11 Angelis Lotus/Renault 30 :36'13.526"
6 15 Tambay Renault 30 :36'14.500"

At the end of this lap, Prost and Tambay stop at the pits, still about two or three seconds apart, and accelerate back into the race just before Lauda appears to start his 33rd lap (as an aside, just watch Tambay nearly mowing down the McLaren pit crew :eek: :eek:), while Rosberg, Warwick and de Angelis continue around, gobbling up Bellof in the process, who’s still following them when they complete their 32nd. Now, take a look what the electronic timing tells us:

15 4 Bellof Tyrrell/Ford 30 :37'23.080"
2 6 Rosberg Williams/Honda 31 :37'24.284"
3 15 Tambay Renault 31:37'25.280"
4 16 Warwick Renault 31 :37'25.281"
5 11 Angelis Lotus/Renault 31 :37'25.835"
6 7 Prost McLaren/TAG 31 :37'29.595"

This is the lap when both Prost and Tambay enter the pits – Prost’s time is reasonable, but Tambay???!!!???

2 15 Tambay Renault 32 :38'35.262"
3 6 Rosberg Williams/Honda 32 :38'37.314"
4 16 Warwick Renault 32 :38'38.096"
5 11 Angelis Lotus/Renault 32 :38'38.756"
6 7 Prost McLaren/TAG 32 :38'42.495"
17 4 Bellof Tyrrell/Ford 31 :39'06.686"

Next time around, Tambay’s supposed to have taken over second place, whilst stationary at the pits!! And after his regular “slow in-lap”, Prost appears to be lapping again as if he had never stopped, while Bellof has mysteriously lost half a minute, although we can still see him circulating with the Rosberg/Warwick/ de Angelis group! :drunk:

2 6 Rosberg Williams/Honda 34 :41'02.214"
3 16 Warwick Renault 34 :41'02.737"
4 11 Angelis Lotus/Renault 34 :41'03.242"
19 4 Bellof Tyrrell/Ford 32 :41'07.759"
9 7 Prost McLaren/TAG 34 :41'23.000"
10 15 Tambay Renault 34 :41'26.930"

It takes another couple of laps before the relative positions of the cars agree again with what’s shown on TV, but now Bellof is two laps down instead of only one - and it gets worse: five laps later, and for no obvious reason, he’s credited with a lap time of 2’32.945”, pretty much exactly the time it usually takes for him to complete two laps, so now the Olivetti/Longines timing has him three laps down in last place but one, instead of running one lap down, and in 14th!




Having screwed up Bellof’s race by docking him two laps, the timekeepers then, perhaps by way of compensation :rolleyes:, go on and ignore his pit stop to take on ballast… sorry, water replenishment on lap 57 (actually, lap 59!). Here, we have to refer to the Globo broadcast, although the stop itself is not captured, but we can see Bellof following Lauda into Crowthorne just after the commercial break at 1:23’50”, whereas the Olivetti/Longines timing has him running more than twenty seconds up on the road at this time (Lauda’s 63rd lap). The eagle-eyed will also spot him passing in the background during the coverage of the Laffite retirement, again some 25 seconds after he is supposed to appear according to the electronic timing!

And finally, at about 1:26’10”, we can see Alboreto, Warwick and de Angelis lap him again just before he spins out on what is supposedly his 60th lap, but is in fact his 62nd! Here are the relevant time stamps by Olivetti/Longines:

16 4 Bellof Tyrrell/Ford 59 1:16'02.258"
6 22 Patrese Alfa Romeo 62 1:16'25.723"
4 27 Alboreto Ferrari 63 1:16'27.496"
5 16 Warwick Renault 63 1:16'27.572"
15 11 Angelis Lotus/Renault 60 1:16'28.011"
8 19 Senna Toleman/Hart 61 1:16'29.988"

From the TV feed, we can see that Bellof’s actual time for 61 laps (not 59) is roughly 1:16’27.8”, not 1:16’02.258”, and that he was still running 8th, just ahead of Ayrton Senna, at the time of his accident.

Some cock-up, innit?

Somehow not surprisingly, there are none of the usual Olivetti/Longines graphics detailing the current race standings during the broadcast, only some “homemade” graphics giving the top six placings, but no total time or gaps. There are, however, a very few occasions when the then current fastest lap was shown, and those are interesting, again: shortly before the pit stops of Prost and Tambay, the latter was credited with fastest lap of 1’09.458” on his 24th, however the individual lap time report has him doing that lap in 1’12.052”! In fact, according to the Olivetti/Longines data, Tambay’s second lap in 1’11.341” remained his fastest until doing 1’10.780” and 1’09.982” on laps 31 and 32, respectively – the in- and out-laps of his pit stop!!! A time of 1’09.458” is not recorded for Tambay during the entire race, nor for any other driver!

And finally, the fastest lap of the race overall (1’18.887”) is credited to Tambay on his 65th lap on the screen, yet it is lap 64 in the electronic timing report! Autocourse also has it on lap 65, and shows a few other differences to the official figures: Prost is credited with a 1’08.961” on lap 33, but only 1’09.398” (on lap 41) in the lap-by-lap data (where his 33rd is given as 1’29.807”, his apparent out-lap – which we know from the TV feed to have happened a lap earlier!), while Lauda and Warwick (third and fourth fastest laps) are given the same time in both lists but, like Tambay, one lap later in Autocourse than in the Olivetti/Longines data, and finally Mauro Baldi (of all people!) with the same time, but on lap 69 instead of 70 – enough to make your head spin!

Needless to say, I am at a complete loss as to even only understand how these errors could have originated, but if I have shaken anybody’s belief in the accuracy of electronic timekeeping then it was probably worth it! :)

Edited by Michael Ferner, 29 September 2020 - 22:01.


#10 dgs

dgs
  • Member

  • 249 posts
  • Joined: November 12

Posted 30 September 2020 - 08:31

The prize for the biggest balls-up in the history of electronic timing must go to the 1984 South African GP, I guess. Some of you may be aware of the Boutsen snafu, as explained in Autocourse and Mike Lang's Grand Prix: he actually finished 9th with 71 laps completed, but the provisional results handed out after the race placed him 13th, with only 70 laps, which is what all the magazines and newspapers published. However, this error was corrected when the Formula One - 1984 Championship Information Service was published, and I have 71 individual lap times for him, although I'm not really sure if they are in the right order – as it is, the difference between his overall time in the provisional results and the final, official ones is given as the time for his 71st lap, alas there’s evidence that the lap that was “lost” by the timekeepers was an earlier one, maybe his 51st. But in any case, this is only small fry when compared to some of the other timekeeping errors!

When entering the individual lap times in my specially prepared spread sheet, I already noticed some “weird data” – e.g. normally, a pit stop can be easily identified by a couple of lap times that are slower than usual, i.e. a long in-lap/short out-lap or short in-lap/long out-lap couple, depending on the location of the pits relative to the timing line. A spin or other driving error can cause similar lap times, but they rarely happen on consecutive laps, in addition to which I can always consult race reports to make sure I get it right. At Kyalami, now, the pits are located behind the timing line, so it’s short-in and long-out, but on a few occasions this pattern was not clearly visible, or not happening at the reported moments, so I decided to cross check my data with the TV feed. Luckily, I found a YouTube video of the Brazilian Globo (I think) broadcast with a continuous feed, only interrupted by three or four commercials of 30 or 45 seconds, which I could clearly identify even though all but one of them were cut from the video. Thus, I was able to check a good number of time stamps from my spread sheet against the running clock, and whilst not a very precise procedure, it still enabled me to see that the vast majority of the times checked that way were at the very least “reasonable”, i.e. within a margin of error of less than one second or so.

Some, however, were not, and concerned the time stamps (and thus, the individual lap times) of only three drivers: Patrick Tambay, Stefan Bellof and, to a lesser degree, Alain Prost, but since the fastest lap of the race was credited to Tambay, and Prost finished second besides recording the second fastest race lap time, those errors are anything but trifle! Beginning with the relative minor issue of the running order on the first lap, the video and a couple of photographs published in Motorsport aktuell at the time, show clearly that Bellof was running in 17th position through the kink on the main straight and arriving in 18th at Crowthorne, having been passed by Hesnault, while Prost was in 25th position coming out of Leeukop, and two places further up at Crowthorne, having passed both Arrows, yet the Olivetti/Longines timing data shows them in 23rd (Bellof) and 26th (Prost) positions, respectively. As a result, Bellof is credited with a Chico-Serra-type of personal best time on lap 2, almost two seconds faster than any other he recorded that afternoon, and less than half a second slower than his qualifying time. Granted, not an earth-shaking error, but stay tuned – it gets worse, much worse in fact!

A mistake like that is likely the product of a “hiccup” in the timing and/or data processing, and apparently not that rare – cf. the Chico Serra and Derek Warwick incidents mentioned in my earlier post. Furthermore, it is quite likely that a number of such “hiccups” will never be detected, e.g. if a car is not running in the close proximity of other cars, and outside of view of the TV cameras. It’s impact during a race will normally be limited to the accuracy of the lap chart generated from the timing data, and perhaps the fastest race lap of a competitor, however in practice and qualifying such an error could conceivably be much more “vicious”, in that it could result in shifts of the starting order, or even decide whether a competitor qualifies to start a race or not! Incidentally, I believe that an error like the one in the Tambay/Lauda case from my previous post may be the result of one car running closer to the receiver station than the other – if you have watched the video, you will have noticed that Lauda was forced onto the grass strip between the pit wall and the track, which may have triggered his transponder signal a fraction of a second earlier – at that speed, one thousandth of a second is equivalent to three inches on the track, so it doesn’t take much to create an error in that range!



To get back to the South African GP, here’s a YouTube video from the CBC broadcast with BBC commentary, which is of better quality than the Globo broadcast, but incomplete. If you wind forward to about 20’15”, you can watch the group fighting over second place close in on the lapped Tyrrell of Bellof – up to this point of the race (lap 31), I can’t find any further obvious errors, so here’s the time stamps from the Olivetti/Longines data of the previous lap (position, #, driver, car, lap, overall time):

16 4 Bellof Tyrrell/Ford 29 :36'08.768"
2 6 Rosberg Williams/Honda 30 :36'11.825"
3 7 Prost McLaren/TAG 30 :36'12.303"
4 16 Warwick Renault 30 :36'12.884"
5 11 Angelis Lotus/Renault 30 :36'13.526"
6 15 Tambay Renault 30 :36'14.500"

At the end of this lap, Prost and Tambay stop at the pits, still about two or three seconds apart, and accelerate back into the race just before Lauda appears to start his 33rd lap (as an aside, just watch Tambay nearly mowing down the McLaren pit crew :eek: :eek:), while Rosberg, Warwick and de Angelis continue around, gobbling up Bellof in the process, who’s still following them when they complete their 32nd. Now, take a look what the electronic timing tells us:

15 4 Bellof Tyrrell/Ford 30 :37'23.080"
2 6 Rosberg Williams/Honda 31 :37'24.284"
3 15 Tambay Renault 31:37'25.280"
4 16 Warwick Renault 31 :37'25.281"
5 11 Angelis Lotus/Renault 31 :37'25.835"
6 7 Prost McLaren/TAG 31 :37'29.595"

This is the lap when both Prost and Tambay enter the pits – Prost’s time is reasonable, but Tambay???!!!???

2 15 Tambay Renault 32 :38'35.262"
3 6 Rosberg Williams/Honda 32 :38'37.314"
4 16 Warwick Renault 32 :38'38.096"
5 11 Angelis Lotus/Renault 32 :38'38.756"
6 7 Prost McLaren/TAG 32 :38'42.495"
17 4 Bellof Tyrrell/Ford 31 :39'06.686"

Next time around, Tambay’s supposed to have taken over second place, whilst stationary at the pits!! And after his regular “slow in-lap”, Prost appears to be lapping again as if he had never stopped, while Bellof has mysteriously lost half a minute, although we can still see him circulating with the Rosberg/Warwick/ de Angelis group! :drunk:

2 6 Rosberg Williams/Honda 34 :41'02.214"
3 16 Warwick Renault 34 :41'02.737"
4 11 Angelis Lotus/Renault 34 :41'03.242"
19 4 Bellof Tyrrell/Ford 32 :41'07.759"
9 7 Prost McLaren/TAG 34 :41'23.000"
10 15 Tambay Renault 34 :41'26.930"

It takes another couple of laps before the relative positions of the cars agree again with what’s shown on TV, but now Bellof is two laps down instead of only one - and it gets worse: five laps later, and for no obvious reason, he’s credited with a lap time of 2’32.945”, pretty much exactly the time it usually takes for him to complete two laps, so now the Olivetti/Longines timing has him three laps down in last place but one, instead of running one lap down, and in 14th!




Having screwed up Bellof’s race by docking him two laps, the timekeepers then, perhaps by way of compensation :rolleyes:, go on and ignore his pit stop to take on ballast… sorry, water replenishment on lap 57 (actually, lap 59!). Here, we have to refer to the Globo broadcast, although the stop itself is not captured, but we can see Bellof following Lauda into Crowthorne just after the commercial break at 1:23’50”, whereas the Olivetti/Longines timing has him running more than twenty seconds up on the road at this time (Lauda’s 63rd lap). The eagle-eyed will also spot him passing in the background during the coverage of the Laffite retirement, again some 25 seconds after he is supposed to appear according to the electronic timing!

And finally, at about 1:26’10”, we can see Alboreto, Warwick and de Angelis lap him again just before he spins out on what is supposedly his 60th lap, but is in fact his 62nd! Here are the relevant time stamps by Olivetti/Longines:

16 4 Bellof Tyrrell/Ford 59 1:16'02.258"
6 22 Patrese Alfa Romeo 62 1:16'25.723"
4 27 Alboreto Ferrari 63 1:16'27.496"
5 16 Warwick Renault 63 1:16'27.572"
15 11 Angelis Lotus/Renault 60 1:16'28.011"
8 19 Senna Toleman/Hart 61 1:16'29.988"

From the TV feed, we can see that Bellof’s actual time for 61 laps (not 59) is roughly 1:16’27.8”, not 1:16’02.258”, and that he was still running 8th, just ahead of Ayrton Senna, at the time of his accident.

Some cock-up, innit?

Somehow not surprisingly, there are none of the usual Olivetti/Longines graphics detailing the current race standings during the broadcast, only some “homemade” graphics giving the top six placings, but no total time or gaps. There are, however, a very few occasions when the then current fastest lap was shown, and those are interesting, again: shortly before the pit stops of Prost and Tambay, the latter was credited with fastest lap of 1’09.458” on his 24th, however the individual lap time report has him doing that lap in 1’12.052”! In fact, according to the Olivetti/Longines data, Tambay’s second lap in 1’11.341” remained his fastest until doing 1’10.780” and 1’09.982” on laps 31 and 32, respectively – the in- and out-laps of his pit stop!!! A time of 1’09.458” is not recorded for Tambay during the entire race, nor for any other driver!

And finally, the fastest lap of the race overall (1’18.887”) is credited to Tambay on his 65th lap on the screen, yet it is lap 64 in the electronic timing report! Autocourse also has it on lap 65, and shows a few other differences to the official figures: Prost is credited with a 1’08.961” on lap 33, but only 1’09.398” (on lap 41) in the lap-by-lap data (where his 33rd is given as 1’29.807”, his apparent out-lap – which we know from the TV feed to have happened a lap earlier!), while Lauda and Warwick (third and fourth fastest laps) are given the same time in both lists but, like Tambay, one lap later in Autocourse than in the Olivetti/Longines data, and finally Mauro Baldi (of all people!) with the same time, but on lap 69 instead of 70 – enough to make your head spin!

Needless to say, I am at a complete loss as to even only understand how these errors could have originated, but if I have shaken anybody’s belief in the accuracy of electronic timekeeping then it was probably worth it! :)

Michael has done a fine job, in trying to understand the differences that appear in the early days of electronic timekeeping by Olivetti-Longines.

I also have the Olivetti-Longines timing book published after the end of the 1984 season. What I also have for some Grand Prix's (first half of year) is a set of timings from Olivetti-Longines sent to me by post (remember that!) after each race. These are even more comprehensive than yearbook in two ways. One, these show untimed sessions (the practice sessions not counting for grid positions) two they also show all drivers times, the Tyrrell drivers at South African race, Martin Brundle and Stefan Bellof times are not shown in yearbook.

Michael wonders if times were amended for  Thierry Boutsen, when yearbook was published and actual individual lap times issued at time of race. I have checked both timing set sent after race and yearbook and both are exactly the same on each lap.

I cannot see a race lap time for Bellof of 2'32.945" on the sheet sent after race (Bellof not shown in yearbook), but agree with Michael both race tines sent after race only credit Bellof as completing 59 laps in time of 1:13'31.065". Autocourse Annual, Motor Sport and Autocar all show him as completing 60 laps. 



#11 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 30 September 2020 - 19:02

Well, that's interesting! 1:13'31.065" is the time I have for Bellof completing 57 laps!! That also explains why Mike Lang wrote that Bellof's accident happened "at almost precisely the same moment" as the Laffite retirement, whereas the TV feed shows that more than three minutes passed between both incidents - very obviously, Lang took his information from the published race times, and calculated that with Laffite triggering the timing for the last time at 1:12'29.196", and being almost a minute into his last lap at the time of his retirement (between The Esses and Leeukop), Bellof would have reached Crowthorne (where he spun off) only seconds later!

 

You are right, in that Bellof's times are not shown in the yearbook, from which I have the majority of my information. I got the Tyrrell times from this YouTube "video", mimicking the official FIA timing screen throughout the race:

 

 

I don't know about the source for this "video", but I couldn't find anything wrong with this or any other video of this series so far - I'll send you a private message.



#12 Michael Ferner

Michael Ferner
  • Member

  • 7,151 posts
  • Joined: November 09

Posted 02 October 2020 - 18:29

I've constructed a gap chart with corrected times for South Africa '84. A gap chart is a lap chart which shows approximate gaps between cars as well as positions. I'd like to think it's self-explanatory, but feel free to ask!

 

gapchart-1984-ZA.jpg