question on hand timing and FAT 1952-64


Forum devoted to track & field items of an historical nature.

question on hand timing and FAT 1952-64

Postby johnclark » Tue Aug 07, 2012 5:01 am

Hi folks,

I am new here and have a question for you. The (unofficial) FAT results from the 1952-1964 Olympics have an odd trend. I'll use the 100m results as an example. In the final of 1952, the hand timed and FAT results are:

10.4 = 10.79
10.4 = 10.80
10.4 = 10.83
10.4 = 10.84
10.5 = 10.88
10.6 = 10.91

You can see that the difference between the hand and FAT times is around 0.4 seconds on average.

Compare that now to the results in 1960:

10.2 = 10.32
10.2 = 10.35
10.3 = 10.42
10.3 = 10.44
10.3 = 10.46
10.4 = 10.50

You can see that the difference is around 0.15 seconds on average, much smaller than in 1952!

My question is why does the difference between hand and FAT times decrease? The other years are consistent as well - the differences in 1956 are between the 1952 and 1960 differences and the difference in 1964 is a bit smaller again - why?

My understanding is that FAT technology did not change much over those years, and I doubt that hand timing improved, so is there something I am missing?

Personally, I find the FAT results strange - they show that between 1952 and 1960 times in the 100m final fell by 0.4 seconds, which is a huge decrease. I find the hand-timed decrease of around 0.2 seconds in the eight years more believable but perhaps I am wrong.

Thanks in advance!

John.
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby gh » Thu Aug 09, 2012 2:18 pm

The meets were held in different cities so there were different timers.

Merely proves that NorCal timers understood proper timing procedure better than their SoCal counterparts.
gh
 
Posts: 46314
Joined: Sat Oct 08, 2005 4:31 am
Location: firmly at Arya's side!

Re: question on hand timing and FAT 1952-64

Postby gh » Sat Aug 11, 2012 3:29 am

remind me never to answer questions in the middle of the night at the Olympics while operating on 4-5 hour sleep a day!

For some reason I read the original Olympic TRIALS, not Games, hence the Cali reference.

But the answer remains the same: different timers different levels of competence from Helsinki to Australia (undoubtedly small because they were influenced by British methodology, which tried to eliminate reaction time) to Rome.

Tokyo '64 is a totally different kettle of fish. As I recall, even though the times were expressed in 10ths, those reflect roundings from autotiming equipment.
gh
 
Posts: 46314
Joined: Sat Oct 08, 2005 4:31 am
Location: firmly at Arya's side!

Re: question on hand timing and FAT 1952-64

Postby dj » Sat Aug 11, 2012 4:15 am

GH is correct about Tokyo.

Another difficulty on analyzing the times may be the hand times you're using. Are you certain these are the original hand times? They appear too internally consistent, and I suspect the hand times may have been adjusted based on photo analysis of the time behind the winner, which was a common statistical tool in the days when photos were available but not used for times.

I've had difficulty finding a good source where I can confirm the unadjusted hand times.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby dj » Mon Aug 13, 2012 6:30 am

I've finally been able to confirm that the hand times you are using are the photo-adjusted hand times as they differ from the originals for 1952:

Yours for 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.6
Original 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.5

As for 1960, it's apparent the hand times were in relative agreement with the photo times, thus requiring no adjustment:

Yours for 1960: 10.2, 10.2, 10.3, 10.3, 10.3, 10.4
Original 1960: 10.2, 10.2, 10.3, 10.3, 10.3, 10.4


Ultimately, the best means of checking what you want is not looking at place times in the final, but looking at the average difference between hand and auto of all the winning times from all four rounds.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby johnclark » Wed Aug 15, 2012 6:05 am

Thanks heaps for the replies! I think I got the times out of the Wallenchinsky book on the Olympics. I understand better now how the times were done. The FAT times are very accurate for the differences between the athletes but the early FAT times were a bit mixed for the absolute time of the winner (which is then the benchmark for comparison with the others).

If the winners FAT time is a bit out then the times for all of the others in the same race will be out as well.

Thanks for the info on Tokyo - that seems to make a lot of sense of the times.
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby johnclark » Wed Aug 15, 2012 7:13 am

OK, as suggested, I've now compared the (original) hand times with the FAT times just for the winners of each race (final, semifinals, quarterfinals, heats) for 1952, 1956, 1960.

The differences were:

1952: average of 0.275 with a standard deviation of 0.065 (max=0.43, min=0.18)
1956: average of 0.205 with a standard deviation of 0.059 (max=0.29, min=0.12)
1960: average of 0.163 with a standard deviation of 0.040 (max=0.22, min=0.11)

(FAT times always higher than the hand times)

I guess a standard deviation of around 0.05 is to be expected given that the hand times are effectively rounded to the nearest 0.1.

I've also calculated the differences between FAT and hand timing for the margins from the winner to the other runners. These are:

1952: average = -0.036, standard deviation = 0.095 (range of -0.24 to +0.29)
1956: average = -0.018, standard deviation = 0.081 (range of -0.23 to +0.15)
1960: average = -0.010, standard deviation = 0.046 (range of -0.22 to +0.07)

We should expect here an average of zero, and it is roughly that (1952 is pushing it). The standard deviation should again be around 0.05 - the 1952 and 1956 results are big. The range of results is very interesting - the worst case overestimate decreases over time but the worst underestimate is roughly constant!

I can't find any evidence that the accuracy changes with place: the margin to 2nd place is measured just as accurately as the margin to 6th place.

The hand timing of 1960 seems to be more accurate than 1952 or 1956.
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby dj » Wed Aug 15, 2012 1:34 pm

johnclark wrote:I can't find any evidence that the accuracy changes with place: the margin to 2nd place is measured just as accurately as the margin to 6th place.

The hand timing of 1960 seems to be more accurate than 1952 or 1956.


This begs my original question, where did you get the place times? My guess is that if they were derived from Wallechinsky or any modern compilation, there is the strongest likelihood that these are adjusted place times, based on the photo-timing evidence showing the time behind the winner. They do not represent what the hand-timers themselves too.

Earlier research shows there is much greater variation from hand to auto among the place times for various reasons. First place "always" had three timers (or more), with the middle time used. Places were often fortunate to have two timers, with the longer time used. The best timers were often assigned to first place, the lesser timers to 2nd place and lower.

First place is the easiest place to time, last is the next easiest. Everything in between is difficult as the timer has to look through the pack, choose the person who's going to finish in the place assigned the timer, and stop the watch. Because there is fear of missing the person, there is a greater likelihood of hurrying the process of stopping the watch, thus recording a shorter time than when would would properly wait to see the finish and not anticipate it.

As a result, the middle times can show wider variation between hand and auto than the winning times.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby johnclark » Fri Aug 17, 2012 12:01 am

dj wrote:
johnclark wrote:I can't find any evidence that the accuracy changes with place: the margin to 2nd place is measured just as accurately as the margin to 6th place.

The hand timing of 1960 seems to be more accurate than 1952 or 1956.


This begs my original question, where did you get the place times? My guess is that if they were derived from Wallechinsky or any modern compilation, there is the strongest likelihood that these are adjusted place times, based on the photo-timing evidence showing the time behind the winner. They do not represent what the hand-timers themselves too.


DJ, thanks for extra info. As I said above, I did get those times from Wallechinsky. However, for the later analysis (in the third (?) post above) of margins, I used the FAT times for all places, where obviously the absolute times are irrelevant.
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby dj » Fri Aug 17, 2012 5:00 am

johnclark wrote:
dj wrote:
johnclark wrote:I can't find any evidence that the accuracy changes with place: the margin to 2nd place is measured just as accurately as the margin to 6th place.

The hand timing of 1960 seems to be more accurate than 1952 or 1956.


This begs my original question, where did you get the place times? My guess is that if they were derived from Wallechinsky or any modern compilation, there is the strongest likelihood that these are adjusted place times, based on the photo-timing evidence showing the time behind the winner. They do not represent what the hand-timers themselves too.


DJ, thanks for extra info. As I said above, I did get those times from Wallechinsky. However, for the later analysis (in the third (?) post above) of margins, I used the FAT times for all places, where obviously the absolute times are irrelevant.


Perhaps I'm missing your point, but if you're trying to see if the difference between hand and auto times for 2nd through 6th is different than that for 1st, you have to be using the original hand times that were taken at the finish line. What you're used are times that have been adjusted to provide internal consistency and therefore do not always reflect what the timers recorded.

Are you trying to analyze something different, which I'm just not seeing?
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby Conor Dary » Fri Aug 17, 2012 5:32 am

dj wrote:

Perhaps I'm missing your point, but if you're trying to see if the difference between hand and auto times for 2nd through 6th is different than that for 1st, you have to be using the original hand times that were taken at the finish line. What you're used are times that have been adjusted to provide internal consistency and therefore do not always reflect what the timers recorded.


Isn't that the truth! I remember quite a few times working meets where 4th or 5th, say, had a faster time than 3rd or above. Needless to say the times got adjusted.
Conor Dary
 
Posts: 6297
Joined: Sat Oct 08, 2005 4:32 am
Location: कनोर दारी in Ronald MacDonald's Home Town, and once a Duck always a Duck.

Re: question on hand timing and FAT 1952-64

Postby lonewolf » Fri Aug 17, 2012 7:34 am

Maybe slightly off-topic but hand time related... when my granddaughter was swimming in hand timed youth meets, places were awarded by parent hand-times instead of who reached the finish first.. often by indisputable margins. :?
lonewolf
 
Posts: 8814
Joined: Sat Oct 08, 2005 4:32 am
Location: Indian Territory

Re: question on hand timing and FAT 1952-64

Postby johnclark » Mon Aug 20, 2012 10:34 pm

dj wrote:Perhaps I'm missing your point, but if you're trying to see if the difference between hand and auto times for 2nd through 6th is different than that for 1st, you have to be using the original hand times that were taken at the finish line. What you're used are times that have been adjusted to provide internal consistency and therefore do not always reflect what the timers recorded.

Are you trying to analyze something different, which I'm just not seeing?



No, its probably me missing something. For that analysis I have used the hand times given here:

http://www.sports-reference.com/olympic ... final.html

These times correspond to what you showed as the 'original' hand times - 6th place is 10.5 rather than 10.6. I've assumed that all the times on this website are the original hand times. I should have made that clearer - sorry.
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby dj » Tue Aug 21, 2012 1:10 am

johnclark wrote:
dj wrote:Perhaps I'm missing your point, but if you're trying to see if the difference between hand and auto times for 2nd through 6th is different than that for 1st, you have to be using the original hand times that were taken at the finish line. What you're used are times that have been adjusted to provide internal consistency and therefore do not always reflect what the timers recorded.

Are you trying to analyze something different, which I'm just not seeing?



No, its probably me missing something. For that analysis I have used the hand times given here:

http://www.sports-reference.com/olympic ... final.html

These times correspond to what you showed as the 'original' hand times - 6th place is 10.5 rather than 10.6. I've assumed that all the times on this website are the original hand times. I should have made that clearer - sorry.


Exactly. As I pointed out earlier, only the first five times correspond to what was actually timed by the finish line hand timers. Sixth place was originally timed at 10.6. What the Olympic Sports site is showing is 10.5, a time subsequently adjusted by statisticians/historians after consulting the photo-timer.

As a result, you're giving credit for better hand timing than truly existed. If you want to find out what the hand timers recorded, you'll have to go to the contemporary newspaper reports, which almost always report the original hand times, unadulterated by subsequent reference to the photo-times.

Almost any website is going to give you "better" times, times that are closer to what the reality of the race and the difference between runners was. But these are farther from the reality of what the hand timers recorded.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby johnclark » Tue Aug 21, 2012 9:49 pm

dj wrote:Exactly. As I pointed out earlier, only the first five times correspond to what was actually timed by the finish line hand timers. Sixth place was originally timed at 10.6. What the Olympic Sports site is showing is 10.5, a time subsequently adjusted by statisticians/historians after consulting the photo-timer.


Oh now I am getting confused! You had previously listed 10.5 as the 'original' time and 10.6 as the adjusted time:

dj wrote:I've finally been able to confirm that the hand times you are using are the photo-adjusted hand times as they differ from the originals for 1952:

Yours for 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.6
Original 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.5


So is 10.5 the original time for 6th place John Treloar (as shown in the website) or is 10.6 the original time (as shown by Wallechinski)?
johnclark
 
Posts: 51
Joined: Tue Aug 07, 2012 4:46 am

Re: question on hand timing and FAT 1952-64

Postby dj » Wed Aug 22, 2012 6:25 am

johnclark wrote:
dj wrote:Exactly. As I pointed out earlier, only the first five times correspond to what was actually timed by the finish line hand timers. Sixth place was originally timed at 10.6. What the Olympic Sports site is showing is 10.5, a time subsequently adjusted by statisticians/historians after consulting the photo-timer.


Oh now I am getting confused! You had previously listed 10.5 as the 'original' time and 10.6 as the adjusted time:

dj wrote:I've finally been able to confirm that the hand times you are using are the photo-adjusted hand times as they differ from the originals for 1952:

Yours for 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.6
Original 1952: 10.4, 10.4, 10.4, 10.4, 10.5, 10.5


So is 10.5 the original time for 6th place John Treloar (as shown in the website) or is 10.6 the original time (as shown by Wallechinski)?



Sorry, I've mucked this up pretty good with this last post. I was right the first time, and the got it backwards the last time.

The originally recorded 6th place time in 1952 was 10.5. The 10.6 is an adjustment made by statisticians based on the photo times, and created to bring better internal consistency to the hand times.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby James Fields » Sat Aug 25, 2012 1:01 pm

dj wrote:
First place is the easiest place to time, last is the next easiest. Everything in between is difficult as the timer has to look through the pack, choose the person who's going to finish in the place assigned the timer, and stop the watch. Because there is fear of missing the person, there is a greater likelihood of hurrying the process of stopping the watch, thus recording a shorter time than when would would properly wait to see the finish and not anticipate it.

________

Agreed. I much prefer the system used years ago when I served as a volunteer official at Eugene meets: Each of us [hand-timers] as assigned a lane instead of a place for the "lanes all the way" sprint races.
James Fields
 
Posts: 775
Joined: Sat Oct 08, 2005 4:32 am
Location: Knoxville and Seattle

Re: question on hand timing and FAT 1952-64

Postby pinoyathletics » Mon Sep 03, 2012 4:51 pm

Would the number of stop watches used effect the differential between a hand time and electronic timing?. Lets take for example in the 100m.

One stop watch used 10.87 = 10.9ht.
Two stop watches used 10.87 and 10.91 = 11.0ht (10.91 must round up)
Three stop watches used 10.80, 10.87 and 10.91 = (10.9ht)

The standard addition is +0.24. But i'm presuming this is only a guide if 3 stop watches are used? which is usually more accurate. So is it possible it could be +0.3 with two stop watches and +0.4 or more with just one stop watch??
pinoyathletics
 
Posts: 26
Joined: Mon Sep 03, 2012 4:36 pm
Location: Australia

Re: question on hand timing and FAT 1952-64

Postby dj » Tue Sep 04, 2012 2:42 am

pinoyathletics wrote:Would the number of stop watches used effect the differential between a hand time and electronic timing?. Lets take for example in the 100m.

One stop watch used 10.87 = 10.9ht.
Two stop watches used 10.87 and 10.91 = 11.0ht (10.91 must round up)
Three stop watches used 10.80, 10.87 and 10.91 = (10.9ht)

The standard addition is +0.24. But i'm presuming this is only a guide if 3 stop watches are used? which is usually more accurate. So is it possible it could be +0.3 with two stop watches and +0.4 or more with just one stop watch??


What matters most is the number of GOOD timers. One bad timer can produce a faulty time. However, a head timer might just throw it out and create a reasonable time. For instance, the times in a high school boy's 100 might come to the head timer as 11.0, 11.1, 11.2, 10.9, and 11.4 for the first five placers. No head timer is going to say that the fourth place time is faster than the winner; it's going to be adjusted to either 11.2, 11.3 or 11.4 and no one will be the wiser.

Where there are multiple times, the longer times (almost always those closer to reality) will be taken. The longer of two watches, the agreement of two or the middle time where there are three watches. More watches makes for better timing IF there are more good hand timers.

But the best use of a small number of hand timers where there aren't enough to get three watches on each place is to pair a good timer with a bad timer, thereby overcoming the effect of the bad timers. This becomes critical in high school meets where two runners might be advancing on place and another several are advancing on time. It becomes critical to get those times right, so a head timer might shift the best timers to taking 3rd and 4th places during the heats, then move them back to first, second and down the line for the final.

Mind you, none of this is prescribed in the rules, and some of it creates situations that are against the rules (two of fewer timers on a place, for instance), but it's the reality of the situation.
dj
 
Posts: 6198
Joined: Sat Oct 08, 2005 4:31 am

Re: question on hand timing and FAT 1952-64

Postby Marlow » Tue Sep 04, 2012 3:51 am

In the 1960s Florida was famous for producing 9.4 and 9.5 HS 100y sprinters. It wasn't till I got here that I saw just how bad HS hand timing is here. I've trained myself to be pretty consistent around the +0.1 sec differential from FAT (spent a lot of time with fellow coaches at meets with FAT, trying to see who could get the closest - if I added 0.1 in my head to my stopwatch, I could predict the correct time +/- .05 about 80% of the time). When I go to local meets, I'll time the 100 and 110H and the 'official results' are typically .2 - .4 seconds faster than what I have (which is already +.1). :(
Marlow
 
Posts: 21079
Joined: Thu Jan 24, 2008 5:00 pm
Location: Somewhere over the . . . hill


Who is online

Users browsing this forum: No registered users and 7 guests