9.77!

[Daisy]

So all this being said, does anyone know WHO ran faster: Justin or Asafa?

Head to heads are what counts now.

[SQUACKEE]

I know this gonna sound trite but can we all agree Gatlin ran fast, real fast.

[Daisy]

edit - just read gh's post and I will have to defer to his judgement of the 1/1000ths resolution accuracy (still not fully convinced). If .01 is usually equated to 3 inches, then we're talking about 1 cm (approx) in thousandths. Y'all have seen finish photos - do we really have 1cm accuracy in reading of the athletes torsos? Really?

Tafnut, did you read the links i posted for comments from jla at this link and tandfman at this link?

[tafnut]

Tafnut, I have voiced these concerns in the past and have had some very good responses from jla
at this link and tandfman at this link

Thanks for the link, Daisy, I remember those discussions and while I fully agree that the EQUIPMENT may be accurate to 5/10000s of a second, that sure as HECK does NOT mean that the human being with his little cursor is. Those finish photos in the first link corroborate my point. There's just no way someone can distinguish the flailing arms, shoulders, necks and chests down to 1/1000s (IMO). I remain skeptical.

[guru]

Accuracy to 1/1000s is simply not possible IMO! I've seen 1/10s errors just from an errant eyeball.

Not exactly true. It's very easy for a competent operator with good equipment to distinguish 1000ths. Used to break ties all the time.

gh is absolutely correct. In fact, one of my 4x400 teams was involved in a barnburner in 2004.

At the Ohio state champs, in a race for second, we nosed out third place 3:50.275 to 3:50.289. I was told it amounted to about 1 inch. Resulted in the third place team being knocked back into a tie for the team championship, as well as preserving our Cincinnati city record(they were also from Cincy) rounded to 3:50.28. The other coach contested the result, but to no avail. Probably helped that the FinishLynx was being run that day by FinishLynx personnel.

[gh]

There's just no way someone can distinguish the flailing arms, shoulders, necks and chests down to 1/1000s (IMO). I remain skeptical.

Have you ever actually looked at high-end photo-timing system on-screen?

[Daisy]

There's just no way someone can distinguish the flailing arms, shoulders, necks and chests down to 1/1000s (IMO). I remain skeptical.

Have you ever actually looked at high-end photo-timing system on-screen?

This is a point that someone else made previously (can't fiond it in the search). What we see in the press release is a very low resolution version of what they get in the photofinish booth.

[tafnut]

And I have no problem with breaking ties down in the sub-01 realm, as long as the coaches can agree that one athlete put some part of his torso in front of someone else's. But TIMING ACCURACY to 1/1000s is TOTALLY dependent on HUMAN JUDGEMENT and sorry, folks, I ain't buying it. I suppose there are very experienced readers who have been through this enough to pretty darn good at it, but essentially we're back to 'trusting' people to get it right, not machines, and humans are prone to ERROR. Especially down in the 1/1000s range.

I'm probably in a minority of 1 here, but that's OK, I'm used to it! :-)

[bekeselassie]

So all this being said, does anyone know WHO ran faster: Justin or Asafa?

Head to heads are what counts now.

Oh I don't want to get dogmatic about saying so-and-so is the "real" world record holder. I'm just curious. Given this method of determining the time to the hundredths of a second, is there any record of what Asafa's time was to the thousandth. It's just for kicks.

Squackee, now we know that Gatlin did NOT run fast. Got it now?

[tafnut]

There's just no way someone can distinguish the flailing arms, shoulders, necks and chests down to 1/1000s (IMO). I remain skeptical.

Have you ever actually looked at high-end photo-timing system on-screen?

And there's the rub . . . we have a new FinishLynx system, but I'm sure that the BigBoiz have substantially better screen resolution. Are you telling me that the resolution is high enough to distinguish body parts to less than a cm? Assuming that's a 'yes', isn't it still a judgement call as to what constitutes the 'first part of the torso' as evidenced in Daisy's first link above? My skepticism is still born of the human margin of error, NOT the machine's.

[gh]

And I have no problem with breaking ties down in the sub-01 realm, as long as the coaches can agree that one athlete put some part of his torso in front of someone else's. But TIMING ACCURACY to 1/1000s is TOTALLY dependent on HIMAN JUDGEMENT and sorry, folks, I ain't buying it. I suppose there are very experienced readers who have been through this enough to preet darn good at it, but essentially we're back to 'trusting' people to get it right, not machines, and humans are prone to ERROR. Especially down in the 1/1000s range.

I'm probably in a minority of 1 here, but that's OK, I'm used to it! :-)

I repeat my question: have you ever actually looked at one of these timing systems, or are you just talking through your hat?

[JRM]

Ideally, this error should not have happened. Their software was faulty, for some reason.

I am still very skeptical that a person can line up a cursor with thousandths accuracy, just by looking at the (somewhat small, somewhat blurred) picture and 'judging' the torso (shoulder, neck and arm often confuse the picture).
...
If .01 is usually equated to 3 inches, then we're talking about 1 cm (approx) in thousandths. Y'all have seen finish photos - do we really have 1cm accuracy in reading of the athletes torsos? Really?

Two things: if you have a sufficiently high-powered computer, photo finish images have exceedingly high resolution. It becomes very easy to see the person's outline. Choosing where on the torso you should click also doesn't affect the results (probably less than a 0.001s difference).

Also, the photo finish is not a "photo" in the usual sense. Rather, it's a more of a "graph" that represents what was crossing the finish line at successive times. The camera is taking a picture of the line -- it's a very thin aperture. The image of the athlete is actually a trace of his/her body crossing the finish line at different times. That's why you get funky "stretching" or "warping" effects, particularly ini the sprinters at the end (because they slow down as the cross the line). If you look at the image in question, you'll see that Gatlin began crossing the line at about 9.74s (right foot), and finished crossing at about 9.83s (left foot).

Now, all this being said, just eyeballing that lower-resolution photo finish makes it quite clear that the time should have been 9.77 from the get-go, no ifs, ands, or buts. The red torso line on Gatlin falls cleanly between 9.76 and 9.77s, which according to the rules must become 9.77s. In fact, even without the IAAF rounding rule, standard measurement techniques say that mark should have been 9.77 -- it's closer to .77 than .76.

[tafnut]

I repeat my question: have you ever actually looked at one of these timing systems, or are you just talking through your hat?

I think you are talking nuts and bolts and I am talking flesh and blood. Yes, the SYSTEM is that accurate - no, the people are not. As for talking though my hat, I rarely wear one, but as we know, I am human, therefore prone to error, so there is a distinct possibility (probability . . . certainty?) that I'm wrong here too! Sigh. The VERY FACT THAT WE'RE HAVING THIS DISCUSSION IS BECAUSE SOMEONE SCREWED UP THE READING OF THE PHOTO!! People mess up. That should be a clue. :-)

[Daisy]

I'm just curious. Given this method of determining the time to the hundredths of a second, is there any record of what Asafa's time was to the thousandth. It's just for kicks.

I understand your curiosity bu the environmental differences mean the thousandths are kind of moot, even the hundredths, which is why people are always refering to the basic times.

[tafnut]

Here's a concession - if JRM (or gh) is telling me that the finish photo can be blown up so that 1cm of athlete = 3 inches of monitor image, and that the image is crystal clear, then yes, even I could probably distinguish 1/1000s - ASSUMING THERE IS NO CONTROVERSY OF WHAT CONSTITUTES THE TORSO. How's that?