Paul
Kraght gave me permission to post this discussion of HIS timing method. Use
the "Back" button on your browser to return to my web page after reading
about Paul's method.
Jay O'Brien
-------- Original Message --------
Subject: Congratulations!!
Date: Mon, 23 Sep 2002 15:00:20 -0700
From: phkraght@beckman.com
To: jayobrien@att.net
Jay,
Congratulations on your well-deserved victory!! You've just been getting
closer and closer, and now it will be a lot harder to get even closer...
You and Ted Zaleski are putting the heat on me! I'm going to have to
start driving with my Walkman-navigator again. That's the only way
I've been able to achieve precision timing. Of course, there's the
matter of having a car to drive...
I'd love to pick your brain, but I suspect there's a lot of well-guarded
secrets that you wouldn't want to divulge to the teacher of the navigators'
class at MKM! Again, well done! My only regret was that I wasn't
there!
-- Paul
-------- Original Message --------
Subject: Re: Congratulations!!
Date: Mon, 23 Sep 2002 23:17:29 -0700
From: Jay O'Brien <jayobrien@att.net>
To: Paul Kraght <phkraght@beckman.com>
Paul,
Thank you very much. There really aren't any secrets, except perhaps the
actual measured course distances, which anyone can determine independently.
I plan to discuss our process in the web page I will put together for this
event. Have you documented your process anywhere on the web? It would be
informative to include a link to your work if it is available.
Jan and I really appreciate your comments. You make us feel as if we have finally "arrived"! We hope to see you soon.
Regards,
Jay
-------- Original Message --------
Subject: Re: Congratulations!!
Date: Tue, 24 Sep 2002 11:07:54 -0700
From: phkraght@beckman.com
To: "Jay O'Brien" <jayobrien@att.net>
Jay,
No, I haven't documented my process anywhere, though I have discussed it
with a number of people. In my case, I'm not worried about giving away
any secrets because my method only works for solo drivers who are willing
to fiddle with a Walkman while they drive!
Basically, there are three steps. The first is to measure the course.
The distance measurements can be done a variety of ways, and need not be
very accurate except close to the end. I've gotten my best measurements
by strapping on a measuring wheel to my car wheel. I forget the brand
name, but I got it from AW Direct, a towing-equipment supply house that somehow
got my name on their mailing list -- I have no idea how it got there!
This measuring wheel is the sort with a telescoping handle that you're supposed
to roll along the ground, and it reads the distance in feet-inches, or meters-centimeters.
It has a 14.5" diameter, so when I strap it to my car wheel, it still thinks
it's measuring with the 14.5" diameter, which is just a little over half
the tire's actual diameter. But, other than a scalefactor, the measurements
are surprisingly precise. For instance, I measured the last several
miles of Gambler's Run in 2000, but when the finish line was moved two days
before the race in 2001, I had to remeasure the section. When I adjusted
for the scalefactor, the measurements differed by an average of only 1.5
feet over the several miles! However, I don't have as good of measurements
for SSCC. There, I counted tire revolutions for the last 3 miles, so
the error is within 6 feet or so. For the last 3 miles of every course,
I measure all the roadposts, not just the mileposts.
Additionally, for the speeds I like to drive, I need to measure the curve
radii and lengths, so I have an objective way of giving speed ratings to
the curves. I then adjust for any road conditions, such as off-camber, rough
surface, blind entrance, etc. This isn't necessary for timing, but
it gives me a much smoother ride and allows me to go faster than my driving
skill would otherwise permit. "It's easy to go fast when you know when to
slow down."
The second step is to prepare a driving plan. I have software which
simulates a run through the course according to the driving instructions
I supply and a model of the car's performance. The performance data
includes weight, torque curve, gear ratios, and aerodynamic drag. Also,
I include information about the altitudes, as this affects the performance
in two ways: 1) by affecting the engine output, and 2) taking or giving back
power due to going up or down hills. The driving instructions include the
speeds around the turns, shift speeds, cruise speeds, and if and how hard
I want to brake. I adjust the driving instructions until the program
tells me I have achieved the exact target time. The program generates a list
of target times for all the measured roadposts.
The third step is to prepare the navigator recording. The recording
has three parts: 1) Spoken driving instructions, such as "righthander coming
up, right after the crest, rated 136 mph. It is 0.4 miles long."
Also timing info such as "Mile 36 time is 37 minutes and 9 seconds."
2) Computer-generated beeps at the target times for the roadposts, as computed
by the previous step. 3) Timing track. This consists of one beep
every second. Every tenth second it has a different tone, and the first
beep of each minute has yet a different tone. This allows me to synchronize
the tape to my stopwatch. Incidentally, my Walkman has a speed adjustment
which is normally turned off. About 12 miles before the finish, I turn
it on and adjust the tape until the timing track exactly synchronizes with
the stopwatch. Then I turn it back off. The beeps at the roadposts
guide me the rest of the way in. The reason I measure all the roadposts
near the end is so I get timing guidance every few seconds, which keeps me
from drifting between the mileposts. In fact, in the last few miles,
I don't use either my speedometer or stopwatch -- it's solely by ear and
eye! (Of course, I try to remember to get a split at the finish line...)
As with you, I have had to determine my stopwatch delay. The 2000 Gambler's
run allowed me to characterize this, especially since it was a 2-way event,
so I got two timing-delay measurements. Using this data, the next two
events I drove were both within 0.1 second error.
Unfortunately, I can't seem to get things to work anywhere near this well
as a navigator, as my last 4 runs demonstrate! Interestingly, it seems
that my timing delay was all over the map as a navigator, but very consistent
as a driver. Having lost all faith in my reactions, I wrote some software
to simulate the timing lights and time my delay. I learned two things:
1) my reactions were remarkably consistent on the computer, and 2) the delay
depends somewhat on how much dead time there is between the last amber going
out and the green going on. With 0.1 seconds of dead time, I averaged
about 0.08 seconds, with a SD of about 0.02. With 0.2 seconds dead time,
I averaged closer to 0.13 seconds. I didn't do the longer delay enough
times to get a reliable SD, but the results were definitely less well controlled.
Incidentally, Dabney told me that he might be fixing up an LED timing tree
next year for MKM to use. If that is so, they will have virtually zero
dead time.
Anyway, that's my story. Lots of preparation, but it's sure fun when everything comes together!
-- Paul
-------- Original Message --------
Subject: Re: Congratulations!!
Date: Fri, 27 Sep 2002 10:26:00 -0700
From: Jay O'Brien <jayobrien@att.net>
To: Paul Kraght <phkraght@beckman.com>
Hi Paul,
Would it be ok with you if I put the message thread (above) on my web site and linked to it from my September 2002 page?
Jay
-------- Original Message --------
Subject: Re: Congratulations!!
Date: Fri, 27 Sep 2002 10:45:53 -0700
From: phkraght@beckman.com
To: "Jay O'Brien" <jayobrien@att.net>
I would be honored!
-- Paul