I pulled the rental onto the Virginia highway and saw no other cars, so I stomped the gas and enjoyed the kick. That particular Malibu seemed well set up and the needle headed for sixty pleasingly fast. A little too pleasingly, maybe. It occurred to me that automakers have a systematic bias in favor of making their speedometers lie, on the high side of course; it makes the customers happy and the police happy and probably saves some lives. For rentals, bought a fleet at a time, the pressure would be higher. I wouldn’t be surprised if some Operations-Research wonk in a back-room at Hertz or Avis or somewhere has done the numbers: for every mile per hour of positive error on the dial, so many fewer accidents, so many tens of millions of dollars of savings in repair bills, so many dozen fewer lawsuits. Plausible?
Contributions
Comment feed for ongoing:
From: Chad Okere (Oct 13 2006, at 01:45)
Get a GPS device if you want to <i>really</i> know how fast you're going. That's what I did when my speedometer flat out broke down. GPS will read a few miles an hour slower then my new speedometer.
[link]
From: Ryan Cousineau (Oct 13 2006, at 01:48)
Speedometers do indeed tend to err on the side of caution, for several reasons.
T & B had a <a href="http://truckandbarter.com/mt/archives/2004/06/on_the_accuracy.html">pretty good post</a> two years ago on the speedo/odo issue, but Car and Driver had an <a href="http://www.caranddriver.com/features/1906/speedometer-scandal.html">excellent article</a> on this in 2002.
Here's their <a href="http://www.caranddriver.com/features/1907/speedometer-scandal.html">summary chart of measured errors</a>, including breakouts by maker, continent, cost, and vehicle type.
Short version: the European standard requires that speedo error be on the high side. US standards are more complicated, but also encourage a bias towards faster-than-true (and on the odo side, further-than-true) readings.
GM was the most accurate manufacturer of those C&D tested.
[link]
From: Gareth Simpson (Oct 13 2006, at 02:23)
If you have certain models GPS navigation, they can tell you your speed independantly of the speedo, so you could always test your theory.
[link]
From: Rimantas (Oct 13 2006, at 03:08)
in EU we have directive http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32000L0007:EN:HTML
which says:
2.3.7. the speed displayed must never be lower than the actual speed. Between speed V1, read on the speedometer and actual speed V2 there must be the following relationship with the test values specified in item 2.3.5 and between those values:
0 <= (V1 - V2) <= 0,1 · V2 + 4 km/h.
[link]
From: james governor (Oct 13 2006, at 04:15)
Your Mileage May Vary....
[link]
From: Eamonn Sullivan (Oct 13 2006, at 04:48)
I suspect the same laws prohibiting mileage-fiddling might ensure that the speedometer is right.
[link]
From: Jesús Cea (Oct 13 2006, at 04:58)
You can't tweak "too much" that. If going from city A to city B, X km (I'm european, you know :) away (you see it in the highway signs) is taking too long, the tweak would be a bit obvious.
In a time with googleearth and GPS navigators, changing traffic signs to accomodate "new reality" couldn't work either :-p
[link]
From: james governor (Oct 13 2006, at 05:15)
Your Mileage May Vary....
[link]
From: Brett Duncan (Oct 13 2006, at 05:24)
[link]
From: Chuck Hinson (Oct 13 2006, at 07:09)
Don't know about the rental car companies, but I do know that the last three Hondas I've owned have had speedometers that were between one and five percent fast. I always figured they were just trying to make people think they were getting better gas mileage.
Like setting your clock to be five minutes fast, I'm not sure that there's really any long term benefit in it. The part that bothers me is that the odometer and speedometer are typically tightly coupled - if one is fast, so is the other. So, while I can compensate for the fact that my speedometer is fast, there's nothing I can do about the fact that my odometer will say that I've driven more miles than I actually have - and that's not a pleasant thought when you're pushing the mileage cap on your lease or contemplating resale value.
[link]
From: Dan Davies Brackett (Oct 13 2006, at 07:15)
It's reasonable, and it seems like there'd be no easy way to check that systematic bias without a friendly police officer. One way to check would be to drive a rental in front of a known vehicle, and see whether their spedometers match...
Of course, this kind of tomfoolery is only as effective when it's not acted on.
[link]
From: Curtis Pew (Oct 13 2006, at 07:32)
I’ve noticed that whenever I go past one of those “your speed is” installations that the police put up the value that flashes there is 2 to 5 MPH slower than what my speedometer reads. This has been true for both my Ford van and my Chevy Cavalier, and in several jurisdictions. I can’t really see the police setting their speed checkers to read lower than the true value, so I’ve also suspected that the speedometers are overestimating my speed.
[link]
From: John Cowan (Oct 13 2006, at 07:44)
U.S. law, FWIW, limits speedometer errors to 5%.
[link]
From: David Nesting (Oct 13 2006, at 08:07)
I suspect it has more to do with the question of legal liability.
If your speedometer calibration technique has a certain intrinsic margin of error, and tire variations increase that margin further, you are going to have some people with vehicles underreporting their speed. Some percentage of these people will get into accidents. Some percentage of those will be litigious and try to blame/extort money from the car manufacturer, suggesting that the driver's high speed was due to the false reading on the speedometer, and that's what caused baby Joseph to die. The car manufacturer settles for $NNN,000 to avoid a higher penalty that would be imposed by sympathetic/angry jurors trying to "send a message".
It's far cheaper to avoid this scenario altogether by figuring out the possible margin of error and causing the speedometer to underreport.
[link]
From: Gregory Williams (Oct 13 2006, at 08:24)
I rented a Kia Sorento from Hertz just last week. Using a GPS I couldn't see any error at 65-70 mph. It still seems reasonable to me that this might happen in certain situations, but I've never noticed it myself.
[link]
From: Ryan Cousineau (Oct 13 2006, at 09:36)
Dan: are there no mileage markers? Are there no stopwatches?
[link]
From: Hub (Oct 13 2006, at 09:46)
The other day on the Ottawa River Parkway, there was that big speedometer telling you how fast you were driving. It was showing 52kmph. My car showed 60kmph. The speed limit is 60kmph.
Go figure.
[link]
From: Frederick Dean (Oct 13 2006, at 11:07)
The speedometer reading depends on tire wear because worn tires have a smaller diameters. Likewise, tire inflation also affects reported speed.
[link]
From: Rob (Oct 13 2006, at 11:15)
The reverse might also be true on occasion: with a family of five I am a confirmed minivan driver. When we got our Chevy Lumina, I was startled to notice that the speedometer only went up to 140K, usually speedometers are insanely optimistic about what old Bessie could do if you really actually wanted to. I had that thing cranked up above 140 a couple of times, and it felt like it had plenty of oomph left, it had a pretty gutsy motor.
So if the speedometer range can be set low to reassure my inner soccer-mom, perhaps so can the speedometer measurement, to let my her feel better about driving too fast?
[link]
From: Mark Levison (Oct 13 2006, at 11:49)
That's what mile/kilometer posts are for. When you're driving on a highway (in Ontario at least) you will find a small green sign every kilometer. Assuming these are accurate you check how much your odometer differs - mine overestimates by 2-3%. So I assume the same is true for the speedometer.
Cheers
Mark Levison
[link]
From: David Singer (Oct 13 2006, at 14:03)
One might also consider that having the odometer read a little high makes service intervals come more often and warranties expire sooner.
[link]
From: Simon Brocklehurst (Oct 14 2006, at 05:34)
Tim - is your hypothesis plausible? I'm not sure that it is. It seems to rely on there being a strong correlation between speed and number of accidents. I don't think there is: the roads on which there are the *smallest* numbers of accidents tend to be the roads which have the *highest* speed limits.
[link]
From: John Cowan (Oct 14 2006, at 08:09)
There is certainly a correlation between speed and severity (perhaps number) of accidents *at a given location*. That says nothing about high-speed highways being less safe than low-speed ones (probably safer because more modernized in other ways).
[link]
From: Tim (Oct 14 2006, at 20:19)
Interesting post on the speedometer issue - and coming right after another manufacturer-installed measurement "error" post too!
Not on speedometers per se, but Dad had a Cadillac (Cimarron? Catera? something two-door) in the eighties that drove him nuts: it had an electronic miles per gallon reading which changed as you were driving along, supposedly to help you manage fuel consumption better. What drove him crazy? When stopped at a stop sign, the MPG calculator <b>refused</b> to read zero! Dad always figured, correctly to my mathematically-inclined mind, that if you were going zero miles, your MPG was zero as well.
[link]
From: Oliver (Oct 15 2006, at 05:20)
here in Australia, some freeways and highways have speed readouts to check this very thing. while it would be a concern if the speedo read under the actual speed, most cars I have driven are 5 Kph (kilometers per hour) OVER the actual speed measured and displayed.
cheers,
- Oliver
[link]
From: ben lowery (Oct 16 2006, at 11:50)
According to my GPS (Garmin Nuvi), my wife's 2004 Mini Cooper is pretty consistently 2mph fast (reports 60 when I'm going 58) and my 1999 Audi A4 is spot on.
[link]
From: Eric (Oct 20 2006, at 17:04)
I tend to check any new car I drive (either our own or any rental).
Most US-built cars and some Japanese cars read 2-4 mph fast at highway speeds.
Our Audi and Lexus are spot on.
[link]