snopes.com Post new topic  Post a reply
search | faq | forum home

  next oldest topic   next newest topic
» Hello snopes.com » Urban Legends » Trivia » If 99.9% Is Good Enough, Then . . .

 - UBBFriend: Email this page to someone!    
Author Topic: If 99.9% Is Good Enough, Then . . .
snopes
Return! Return! Return!


Icon 203 posted      Profile for snopes   Author's Homepage   E-mail snopes       Edit/Delete post   Reply with quote 
Comment: Someone sent this to me, and I've seen it used in training
classes a number of times. Are these figures accurate or just made up?
Some of them seem like they've got to be somewhat exaggerated.

If 99.9% Is Good Enough, Then . . .


12 newborns will be given to the wrong parents daily.

114,500 mismatched pairs of shoes will be shipped/year.

18,322 pieces of mail will be mishandled/hour.

2,000,000 documents will be lost by the IRS this year.

2.5 million books will be shipped with the wrong covers.

Two planes landing at Chicago's O'Hare airport will be unsafe every day.

315 entries in Webster's dictionary will be misspelled.

20,000 incorrect drug prescriptions will be written this year.

880,000 credit cards in circulation will turn out to have incorrect
cardholder information on their magnetic strips.

103,260 income tax returns will be processed incorrectly during the year.

5.5 million cases of soft drinks produced will be flat.

291 pacemaker operations will be performed incorrectly.

3,056 copies of tomorrow's Wall Street Journal will be missing one of the
three sections.

Posts: 36029 | From: Admin | Registered: Feb 2000  |  IP: Logged | Report this post to a moderator
TrekkerScout
Deck the Malls


Icon 1 posted      Profile for TrekkerScout     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by snopes:
18,322 pieces of mail will be mishandled/hour.

Prior to the Seattle Bulk Mail Center moving to its new location, there used to be a prominantly displayed sign on top of the building that said "##% accuracy" (where ## was changed periodically). The highest number I ever saw on the sign was 97%.
Posts: 306 | From: Tacoma, WA | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Troberg
Angels Wii Have Heard on High


Icon 1 posted      Profile for Troberg     Send new private message       Edit/Delete post   Reply with quote 
I would guess that most of the numbers quoted are actually worse in real life. Probably much worse.

--------------------
/Troberg

Posts: 4360 | From: Borlänge, Sweden | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
Esprise Me
We Wish You a Merry Giftmas


Icon 1 posted      Profile for Esprise Me     Send new private message       Edit/Delete post   Reply with quote 
That line about the soft drinks makes me think this was made up. In between "your planes will be unsafe" and "they'll screw up your surgery" is..."your soda might be flat!" Oh, my God! Whatever shall we do?

--------------------
"If God wrote it, the grammar must be infallible. Perhaps it is we who are mistaken." -MapleLeaf

Posts: 977 | From: Boston, MA | Registered: Oct 2005  |  IP: Logged | Report this post to a moderator
Grand Illusion
Jingle Bell Hock


Icon 1 posted      Profile for Grand Illusion   Author's Homepage   E-mail Grand Illusion   Send new private message       Edit/Delete post   Reply with quote 
The business concept behind this piece is the "rule of nines." In the real world, it's impossible to achieve a 100% accuracy rate on anything of significant quantity, but you can get to 99% or 99.9% or 99.999999999999999%. The rule of nines states that the more nines you add, the more costly and complex the guarantee is.

To someone who doesn't understand business process, 99.9% may sound like an extremely high ratio. In reality, it may or may not be, depending on the context. If the police solve 99.9% of all crimes, that's pretty good, and I doubt that any major city's cops have that good a record. And a 99.9% recovery rate for cancer is also much better than reality. However, the OP states some cases where 99.9% may not be very good. Having too much tolerance for error or having too little tolerance can both be detrimental to a product or service, and it's a business decision to determine how many nines to build into a process.

--------------------
There are 10 types of people in the world: those who understand binary and those who do not.

"Are you pondering what I'm pondering?" - The Brain

Posts: 587 | From: Colorado | Registered: Jul 2000  |  IP: Logged | Report this post to a moderator
Aptenodytes_Forsteriis
We Wish You a Merry Giftmas


Icon 1 posted      Profile for Aptenodytes_Forsteriis     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Troberg:
I would guess that most of the numbers quoted are actually worse in real life. Probably much worse.

]

I have to agree, I don't believe a human being can achieve 99.9% accuracy on any complicated multi-step task, at least not for long.

--------------------
'Hello, assorted humanoid strangers. You are standing casually in our forest. This bewilders us.' Blatherskite

Posts: 950 | From: Cincinnati, Ohio | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Richard W
Ding Dong! Merrily on High Definition TV


Icon 1 posted      Profile for Richard W   E-mail Richard W   Send new private message       Edit/Delete post   Reply with quote 
quote:
Two planes landing at Chicago's O'Hare airport will be unsafe every day.
What does that mean?

They must have phrased it that way because it's obvious that two planes a day don't crash and that this is because airlines actually have a much better than 99.9% safety record...

Posts: 8725 | From: Ipswich - the UK's 9th Best Place to Sleep! | Registered: Feb 2000  |  IP: Logged | Report this post to a moderator
jimmy101
The First USA Noel


Icon 1 posted      Profile for jimmy101   E-mail jimmy101   Send new private message       Edit/Delete post   Reply with quote 
Lets see, about 5.9 million votes were cast in the Florida 2000 presidential election.

99.9% accuracy of 5.9 million is 5,900 votes.

Bush won by 540 votes.

To get an overall accuracy of +/- 540 votes you need an accuracy rate of 99.99%.

Few things are 99.99% accurate.

edit: one <> won, plus another typo

Posts: 629 | From: Greenwood, IN | Registered: Dec 2005  |  IP: Logged | Report this post to a moderator
Mr. Billion
The First USA Noel


Icon 1 posted      Profile for Mr. Billion   Author's Homepage   E-mail Mr. Billion   Send new private message       Edit/Delete post   Reply with quote 
Sounds fairly reasonable, although of course there's no sources for the initial figures multiplied by .999 to checj it all against.

I've read a mathematical demonstration that with a 98% accurate test for AIDS, still most of the positive results will be wrong. The logic is sort of convoluted.

--------------------
"For the U.S. to get involved militarily in determining the outcome of the struggle over who's going to govern Iraq strikes me as a classic definition of a quagmire." ~Dick Cheney.

Posts: 747 | From: Kansas | Registered: Jul 2005  |  IP: Logged | Report this post to a moderator
Grand Illusion
Jingle Bell Hock


Icon 1 posted      Profile for Grand Illusion   Author's Homepage   E-mail Grand Illusion   Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Richard W:
quote:
Two planes landing at Chicago's O'Hare airport will be unsafe every day.
What does that mean?

They must have phrased it that way because it's obvious that two planes a day don't crash and that this is because airlines actually have a much better than 99.9% safety record...

That probably means that two planes would not pass a safety inspection even if they never crashed.

--------------------
There are 10 types of people in the world: those who understand binary and those who do not.

"Are you pondering what I'm pondering?" - The Brain

Posts: 587 | From: Colorado | Registered: Jul 2000  |  IP: Logged | Report this post to a moderator
WingedBear
A View to a Krill


Icon 1 posted      Profile for WingedBear   E-mail WingedBear   Send new private message       Edit/Delete post   Reply with quote 
Why are the numbers so off for drug prescriptions and credit cards? Do most people really have 44 times more credit cards on hand at any one time than they recieve drug prescriptions over the course of a year?

I've gotten at least two prescriptions this year already. Am I supposed to have 88 credit cards? I really need to start accepting all those pre-approved credit card offers I get in the mail if I'm going to catch up!

--------------------
If the sum of the hypotenuse is equal to the sum of the square on the other two sides, why is a mouse when it spins?

Posts: 90 | From: Cleveland, OH | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
kanazawa
I'll Be Home for After Christmas Sales


Icon 1 posted      Profile for kanazawa   E-mail kanazawa   Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Grand Illusion:
[That probably means that two planes would not pass a safety inspection even if they never crashed.

It's more complicated than that. I'm sure they have a system in place to inspect and maintain the aircraft before anything becomes unsafe.

From www.airnav.com: Average 2627 operations a day at O'Hare. I think an operation is either take off or landing, so that's about 1313 landings. So if 2 crash (correct my math if need be) that comes out to something like 99.8%.

I could try to look up how many planes actually crash at O'Hare each year (and average it over days in a year), but I'm thinking most plane crashes don't happen at the airport (or landing anyway). Takeoffs are the most dangerous...

--------------------
If I say it's safe to surf this beach, then it's safe to surf this beach...

Posts: 127 | From: Cupertino, CA | Registered: Mar 2006  |  IP: Logged | Report this post to a moderator
kanazawa
I'll Be Home for After Christmas Sales


Icon 1 posted      Profile for kanazawa   E-mail kanazawa   Send new private message       Edit/Delete post   Reply with quote 
Hmmm...I checked between years 2000 and 2001 (at www.ntsb.gov) for all of the accidents/incidents in chicago, and came up with 6 (2 incidents, meaning something bad came close to happening, and 4 non fatal accidents). So (again, help with the math) that comes out to being something like 99.999% , right?

--------------------
If I say it's safe to surf this beach, then it's safe to surf this beach...

Posts: 127 | From: Cupertino, CA | Registered: Mar 2006  |  IP: Logged | Report this post to a moderator
Grand Illusion
Jingle Bell Hock


Icon 1 posted      Profile for Grand Illusion   Author's Homepage   E-mail Grand Illusion   Send new private message       Edit/Delete post   Reply with quote 
On a layover in Salt Lake City once, my wife and I had a flight delayed two hours because one of the doors in our plane wasn't sealing properly. They rearranged the seating so that nobody was in that section and they roped the section off and they allowed us to get on and take off. That was an unsafe enough condition to warrant action, but obviously not severe enough to ground the plane. If that's the definition of "unsafe" that the OP used, I would have no problem believing that 2 of that happens every day at O'Hare.

--------------------
There are 10 types of people in the world: those who understand binary and those who do not.

"Are you pondering what I'm pondering?" - The Brain

Posts: 587 | From: Colorado | Registered: Jul 2000  |  IP: Logged | Report this post to a moderator
James D
Deck the Malls


Icon 203 posted      Profile for James D   E-mail James D   Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Mr. Billion:
Sounds fairly reasonable, although of course there's no sources for the initial figures multiplied by .999 to checj it all against.

I've read a mathematical demonstration that with a 98% accurate test for AIDS, still most of the positive results will be wrong. The logic is sort of convoluted.

It sounds quite reasonable, noting the large numbers involved. For those who didn't notice, they are not saying that these things are happening, just that if there was a 99.9% success rate this many errors would sitll occur. I would be feasable with a bit of digging to find if the assumptions for the totals are correct (for a 0.1% failure rate, multiply each number by 1000 to get the assumed totals.

To explain one out longhand, if 12,000 infants are born daily, and 99.9% are given to the correct parents, 12 still end up in the wrong hands.

The AIDS test (this is also often applied to arguements about drug tests) involves a piece of math known as Bayes theorem

Now, according the The CIA world factbook, about 0.6% of the US population has AIDS. For the purpose of this example, we will assume the test 98% accurate, rounding where nessisary.

Take a sample of 1000 Americans, chosen at random. Out of that thousand, about 6 will have AIDS. Now we apply the 98% accurate test. Of the 994 who do not have AIDS, about 974 will test negative and 20 will test positive.

Of the six that have AIDS, All will probably test positive.

So, out of 26 positive tests 6 are true positives (about 23%) and 20 (about 77%) will be false positives.

--------------------
The best measure of a man's honesty isn't his income tax return. It's the zero adjust on his bathroom scale.
Arthur C. Clarke (1917 - )

Posts: 244 | From: Ventura, CA | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Mr. Billion
The First USA Noel


Icon 1 posted      Profile for Mr. Billion   Author's Homepage   E-mail Mr. Billion   Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by James D:
It sounds quite reasonable, noting the large numbers involved. For those who didn't notice, they are not saying that these things are happening, just that if there was a 99.9% success rate this many errors would sitll occur. I would be feasable with a bit of digging to find if the assumptions for the totals are correct (for a 0.1% failure rate, multiply each number by 1000 to get the assumed totals.



Yes, that's just what I was getting at.

quote:
Originally posted by James D:
The AIDS test (this is also often applied to arguements about drug tests) involves a piece of math known as Bayes theorem

Again, exactly what I was referring to, though I didn't remember the name. Thanks.

--------------------
"For the U.S. to get involved militarily in determining the outcome of the struggle over who's going to govern Iraq strikes me as a classic definition of a quagmire." ~Dick Cheney.

Posts: 747 | From: Kansas | Registered: Jul 2005  |  IP: Logged | Report this post to a moderator
Towknie
We Three Blings


Icon 1 posted      Profile for Towknie   E-mail Towknie   Send new private message       Edit/Delete post   Reply with quote 
Our company reached 99.5% of revenue goal last year, and you would have thought the sky was falling. Major catasrophy, heads rolling, feet stomping. I guess in some aspects, anything less than 100% is unacceptable.

--------------------
Towknie: Ryda-certified as wonderful, enlighted, and rational.

Posts: 1011 | From: Frisco, TX | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Singing in the Drizzle
Jingle Bell Hock


Icon 1 posted      Profile for Singing in the Drizzle     Send new private message       Edit/Delete post   Reply with quote 
So if there is a 99.9% chance that there is no life around another sun. There will still be life around aproximatly 100,000,000,000,000,000 suns in the univers.
Posts: 597 | From: Bellingham, WA | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
It only depends how you measure whatever you're talking about. For example, in the dictionary example, perhaps 99.9% could mean 99.9% of the entry titles (in which case the OP may be correct) or it may mean 99.9% of printed editions (in which case you may have to wait a few hundred years before seeing something not "good enough"... whatever that means) or it could mean 99.9% of the definitions are correct (again, with the ambiguous definition of "good enough") which, if you are extremely strict about your definition of correct, is probably par for most dictionaries. Anyway, the definition of "good enough" and the criteria for judgement are far too ambiguous even for the examples given.
Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by jimmy101:
Few things are 99.99% accurate.

What makes you say so? My computer is much more accurate than that. And the calculations for shooting rockets at other planets are much more accurate and the motions of all the planets themselves conform very accurately to the same calculations, and the measurement of a micrometer can be much more accurate, and UPC barcode readers are more accurate, and the numbered pages in all of the books in the library are more accurate, and aircraft safety is more accurate... (ad infinitum) What makes you think that this is so rare?
Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Troberg
Angels Wii Have Heard on High


Icon 1 posted      Profile for Troberg     Send new private message       Edit/Delete post   Reply with quote 
quote:
That probably means that two planes would not pass a safety inspection even if they never crashed.
I'd say that it would be flights that had some kind of incident report. Incident reports do not necessarily mean any risk was involved, just the possibility of one.

--------------------
/Troberg

Posts: 4360 | From: Borlänge, Sweden | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
Aptenodytes_Forsteriis
We Wish You a Merry Giftmas


Icon 1 posted      Profile for Aptenodytes_Forsteriis     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Ganzfeld:
quote:
Originally posted by jimmy101:
Few things are 99.99% accurate.

What makes you say so? My computer is much more accurate than that. And the calculations for shooting rockets at other planets are much more accurate and the motions of all the planets themselves conform very accurately to the same calculations, and the measurement of a micrometer can be much more accurate, and UPC barcode readers are more accurate, and the numbered pages in all of the books in the library are more accurate, and aircraft safety is more accurate... (ad infinitum) What makes you think that this is so rare?
Doesn't it depend on the scale at which you examine these things? Your computer certainly performs much higher than 99.9% when examined on a per calculation basis, but I suspect that a computer company would be delighted to have 99.9% of their products (individual computers) work as designed. I suspect that more than 1/1000 has some defect. Just as I suspect (make that have observed) that NASA has more than 1 major failure/1000 missions.

So this debate depends on what level you examine things at. If each unit is one calculation many human activities are probably better than 99.9% accurate, on the other hand if we look at the project level 99.9% may be an unattainable dream. In either case this list in the OP is meaningless. Either we are doing better thna the list suggests or a 99.9% success rate is a good thing.

--------------------
'Hello, assorted humanoid strangers. You are standing casually in our forest. This bewilders us.' Blatherskite

Posts: 950 | From: Cincinnati, Ohio | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Troberg
Angels Wii Have Heard on High


Icon 1 posted      Profile for Troberg     Send new private message       Edit/Delete post   Reply with quote 
quote:
Doesn't it depend on the scale at which you examine these things? Your computer certainly performs much higher than 99.9% when examined on a per calculation basis, but I suspect that a computer company would be delighted to have 99.9% of their products (individual computers) work as designed.
Or 99.9% uptime...

--------------------
/Troberg

Posts: 4360 | From: Borlänge, Sweden | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
jimmy101
The First USA Noel


Icon 1 posted      Profile for jimmy101   E-mail jimmy101   Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Ganzfeld:
quote:
Originally posted by jimmy101:
Few things are 99.99% accurate.

What makes you say so? My computer is much more accurate than that. And the calculations for shooting rockets at other planets are much more accurate and the motions of all the planets themselves conform very accurately to the same calculations, and the measurement of a micrometer can be much more accurate, and UPC barcode readers are more accurate, and the numbered pages in all of the books in the library are more accurate, and aircraft safety is more accurate... (ad infinitum) What makes you think that this is so rare?
Actually, no your computer is not more accurate than that, in any useful way anyway. Yes, give it a set of numbers and it will do a computation to high accuracy. But, how accurate were the original numbers? In the real world, it doesn't matter if your computer is 99.9999999% accurate if the input numbers are only 99% accurate. The "real world" part of the calculation, the original measurements, is where the error comes from.

The calculations for shooting rockets at other planets are really not that much more accurate than 99.9%. That is why all space flights have multiple mid-course corrections. Though the trajectory was calculated to a high degree of precision (which is not the same as accuracy) the actual launch can not match that level of precision or accuracy.

Even the trajectory of the planets is not know to all that great of precision. Astronomers are still refining the orbital trajectories of the planets. But, they've been working on it for hundreds of years so the accuracy is probably a little better than 99.9%.

Micrometers are not generally that much more accurate. They have high precision but relatively low dynamic range. Micrometers and the like have been around for a long time. But engineers still routinely only worked the three decimal places in their calculations (roughly 99.9% accuracy). To use a micrometer to better than 99.9% accuracy (not precision, but precision is misleading anyway) you need to start taking the temperature of the object and micrometer into acocunt.

I doubt UPC readers are anywhere near 99.9% accurate in the real world. For example, when was the last time you had an item at the grocery store that refused to be read by the UPC scanner? That is a good example of an error (inaccuracy) in the UPC reader system.(The US post office uses a UPC like labeling system and that system still gets mis-read and sends something like a couple percent of all letters to the wrong post office).

How many numbered pages are in a library? You don't think that one book in a thousand (0.1%) has a typo in the libraries database?

Ask any practicing scientist or engineer and they'll tell you that for the vast majority of things accuracy is rarely as good as 99.9%. If you want greater than 99.9% accuracy the cost goes up very rapidly.

Airplane safety... yes more than 99.9% of flights end up landing safely. But, I would venture a guess that more than 1% of all flights have a safety related mechanical problem during the flight. (Any snopsters work in aircraft maintenance?) I've been on less than 100 flights and been in one minor crash. So for me, the "didn't crash" percentage is less than 99% (let alone 99.9%).

Heck, the state of Florida had a hell of a time counting votes, even with computers, and even after doing the count several times. Every count came out different. The final margin was, from a statistical perspective, a tie. The difference of ~500 votes between Bush and Gore was not statistically significant since the votes could not be counted to that level of accuracy (99.99%).

So, in the real world, very few things are 99.9% accurate. Heck, even for something as simple as "how long is this piece of 2x4 wood", if you want the length accurate to 99.9% you will also need to specify the relative humidity and temperature of the wood since changes in either of those parameters can change the length enough to make a 99.9% accurate length impossible.

You might want to read up on the difference between "accuracy" and "precision". Accuracy is what you care about, but in most cases the accuracy is unknowable. Precision can be observed and measured but there is no direct relationship between precision and accuracy. The only relationship is that a measurement can be no more accurate than it is precise. However, a measurement can also be much less accurate than it is precise.

Posts: 629 | From: Greenwood, IN | Registered: Dec 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Aptenodytes_Forsteriis:
quote:
Originally posted by Ganzfeld:
quote:
Originally posted by jimmy101:
Few things are 99.99% accurate.

What makes you say so? My computer is much more accurate than that. And the calculations for shooting rockets at other planets are much more accurate and the motions of all the planets themselves conform very accurately to the same calculations, and the measurement of a micrometer can be much more accurate, and UPC barcode readers are more accurate, and the numbered pages in all of the books in the library are more accurate, and aircraft safety is more accurate... (ad infinitum) What makes you think that this is so rare?
Doesn't it depend on the scale at which you examine these things? [...]
Exactly.
Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by jimmy101:
The calculations for shooting rockets at other planets are really not that much more accurate than 99.9%. That is why all space flights have multiple mid-course corrections. Though the trajectory was calculated to a high degree of precision (which is not the same as accuracy) the actual launch can not match that level of precision or accuracy.

Even the trajectory of the planets is not know to all that great of precision. Astronomers are still refining the orbital trajectories of the planets. But, they've been working on it for hundreds of years so the accuracy is probably a little better than 99.9%..

Where do you get your figures from? I don't believe these two points. 0.01% error for the initial trajectory on a mission to Jupiter could never be corrected.

I'd say the accuracy for orbital traqjectories is much much more accurate than 99.99% for planets, not just a little. Even for smaller objects with some uncertainty in measurements, etc, trajectories can be accurately predicted within, for example, half the distance between the Earth and the Moon many years in advance, which -- on the scale of the Solar System and for the next few months -- is still much much better than 99.999%. As I pointed out above and Aptenodytes pointed out again, it all depends how you measure "accuracy". (Yes, I know the difference between accuracy and precision. It doesn't matter. Both can be measured in extremely different ways.)

Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Doug4.7
Angels Wii Have Heard on High


Icon 1 posted      Profile for Doug4.7   E-mail Doug4.7   Send new private message       Edit/Delete post   Reply with quote 
Part of the issue with planetary orbits is they have to use relativity to get it that accurate. In fact, that was one of the areas that "hinted" towards relativity because the math was getting different answers than the planets.

As for rockets, you can live with quite high errors because you have the mid-course corrections and you build in buffers to the course. They've had more than one shuttle flight where the SRBs did not last as long as they should have. They had enough buffer to still get it into orbit.

--------------------
And now for something completely different...

Posts: 4164 | From: Alabama | Registered: Oct 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
quote:
Originally posted by Doug4.7:
As for rockets, you can live with quite high errors because you have the mid-course corrections and you build in buffers to the course. They've had more than one shuttle flight where the SRBs did not last as long as they should have. They had enough buffer to still get it into orbit.

Just getting into orbit and getting to a distant point in the solar system are very different. And compensation very early in a mission is what leads to accuracies of greater than 99.99%. So, again, it all depends how (or, in this case when, you measure it).

(late ETA - I have to take also issue with your representation of relativity as essential for accurately predicting the planet's paths. Again, depending on how far in the future you want to compute, I submit that, for all planets but Mercury, accuracies of 99.99% are perfectly reasonable without considering relativity for many years in advance. Maybe one of our astrophysics experts can correct me if I'm wrong.)

(ETA 2 -- I have read now that the accuracy even of Mercury is well within the 99.99% range for a few years in advance. That is, the error is 43/5600 arc seconds per century, which is slightly better than 99.99% accuracy. The paths of the other planets are far less influenced by the effects of relativity. So, no, relativity is not required for that kind of accuracy for any planet.)

Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Troberg
Angels Wii Have Heard on High


Icon 1 posted      Profile for Troberg     Send new private message       Edit/Delete post   Reply with quote 
quote:
I'd say the accuracy for orbital traqjectories is much much more accurate than 99.99% for planets, not just a little.
I'd say that accuracy as a percentage on an orbital trajectory does not make sense. Percentage of what? How can you have a percentage on a position when the reference point is not defined?

--------------------
/Troberg

Posts: 4360 | From: Borlänge, Sweden | Registered: Nov 2005  |  IP: Logged | Report this post to a moderator
Ganzfeld
Let There Be PCs on Earth


Icon 1 posted      Profile for Ganzfeld     Send new private message       Edit/Delete post   Reply with quote 
Well, that's the main problem, isn't it. Accuracy as a percentage doesn't make sense for most of these stats.
Posts: 4922 | From: Kyoto, Japan | Registered: Sep 2005  |  IP: Logged | Report this post to a moderator
Darth Credence
Deck the Malls


Icon 1 posted      Profile for Darth Credence     Send new private message       Edit/Delete post   Reply with quote 
The place I've seen this is in training for Six Sigma classes. They throw this up to show why Six Sigma is important, because it leads to better control of processes than 99.9%. Many manufacturing processes today are held to a six sigma standard. (This means that the specified limits of the process are six standard deviations from the mean of the process.) If something is processed with a six sigma process, then there are only 2 parts per billion that are out of spec - or 3.4 parts per million if a 1.5 sigma shift is involved. That would be 99.9999998% parts passing specifications.
Where I work, most of our processes are around 6.8 parts per million (ppm) defective. A few are better, a few are worse, but we are trying to improve them all.
When you are talking about things like the length of a 2x4 board or the position of Mercury it is not really relevent to the initial post. The idea behind the original post is that when you are making or processing something, if the output only passes spec 99.9% of the time, there will be problems. The vote one is good, and would belong with the rest of the list as we have a spec of counting correctly and the out of spec condition of wrongly counted or not counted at all. Measuring a board yourself doesn't work as you don't have a spec - the maker of the boards do, and therefore a 2x4 that is 8 ft +/- 1 inch would have 1 board in 1000 be over 8'1" or less than 7'11".
Oh, and there are stats on how accurate a person can be. A person performing a repetitive function, such as data entry or hand cutting boards, will on average only be 95% effective. So if a person is told to type a list of numbers into a computer, and that list has 1 million distinct numbers, that person will, on average, make a mistake on 50,000 entries.

--------------------
Give yourself to the Dark Side. It is the only way you can save your friends.

Posts: 262 | From: Salt Lake City, UT | Registered: Oct 2005  |  IP: Logged | Report this post to a moderator
   

Quick Reply
Message:

HTML is enabled.
UBB Code™ is enabled.

Instant Graemlins
   


Post new topic  Post a reply Close topic   Feature Topic   Move Topic   Delete topic next oldest topic   next newest topic
 - Printer-friendly view of this topic
Hop To:


Urban Legends Reference Pages

Powered by Infopop Corporation
UBB.classic™ 6.7.2