Red and Orange Line riders: Remain calm
By adamg - 9/23/09 - 7:09 am
The National Transportation Safety Administration is urging transit agencies to check their signal systems - including the ones in use on the Red and Orange lines, for possible problems following a subway crash in Washington. The feds say that, under certain circumstances, the systems might give a train driver a phantom all-clear signal. The T says not here; that the system has worked for decades without a hitch.
Comments
Statistically, couldn't that
Statistically, couldn't that also mean we're overdue?
No
The probability of an event doesn't change just because that event has not happened. In and of itself, the fact that an event has not happened yet does not change the probability that it will happen at any given time in the near future - unless the event having happened leads to changes that alter that probability.
So, there "not being an accident" does not increase the chance of an accident. Having an accident may reduce the chance of future accidents if there are changes in the system (more inspections, installation of fail-safes, etc.).
The exception would be if there were no maintenance and we are calculating a "time to fail", where fail is inevitable. Then the probability of failure can rise or change over time. The ultimate expression of this is the mortality curve for a population - in a closed population, everybody eventually dies and the risk of death changes over time.
In my best Harold Hill
In my best Harold Hill impression, "Here we have a piling up of words, words, words, all in service to a New World Order signaled by the airtight packaging of the Uneeda Biscuit."
So, I sorta got a headache trying to grasp the concept, but ... because it hasn't failed, it won't ever fail?
I'm inclined to be influenced by the MBTA's track record on other things rather than rely on Aristotlean concepts of probable impossibilities versus improbable possibilities.
Ugh, too many big words. I'll stick with the music gig!
Hey - isn't "words, words,
Hey - isn't "words, words, words" Eliza Doolittle, rather than Harold Hill? ;-) (And, "Uneeda Biscuit" isn't him either, but at least it's the right show, lol)
I can just only hope that none of us are anywhere near the red or orange lines when they do eventually crash. Until then, I guess we'll have to rely on probability theory - probably. ;-)
Meanwhile - can ANYONE do anything about the increasingly awful musicians who, um, "entertain" us at Downtown Crossing (heading to Forest Hills)? I've been tempted a few times lately to throw them some money on the condition thay they'd STOP...
Same error
It hasn't failed = won't ever failure
It hasn't failed = we are due for a failure
These are both horns of the same "not getting" what a statistical probability implies.
Not failing yet does not mean "more likely to fail now" unless the probability of failure changes.
Not failing yet, does not mean "less likely or not going to fail at all now", either.
The chance of failure at any given time is the same, unless certain circumstances modify the probabilities.
Things that can change the probability over time:
not failing early = less likely to fail later (good quality)
increasing probability of failure over time (mortality curve)
failure of other components or other systems results in maintenence or replacement or enhancements that reduce future failure probability
failure of maintenance programs or reduction of inspections makes failure more likely (think Bhopal)
These are alterations of the probability, though. If the probability is the same over time, you aren't "due" for a failure, not in any statistical sense.
"[T]he system has worked for decades without a hitch."
Really? They said that with a straight face?
Signal problems? On the T? Never!
I'm pretty sure that, had WMATA been tasked with a similar
request to check their signal systems a couple of months prior to their Red Line accident, they probably would have responsed with "Our system has worked for decades without a hitch" as well.
BTW, my first job out of college in the mid 1980s included some PR work. Had I issued a flippant statement like that in response to a potentially serious safety issue, I probably would have been censured at the least, and could have possibly been fired.
Link broken?
Adam, can you double check that link? It's sending me to the wrong place, methinks.
Sorry
Link is fixed.
Thanks
I guess what I was really curious to learn more about were the specifics of the T's response. Did they really make that claim?
Sheesh!
Decades without a hitch!
Maybe without that particular hitch, but certainly with many, many other hitches.