I'm with you 100% on this issue.
Hatcher is the authority on this issue, not de Haas.
And low numbered 03 are
still coming apart when the threshold of safety is exceeded. What's the distance between safe and exceeded? I never heard of one blowing up on a wednesday so I'm only going to shoot mine on a wednesday so I'll be safe. The logic of Darwin candidates is heavy in this forum on this subject.
http://m1903.com/03rcvrfail/
The response to the above by a friend or mine:
"His article is a perfect example of making statistics tell the story you
want them to tell. While I'm not a statistician, I am a mechanical
engineer [NASA] who helps run a hazardous research and development test
operation. I not only have to know statistics but the limitations of
statistics. Often, it's my *** that's going to perforated by high speed
metal chunks if I get it wrong.
In his article, the good doctor shows that he has a good grasp of
statistics, but a poor understanding of the limitations of statistics.
In the article, he takes the known historical data and shows that based
on this data, firing a low-number M1903 is safer than a lot of common
activities. Absolutely true, based on the data at hand.
But here's the limitation - While his conclusion is true for the whole,
his conclusion is faulty for a particular single rifle.
Here's the problem - we don't have enough statistical data to determine
the distribution of M1903 low-number receiver strength. Looking at a
particular specimen, you cannot tell if it's better or worse than
average. Nor can you tell how much better or worse than it is from that
average.
Given the poor process control that Hatcher documented (heat treatment
by eye), I'd say that process variability is quite high, meaning that
there will be many guns that are much better than average. Conversely,
there will be many guns that are much worse than average.
Even if we knew the distribution, we don't know any fatigue behavior: we
don't know the relationship between heat-treatment-related-strength and
how many rounds of a known pressure that receiver design will take
before catastrophic failure at that strength. Could be one. Could be
one thousand. Could be infinite.
So it boils down to this - while on average, firing an M1903 is safer
than some average daily activities, firing a specific M1903 may be far
safer or far less safe. NO ONE CAN TELL!
While the doctor says I may have a 1 in 100,000 chance of one blowing
up, if that one in 100,000 happens to be the one in front of my face, I
have a 1 in 1 chance of getting hurt or killed.
Here's the other little tidbit I'll toss in there that the doctor
doesn't address. Back in Hatcher's time, the Government's assessment of
the worth of a soldier's life was pretty low. I'd bet that back then, a
soldier's life was viewed to be less than that of a rifle. (see note
below) So any sort of judgment of past cost versus benefit (IE:
scrapping rifles versus potential soldier death) must be looked at
through period assessment of soldier life. Or, better said, it's NOT
that the Marine powers that be thought that the chances of failure were
low, it's that they didn't see the
*consequences* (cost of soldier
death) of failure being high, so overall, the risk was acceptably low.
Keep the rifles. Replace the dead soldier as they fail."
by B.H. 1-11-09