View Single Post
 
Old 09-12-2018, 06:31 PM
ggibson511960 ggibson511960 is offline
Member
 
Join Date: Nov 2007
Location: Houston, Texas
Posts: 975
Likes: 1,116
Liked 1,237 Times in 532 Posts
Default Statistical Reliability Comparisons

Quote:
Originally Posted by Practical View Post
No this is not spam.

I was hoping for a more 'engineering' based answer. I am not a mechanical engineer, but have electrical and software engineering experience. My knowledge of mechanical reliability is basically none.

The comments here agree with my experience. I have found Glocks, Pre-lock J-Frames and N Frame competition models with Locks to be extremely reliable. I have only had ONE J-frame fail due to a broken part and I have had my glocks fail in competition due to me breaking off a adjustible rear site.

However, other brands and firearm types have failed me right out of the box or shortly after. Colt's including double action revolvers and 1911s, Kahr's, and a Beretta.

I tend to watch reviews closely and am disheartened when I see brand new guns failing regularly or having minor issues that preclude their use in competition or self defense. The issue of the GP100 7 shot is an example of why I don't think most new revolver's can be trusted based on the engineering processes that bring them to us.

I am not saying all the guns today are bad, I am saying SOME of the processes used to make them are NOT suitable for creating a competition gun or self defense product. I think MOST revolvers today are relegated to the 'consumer' side of the house where the assumption is that these guns may require a customer service call at a rate higher than the Semi-Autos used in LEO.

For the record, I am a mechanical engineer, albeit not a specialist in mechanisms, but a firearms enthusiast, nonetheless, familiar and experienced in many of these issues of comparative reliability and durability. All of the recounts on this thread are anecdotal with zero control. I will confess to say that it is intuitively obvious that the simpler revolver mechanism is inherently more reliable in most situations, but that is my subjective opinion. This argument can never be resolved, for there are many classes of comparison. The one I see neglected is how reliable is the firearm to be drawn and fired six or eight times without reloading and not stop, the charging bear scenario. I wish somebody would sponsor me to evaluate this scenario. For most of us civilians I would assert that this case is the one that matters. We are not fighting a war (yet), battling street criminals (yet), or firing thousands of rounds in a competition without cleaning. Too many of these reliability assertions are predicated on one of these scenarios. Militaries and police forces need high capacity repeaters, which overwhelmingly bends the preference to semi-automatics.

Here's my controlled test. Pick a revolver and semi-automatic from similar price points (very important) with similar downrange ballistics from a similar caliber (no .22's against .45's). Procure five samples of each from a retail outlet, pre-qualify each break in each piece with ammunition that each will feed properly (important for the semi). Then bang away in an environment without variables that influence outcome like snow or mud. Five shots, reload, clean up with a simple wipe down. Repeat until a statistically significant number of failures occur, but a minimum of 1,000 rounds from each pistol, a total of 10,000 rounds. Count all the failures including stupid human mistakes like hitting a mag release in the middle of a string. From this data one can easily calculate the mean number of shots between failure. If you think this scenario is biased in favor of revolvers, I guess it is because there is virtually nothing that can go wrong, especially if you decided the scenario was to be single action only. Complexity is the enemy of reliability. Maybe double action only for the revolver is more valid if you define the scenario to be reasonably rapid fire, i.e. at the charging bear.

Any scenario could be evaluated that would have validity if properly controlled to eliminate extraneous variables. This will likely never happen because manufacturers don't want hard data out there to be used against them in their competitors' ad campaigns.

In the end the real statistical differences in reliability are probably de minimus for we civilians, which leaves us choosing pistols for reasons of personal preference, familiarity, suitability for the mission, emotion or whatever. Those reasons are more valid and pertinent than hard statistics that show up beyond the second decimal place of standard deviation.
Reply With Quote