The Sig P365 and an important lesson in understanding stats/critical thinking. I am not here to bash the P365, I actually want one. I have watched the issues that have been encountered and realized this could be a great chance to explain an important critical thinking concept.

**Don't leave yet!!!** Stats get manipulated all the time and I want to show you all how to see through that a little more.

So there have been issues with the striker and trigger return spring breaking. According to Phil Strader the return rate for both issues combined is 0.25%. Really not bad right? The lesson...

https://www.thefirearmblog.com/blog/2018/06/06/p365-strikers/There is a CONDITION that needs be considered. Most of these P365's tested on the Youtube seem to be ran to 1,000-1,999 or 2,000+ round counts. How many people are really shooting their P365's that much? Thereby is that return rate really representative if I want to shoot my P365 a lot? So what is the probability that a P365 will fail GIVEN that it fires 1,000-1,999 rounds or 2,000+ rounds?

This is called CONDITIONAL PROBABILITY. The math is easy. The main lesson in critical thinking is that people, orgs, businesses, politicians, etc. often try and use a stat that is applied to a group broadly, when there really need to be conditions considered that shrink the group size down and change the story. Here's how it can be applied to the Sig P365.

I can't find sales numbers on the P365 so let's just pretend Sig has sold 300,000 P365's as of this time. Here's how that plays out:

Sig P365's sold: 300,000

Failure Rate (% of pistols returned for return springs and firing pins): 0.25%

# of Pistols Returned: 750

BUT, I want to know what is the probability of my P365 failing if I shoot 1,000-1,999 rounds or 2,000+ rounds. This is where we get hypothetical because neither we nor Sig know the actual number of P365's in the wild that have shot that many rounds. We are going to make some guesses starting with percentage estimates.

P365's with 1,000-1,999 rounds fired: 2.25%

P365's with 1,000-1,999 rounds fired: 10,000 = 300,000 * 2.25%

P365's with >=2,000 rounds fired: 0.67%

P365's with >=2,000 rounds fired: 2,000 = 300,000 * 0.67%

THE FINAL STEP:

Now the lesson on thinking critically. Phil is not being deceiving in sharing the return rate of 0.25% for these two issues. It is the honest rate. BUT we want to know this based on the condition of higher usage. Now we have our hypothetical numbers to use. With our pretend production number we found that 750 P365's were returned for these issues. Return rates based on usage can't exceed overall return rates. BUT what are the numbers of those 750 returned that had fired 1,000-1,999 or 2,000+ rounds? Nobody knows that for certain either... So we will just have to make some guesses again. This is easy to make in Excel and I can adjust numbers across the spectrum, but here's what I picked for the example. It could be higher, but it could be lower and I don't feel like researching and documenting every video and forum post to try and determine these numbers at the moment.

P365's that failed GIVEN they fired 1,000-1,999 rounds: 30% * 750 = 225

P365's that failed GIVEN they fired >=2,000 rounds: 50% * 750 = 375

CONDITIONAL PROBABILITY:I know... this takes a good deal of explaining even though the math is really pretty simple once you understand it. The total group considered is just much smaller for the percentage calculation (numerator/denominator).

P365's that failed GIVEN they fired 1,000 rounds: 225/750 = 2.25%

P365's that failed GIVEN they fired >=2,000 rounds: 375/750 = 18.75%

IF these numbers were correct you would have an 18.75% chance that your Sig P365 would fail if you shot 2,000 or more rounds. That isn't super awesome... But if you keep it between 1,000-1,999 rounds the probability of it failing is only 2.25%.

To be fair to Sig, they would have to test thousands and thousands of their P365's up into the 1,000-1,999 rounds fired range to discover just 1 that had these problems (if normal failure rates played out along these pretend numbers). Now this could lead into testing, testing parts, batch test, and all sorts of other nerdy and fun QA stats stuff. Think Six Sigma, Lean Manufacturing, and all those processes that have improved manufacturing immensely, but that is not what this post is meant to address.

The Final Take Away:We don't have the numbers to get accurate probabilities, but this concept of Conditional Probability is likely why people are scared of the P365 failing more than it actually does. BUT the stats from Phil and Sig swing a little too far the other way and likely make the failure rates look lower than they really are when considering round counts. Ultimately, even with these hypothetical numbers the rates are still really pretty low and I think you should feel confident in getting a P365 (especially because their service has been incredible!). Now if Sig could just use these concepts and real numbers then they could tell us round counts that would elicit needing to replace the striker. This may make people feel a lot more comfortable with their purchases... We love CZ for their incredible service and I think it is fair to embrace another company who is behaving in a similar way. Hopefully this illustrated some ways to think critically, see through some fog, and find some balance.

PS - Phil Strader, if you read this and want some help on using these concepts to ease people's concerns just let me know and I would happily help. Or just credit me when you ask you engineers and marketing teams to figure this out for any sort of press release. You could also send me a P365 and anything else awesome... Haha pipe dreams eh?