Talk:Failure rate

Hazard rate
We should make Failure rate a disambiguation page and move this page to hazard rate. Failure rate is often used confusingly (but not strictly speaking incorrectly!) to refer to either (a) hazard rate or (b) rate of occurrence of failures (ROCoF). The former belongs to parts reliability, the latter to repairable systems reliability. This article is about (a) but might propagate the confusion. Repairable Systems Reliability by Ascher & Feingold puts the matter clearly.

The article also duplicates a lot of stuff that is in survival analysis. I suggest taking the detailed stuff out of survival analysis and putting it into the hazard rate article.Cutler 15:37, Mar 26, 2005 (UTC)

There are several related pages that need to be better organized and cross-referenced: Failure rate, MTBF, Survival analysis, Bathtub curve, Infant mortality and probably some others. I plan to work on this in the near future, starting with merging MTBF with this article. Any comments on the best way to reorganize these pages would be appreciated. Also - nice to see that this page was referenced by the media!

Hazard rate already redirects to this article. DFH 19:02:27, 2005-09-01 (UTC)

As someone who was looking for a widely applicable definition of something called "Failure rate" I found this appropriate to my needs. So, although I took the opportunity to make a few changes for readability/clarity, I would make a strong please for it to stay where it is. (It still needs work done on the references. I don't have time to fix them today.) PeterDz (talk) 16:53, 24 September 2010 (UTC)

This article maintains that there is definitely a difference between failure rate and hazard rate, however, IEC 60050 part 192 (http://www.electropedia.org/iev/iev.nsf/display?openform&ievref=192-05-06) says that failure rate and hazard rate are the same thing and mathematically defines failure rate the same way hazard rate is defined in this article. I think this article is adding to the confusion on this point, not clarifying anything. --Spuzzdawg (talk) 01:26, 31 May 2015 (UTC)

Jargon?
what does "projective manifestations" mean?

Delete this remark when the sentence is fixed.

Matt Whyndham 15:33, 31 May 2007 (UTC)

Fixed it; if others agree with the fixes then please delete all of this. PeterDz (talk) 17:00, 24 September 2010 (UTC)

Hazard rate again
Can we please rename this article "hazard rate" and make failure rate redirect to hazard rate? The books never call it a failure rate.

Ogo (talk) 19:07, 24 June 2009 (UTC)


 * I agree. 'Failure Rate' is never mentioned in any of my textbooks; it's always called Hazard. Paul

Hazard = mean occurence frequency?
Does all that the article explain also mean that the mean frequency of occurence approximates the Hazard? In a neuron that fires 10 spikes per second on average, the Hazard would be 10Hz? Paul

Units - FIT
The given explanation is some how unclear: "The relationship of FIT to MTBF may be expressed as: MTBF = 1,000,000,000 x 1/FIT." To my knowlage this only works for systems with a constant failure rate, however the MTBF is gerally defined for all kind of statistics. The unit of the MTBF may vary. Most of the times it is per year, and I assume that this version is used here, however one should mention the unit used for the MTBF when giving an relation to the FIT value. 194.138.39.52 (talk) 08:39, 26 January 2012 (UTC)


 * "The relationship of FIT to MTBF may be expressed as: MTBF = 1,000,000,000 x 1/FIT."
 * I'm a bit puzzled with the term 10e9 in this equation. IMHO it should be only "MTBF = 1 / LFR" because LFR in FIT already contains "10e-9". In other words, if we have a LFR of 5 FIT it corresponds to MTBF of 2e8 hours and not 2e17 hours, what above equation would give. — Preceding unsigned comment added by 213.164.91.2 (talk) 10:14, 14 October 2015 (UTC)

What is F(t)?
In section "Failure rate in the discrete sense" you say that) R(t) = 1 - F(t), but what is F(t)? It has not been mentioned before that point — Preceding unsigned comment added by Yannis1962 (talk • contribs) 11:49, 25 December 2013 (UTC)

Going from the continuous case, $$F(t) = \sum_{\tau=0}^{t}f(\tau)$$ which is the cumulative distribution of $$f(t)$$ Drevicko (talk) 10:19, 19 July 2015 (UTC)

The section Discrete time is actually continuous time
In discrete time, the variable i should be used and lambda(i)=f(i)/R(i-1).Per W (talk) 17:36, 25 March 2015 (UTC)

Hazard rate and ROCOF
Why are they not the same as the failure rate? Per W (talk) 23:11, 26 April 2015 (UTC)

Personal Computing: Where's MTBF gone? -> Evolution of MTBF in computing over time.
Computers have become pervasive, microprocessors ubitiquous.

As student I worked on a Zuse 23 computer, based on tubes and mechanical devices. Quizz: The world's 1st binary computer?

Later I visited the computing center of a university. It sported the worlds 1st fully transistorized large scale compter. aThey said it was the most performant, being statistically non-working 29 days per month, they said, but faster than ...

The common concern was MTBF.

Today, 60 yrs later: I have a 10 yr old laptop: 8 GB Mem, 200 GB HD: 1st pb: the CD drive quit, but no need any more: now we download apps, have 1TB mSD, etc. I'll toss my PC and have already bought a new one without HD - simply plenty of SSD ...

How about my faithful companion MTBF? When did it become irrelevant?

MTBF is like en ex, it left me, don't miss it. Just wondering where it's gone...

Any graphs available? 2001:9E8:8ACF:AE00:74F0:6592:1246:600D (talk) 00:01, 7 January 2024 (UTC)

Conflicting statements about discrete hazard function
It's stated under "Failure rate in the discrete sense" that:

"Although the failure rate,, is often thought of as the probability that a failure occurs in a specified interval given no failure before time , it is not actually a probability because it can exceed 1."

but later it states:

"...Note that this is a conditional probability, where the condition is that no failure has occurred before time t." Kai (talk) 00:56, 31 March 2024 (UTC)