Boeing ceases publication of annual safety summary?

[paulross]

Boeing has published the "Statistical Summary of Commercial Jet Airplane Accidents" annually going back to 1959. Normally this comes out around July, sometimes as late as October. In 2019 it appears there is no publication. A search of boeing.com shows the latest to be 1959-2017 (https://www.boeing.com/resources/boeingdotcom/company/about_bca/pdf/statsum.pdf) which was published 15 months ago. E-mail to the publisher, [email protected], goes unanswered.

IMHO It is be a shame if this publication, covering the whole commercial jet era, has ended.

[DaveReidUK]

The 1959-2017 was indeed published in October 2018. I suspect that in October 2019 Boeing had other priorities.

[pilotmike]

Wot? Priorities other than safety related matters? PR and spin, possibly? I guess you're right.

[FlightDetent]

I told you so.

[lomapaseo]

Not ended probably just changed to a quarter century periodical of data

[airman1900]

Boeing has finally placed/published its Statistical Summary of Commercial Jet Airplane Accidents Worldwide Operations | 1959 – 2018 50th Edition on its website.

Statistical Summary of Commercial Jet Airplane Accidents 1959 – 2018 (https://www.boeing.com/resources/boeingdotcom/company/about_bca/pdf/statsum.pdf)

Found at:

Boeing's Aviation Safety page (https://www.boeing.com/company/about-bca/aviation-safety.page)

[procede]

Thanks for posting! I think a lot of discussion went on about the graph in page 13, especially the last aircraft type...

[SteinarN]

And next year that graph will go off the scale at a value of around 13...

[procede]

The accidents wil go up by 1, but I think the number of flights increased at a higher rate, as there were more MAXes flying..

[paulross]

Thanks for posting. Publication date September 2019.

[DaveReidUK]

Though the PDF has a "last modified" date of just a couple of weeks ago.

Stats are as of the end of December 2018.

[paulross]

Indeed, I see:
Creation date: 1 Oct 2019 at 18:16
Modification date: 12 Feb 2020 at 18:04

[SteinarN]

procede

The number of accidents will go up by 1 as you say. But the number of flights will not increase that much as the Max was flying for only eleven weeks in 2019.

But you are right, the accident rate can not double, it must be less than a doubling, so maybe around 7 or 8 per million flights. Almost the same as the total hull loss bar for the "Aircraft no longer in service" and the 707/727.

[greeners]

ICreation date: 1 Oct 2019 at 18:16
Modification date: 12 Feb 2020 at 18:04

Interesting. I have been actively searching for this doc since last October - this is the first time I've seen it, despite having been on the Boeing Safety Page many times.

[DaveReidUK]

The number of accidents will go up by 1 as you say. But the number of flights will not increase that much as the Max was flying for only eleven weeks in 2019.

But you are right, the accident rate can not double, it must be less than a doubling, so maybe around 7 or 8 per million flights. Almost the same as the total hull loss bar for the "Aircraft no longer in service" and the 707/727.

The Boeing StatSum shows just over 200,000 Max flights as of the end of 2018 (1 million divided by 4.90). At the time of the grounding, FR24 estimated that there were 8,600 departures per week being flown by the Max - so around 95,000 for the period Jan-Mar 2019.

So roughly 6.7 hull losses per million departures - not that that tells us anything we didn't know already.

[safetypee]

The change in the summary format, as with the Airbus publication, represents the decreasing usefulness of previous approaches.

Many safety management systems focus on fatalities, but in an industry with few fatal accidents this view has little or no value in judging action to reduce future accidents.

Historical risk (outcome - fatalities - probability) can be quantified, although potentially biased by allocation of cause or contributing factors.

Future risk is a judgement - expert guess of what might happen, based on scant historical risk. This approach is open to flawed thinking - that the future will be the same as the past - same number of fatalities, passenger load vs freighter for same type, or not considering intervening aircraft modification or crew training.

An example is with LoC, often cited as the dominant safety issue. This might be true according to a 20 year history, but comparing different generations of aircraft, then the latter are significantly safer (Airbus review).

An alternative approach to safety could minimise cost and effort on training for every type of aircraft by focusing more on older aircraft. Also, a closer study might identify significant differences so that more targeted training can be given; e.g. instead of requiring ill-specified manual flight for everyone, focus on flight without auto-throttle or without FD to encourage instrument scanning, speed awareness, for some types, could help avoid a particular aspect of the LoC threat, which may be well-protected in modern types.

https://www.airbus.com/content/dam/corporate-topics/publications/safety-first/Statistical-Analysis-of-Comercial-Aviation-Accidents-1958-2019.pdf

[infrequentflyer789]

An alternative approach to safety could minimise cost and effort on training for every type of aircraft by focusing more on older aircraft.

Sounds awfully like this would resolve to "need more training for older aircraft" ?

Next $1M question is of course, "what is an 'older aircraft'", 737-classic, 737-NG, 737-MAX, A320, A320neo ???

Think this will lead right back to the same problems it appears the industry has at the moment, with mfrs fighting regulators to reduce training requirements (as they do with certification) so their planes have a competitive total cost.

[safetypee]

iff789, 'more or less training' is a difficult subject.

Re aircraft type divisions; that used by Airbus could be justified by the safety data, however I suspect that there would be many objections.

An alternative is to consider less training for modern aircraft - actually focussed training according to the design, vs existing training for the older types. However, some data suggest that exiting training is not providing the expected results and could benefit from a new approach; e.g. number of approach to stall incidents which could progress to LoC (depending on event definitions) requires more attention on airspeed - awareness and control. Also focus on avoidance might be more effective than training for recovery - avoid and mitigate the threat - thus better use of training time.

Much of this thinking presumes that the modern designs benefit safety; if so - most likely, then the task would be to identify those aspects which technology provides in new designs and consider the extent of human intervention which could help in older aircraft, but this is being questioned after several accidents where the older designs had assumed too much. The alternative is to retrofit appropriate (low cost) technology to aid the human; e.g. conceptually like EGPWS.

[alf5071h]

'not more training; more focussed training'
Recent studies on startle factor considered what aspects might be trained, this also identified areas where existing training could detract from the safety objective.
The industry could better use training time by identifying tasks which no longer apply in modern operations, or the techniques which lead to negative training. Not more training, but eliminate the unnecessary tasks and re-evaluate how training is carried out.

Regulators tend to focus on more training regulations; they need a department for removing regulations, as assessed by safety statistics - but which data to use and with what view of safety in a very 'safe' industry.

[Aero13]

Boeing has published the 2019 Statsum, so the sky isn't falling, yet.

[paulross]

The 1959-2019 publication is here: https://www.boeing.com/resources/boeingdotcom/company/about_bca/pdf/statsum.pdf
Creation timestamp: 2020-12-21 19:23
Modification timestamp: 2020-12-21 20:36