Virgin Blue's new low: 2100 flight cancellations

[gilby08]

BUDGET airline Virgin Blue holds the dubious record of having the highest number of flight cancellations in Australia last year, official figures show.
More than 5700 scheduled domestic flights were cancelled last year, with Sir Richard Branson's budget airline accounting for more than 2100.
But it was fellow budget carrier Tiger Airways that recorded the lowest number of punctual departures and arrivals.
The revelation comes after Tiger was forced to apologise for the cancellation of a Melbourne-bound flight that left more than 30 passengers stranded overnight at Launceston Airport last weekend.
The flight had been scheduled to leave at 5.55pm on Sunday, but was repeatedly delayed until it was eventually cancelled.
Bureau of Infrastructure, Transport and Regional Economics data shows cancellations were highest each way on the country's busiest route - Melbourne to Sydney - with 3.8 per cent of flights cancelled in either direction.

etc etc etc, read more from this wonderful news agency: Virgin Blue's new low: 2100 flight cancellations | News.com.au

Shame they haven't opened comments yet!

 

[Mal]

Perhaps people should read http://www.bitre.gov.au/publications/17/Files/BITRE OTP Report 2009.pdf for a more complete picture.

Yes, DJ did have the most cancellations .. but Qantas was very close behind at 1.4%

What does surprise me is that DJ is now the number 1 carrier (by number of flights) in Australia. (when stating Qantas / Qantaslink / Jetstar separately rather than together - combine Qantaslink and Qantas, and it is a different story)

Selectiveness also counts when comparing figures. If you look at BNE-SYD, both Jetstar and Qantas cancelled more flights than DJ did.

And the highest % of cancelled flights on a route was done by Skywest (Kalgoorlie - Perth @ 6.6% of flights). Qantaslink was close behind MEL-CBR//CBR-MEL with > 6%.

Stats can be twisted any way you want them to appear ...

 

[markis10]

I am surprised, an aviation story that talks about more than one airline, and does not dwell on JQ

 

[JohnK]

Anyone can twist statistics around to suit them.

And the "news" article is not much better. Waste of time reading an article that selectively reports a few anomalies here and there.

 

[samh004]

I recall this reported a couple days ago with more replies on the subject, did the thread get deleted, or has it just been overlooked?

 

[anat0l]

I had to explain some stats to someone, so whilst I still had all the Excel and what not opened I thought, "what the hey, let's put the numbers through."

I did a quick test based on cancellations and on-time departure frequencies for QF (alone, not Link) and DJ (alone). Testing whether the numbers of cancellations and OTDs were significantly different (due to different numbers of flights for each), my quick test showed that they are very different. In other words, the %ge points may be slight, but there is a good difference in the numbers between QF (alone) and DJ (alone).

If one considers the QF group (QF, QFLink and JQ) and DJ group (DJ, ZL and XR), then the QF group actually has more cancellations than the DJ group (1.17% compared to 1.14%). However significance testing shows that this difference is insignificant. The QF group also has better OTD and OTA overall compared to DJ group. However, in testing the significance in the difference of OTD (QF grp 85.33% v DJ grp 85.02%) shows that this difference is actually significant.

Bottom line? QF did better. Very odd.

Happy to share my methods if requested.

 

[Flashback]

I had to explain some stats to someone, so whilst I still had all the Excel and what not opened I thought, "what the hey, let's put the numbers through."

I did a quick test based on cancellations and on-time departure frequencies for QF (alone, not Link) and DJ (alone). Testing whether the numbers of cancellations and OTDs were significantly different (due to different numbers of flights for each), my quick test showed that they are very different. In other words, the %ge points may be slight, but there is a good difference in the numbers between QF (alone) and DJ (alone).

If one considers the QF group (QF, QFLink and JQ) and DJ group (DJ, ZL and XR), then the QF group actually has more cancellations than the DJ group (1.17% compared to 1.14%). However significance testing shows that this difference is insignificant. The QF group also has better OTD and OTA overall compared to DJ group. However, in testing the significance in the difference of OTD (QF grp 85.33% v DJ grp 85.02%) shows that this difference is actually significant.

Bottom line? QF did better. Very odd.

Happy to share my methods if requested.

Seeing you say significance testing takes me back to Uni days.... shoo you!

Reminds me of a "PEAC" course I did at primary school called "Lies, Damn Lies, and Statistics!". Is so true.

 

[anat0l]

Seeing you say significance testing takes me back to Uni days.... shoo you!

Reminds me of a "PEAC" course I did at primary school called "Lies, Damn Lies, and Statistics!". Is so true.

Well it's still better than quoting the average alone That is one of the easiest "stats" to twist.

But apart from that, I agree. You don't see every news article linking to a copy of the BITRE report.

 

[dajop]

However, in testing the significance in the difference of OTD (QF grp 85.33% v DJ grp 85.02%) shows that this difference is actually significant..

Just because their is statistical difference between two results, does not necessarily indicate that the difference is meaningful (in my line of work oft quoted phrase is "statistically different, but not biologically meaningful"). .

Operationally to an airline there may be a meaningful difference over the 000's of flights. The average (even frequently travelling) passenger, a difference between 85.33% and 85.02% would not be meaningful or even be detectable, whereas individual experiences would be.

 

[anat0l]

Just because their is statistical difference between two results, does not necessarily indicate that the difference is meaningful (in my line of work oft quoted phrase is "statistically different, but not biologically meaningful"). .

Operationally to an airline there may be a meaningful difference over the 000's of flights. The average (even frequently travelling) passenger, a difference between 85.33% and 85.02% would not be meaningful or even be detectable, whereas individual experiences would be.

All you say is very true.

As everyone else has pointed out, statistics are a means to support an answer. They aren't an answer in themselves (although pardon the hasty conclusion I've made in my post).

I guess my short analysis was more to answer the differences in figures and provide a bit more numerical context to the article, which is typical of news articles that short sightedly only rely on averages without considering equally (if not more) important contextual aspects: variance, median, etc.

 

[ANstar]

Presumably figures will be high for last year - especially considering the amount of flights both QF and DJ would have had to cancel over the day (and day after) of the dust storms in Sydney last year.

 

[Harry G]

its great we look in depth at the QF/JQ/DJ figures

perhaps a comment about Tiger.......
fly the least sectors with the least number of planes and they stilll cannot get it right!

the media.........and their reporting!!!!!! "DJ" the worst????????????

 

[drron]

I am surprised, an aviation story that talks about more than one airline, and does not dwell on JQ

They didnt let you down-here is today's Nonews article-
Tiger Airways the Aussie airline with the latest planes | Courier Mail

But Qantas subsidiary Jetstar had the highest percentage of cancellations for February at 1.6 per cent, while Regional Express had the lowest at just 0.1 per cent.

 

[JohnK]

Bottom line? QF did better. Very odd.

Happy to share my methods if requested.

Significance testing?

Statistics don't mean much unless applied to our own individual experiences....

 

[Flashback]

Significance testing?

Statistics don't mean much unless applied to our own individual experiences....

Please don't bring back bad uni memories

 

[Mal]

Please don't bring back bad uni memories

lol. Are there really any "bad" uni memories???

Let's see:
- Failing a subject, but emailing the co-ordinator and getting him to push me to a pass.
- Failing another subject (which was critical to my degree), and going through the scheme of "later examinations" and then passing (Note this one did blight my record by recording a "conceded pass" rather than a standard pass.
- Meeting a decent number of young ladies who ....

I think the worst uni memories I had were trying to finance uni textbooks etc, and the small wait for results. Then again, I did have issues with completing some of the tasks involved... and lots of stress etc. But that was nothing compared to the "real world" once I left.

 

[oz_mark]

- Failing another subject (which was critical to my degree), and going through the scheme of "later examinations" and then passing (Note this one did blight my record by recording a "conceded pass" rather than a standard pass.
- Meeting a decent number of young ladies who ....

I think the worst uni memories I had were trying to finance uni textbooks etc, and the small wait for results. Then again, I did have issues with completing some of the tasks involved... and lots of stress etc. But that was nothing compared to the "real world" once I left.

Nothing wrong with 'conceded pass'. I have two of them....

The financial bit was the worst of it - somehow I wonder how I managed to do it.