[IanFromHKG]

KarlMarx – 19/05/2014 15:43 GMT :

“I do think it is a case of being unlucky. There is no reason why your cases won’t be lost on the next sector either”

Oh, thanks! LOL

“in a big system like airport baggage processing, it is truly random thing and some people are going to travel for20 years without ever encountering a lost bag, whilst others may have runs like yours”

I do understand all this, and I acknowledged we have been incredibly unlucky, BUT to have that many experiences, in just four months, on ONE airline, and one airline alone…. And, more to the point, rubbish follow-up in each and every case. As rferguson pointed out recently (http://www.businesstraveller.com/discussion/topic/AA-s-proactive-Customer-Service.), it is not the problems that define an airline, it is how they are dealt with. BA have failed on just about every score, except eventually paying up for our costs in the case of lost bag #3.

By way of contrast – and to illustrate just how unlucky I am being with bags this year – last month the Memsahib and I flew CX to London, and my half-empty bag was crushed. Genuinely ended up smaller. Helpful CX agent (no queue) apologised profusely, took the details, offered a new bag on the spot (which I rejected) and based on our estimate of the bag’s cost, told us we could spend up to GBP200 on a new bag. Am I annoyed? Yes, somewhat. But what have I lost? 20 minutes at the baggage desk, half an hour in a department store. But I now have a new suitcase, and a warm fuzzy feeling that Cathay care when things go wrong and try to fix them (I should add that I only claimed the cost a few days ago and haven’t received the money yet – but I did get a response within 24 hours, clearly crafted by a real human being and not an automated response, saying they would process the claim). What have I lost with BA? Well, see above…

[KarlMarx]

Ian from HKG

You make two points;

Firstly, I understand your frustration about losing your bags on BA, but it is truly random and I think irrational to blame BA for it occurring.

Secondly, with regard to poor customer service, that is under BA’s control – I take it that you will vote with your feet next time?

[nmh1204]

Thanks for clarifying, Ian.
You don’t seem to have much luck with luggage
Hope things get better for you

[IanFromHKG]

KarlMarx, in general we will try to, but unfortunately there are occasions when we have little choice. In this instance, the memsahib had a redemption ticket on CX but wanted to leave earlier to spend more time with the M-I-L, and the only available seating for a European flight was through CDG, and BA was the only feasible connection (and yes we did look at other carriers). Also the oneworld connections push us towards them sometimes. Having said that, for our next flights to GVA, the memsahib and offspring are flying SWISS. Even then, though, I am stuck with BA because of connection times

[DavidGordon10]

I think statements that this is random, or that it is not random, need to be made with care.

To test whether or not it is random, you would need a lot of data about the loss of bags and then to test whether the bag-loss-events deviated from the relevant probability distribution. I hope that someone in a back office at BA is doing this, so that management can search for a cause and correct things if the events are not random. Not many businesses, even those as large as BA, will actually do this.

The relevant probability distribution is the Poisson distribution and the Wikipedia article on this is good – http://en.wikipedia.org/wiki/Poisson_distribution

You will see from that article that one practical application of this distribution was to test the chances of being killed by the kick of a horse. I expect that Ian_from_HKG feels that the horse is kicking him more than it should.

[KarlMarx]

DavidGordon10 – 19/05/2014 23:47 GMT

With the thousands of bags that are carried and lost every day, it is clearly a normal distribution.

The Poisson distribution is not appropriate for this application.

[canucklad]

To TMConsulting, to answer your question about CDG and also to enter into the discussion about randomness, I believe there is a direct correlation between probability and the increased “chance” of a random event occurring.

Randomness suggests being luckily or in Ian’s case unluckily being selected in a normal, nothing out of the ordinary day to day event.

Which brings me to CDG to prove my point, the probability of a random event (losing your bag) increases dramatically when you transfer through an airport that advertises an easy transit time of 90 minutes. When in reality you probably need 3+ hours and that’s being kind.

[IanFromHKG]

In the Memsahib’s case, it was a four-hour layover.

A day has elapsed and BA still have no idea where the bag is…

[DavidGordon10]

Dear Dr Marx

Who am I to argue with someone with such an august handle … but.

If we look at the case of all the bags that are lost everywhere and every day, then the distribution is clearly normal. However, if we look at the number if bags lost on each individual flight, then (we hope) that the number will generally be zero, maybe one, occasionally two; and very rarely many more. That distribution will then be Poissonian, and will allow a test of any hypothesis that higher numbers lost – say on flights from a particular airport – are due to factors other than random variation.

[canucklad]

With all the technology at BA’s disposal, I’m sorry to say Ian, the longer the bag stays lost, the higher the probability it’s not going to be recovered. And normally the disappearance of a bag has to do with opportunistic thieving and thus the removal of identifying tags to cover up the crime!

Fingers crossed, that in this circumstance this isn’t the case, oh, and sorry for the pun

[Bath_VIP]

As I am probably the only professional statistician on this forum, I can confirm that the Binomial distribution is actually the correct distribution! Formally it is actually a Logistic Regression model since the expected % of bags lost can vary from day to day, route to route, etc. The binomial distribution describes the probability of R bags being lost out of a total N bags being transported.

In practice, the Poisson distribution will give very similar results since the probability of N bags being lost is infinitely small though it is not zero! But under a Poisson distribution, it is possible for an infinite number of bags to be lost which clearly can’t happen since at any particular time, there is only N bags being transported.

In both cases, when N gets very large, the Normal distribution can be a good approximation to both the Binomial & Poisson distributions. But it is only an approximation and if the data set contains a large number of R=0 situations (i.e. no bags lost) then it is inappropriate as the Normal can predict a negative number of bags lost which is nonsensical.

[KarlMarx]

Bath_VIP – 20/05/2014 09:54 GMT

In both cases, when N gets very large, the Normal distribution can be a good approximation to both the Binomial & Poisson distributions.

Nice to hear from a professional statistician, I am an economist and it is quite a while since I studied this topic.

Your point, quoted above, is on the money.

The population is every bag being transported on every flight during a day (a very large number) and as we do not have the data, a normal distribution is a good approximation.

Taking your point about the R=0 problem, sadly I doubt that this would be true often enough to cause a problem, but even without the data set, the numbers would be statistically significant to suggest that a normal distribution would approximate reality, unless you have some information that suggests unusual kurtosis combined with a skew – that would be an interesting exercise, if one compared individual airports against the whole population.

Before we get into too much of a focus on method, the point I was making and which DavidGordon10 challenged, was that Ian from HKG and his family suffered from the impact of randomness in the chain of bag losses in a short period of time.

He may never suffer from a bag mislaid again or it may occur on the next flight (I hope not, for his sake.)

I cannot see how one can blame an airline for such randomness, whilst I can totally see how one can blame an airline for failing to communicate and resolve the misplacement.

[IanFromHKG]

All way over my head, since I never studied statistics or economics, but I am nonetheless not wholly in agreement with the notion that I cannot blame the airline. The point being here that I am not blaming them for the randomness per se but (a) for losing the bags (and yes I know the two are related), and (b) doing such a dismal job of following up. Both seem to display failures about which the airline does not seem particularly to care. Add to this the fact that in the period in question (just four months) my family has taken a total of about three dozen flights on various airlines and that the only ones affected are BA short-haul, and to be honest it stops feeling like a random “pattern”, and starts feeling like incompetence on the part of one particular airline

Ah well, having said all that, with – up until this year – only two bags delayed/lost in over twenty years of frequent flying, perhaps it is just fate restacking the odds?? But please, let’s not get into a debate about how THAT probability distribution works. My head hurts… LOL

[canucklad]

Hi Gents, I’m in agreement with Ian on this. BA are at fault for having a system that randomly, and in Ian’s case, with a higher degree of probability have a system in place that makes such incidents occur on a more than regular basis.
Ponder this thought; most weeks in the UK, due to randomness millionaires are created. Now how do these millionaires come about, and whose responsible for creating the millionaires?

[KarlMarx]

I’m intrigued. What exactly is this ‘system’ that BA have created to lose bags? I thought that all airlines used the same systems in the airports?

BA are clearly totally responsible for how they choose to manage the misplacement of baggage and treat their customers.

[Author]

Posts