Why is it so hard to predict arrival times?

[Mark in CA]

I've just had to meet people at the Budapest airport two days in a row. On arrived on a Lufthansa flight from Munich, the other a WizzAir flight from Catania, Italy. It's about a 30-minute drive to the airport, and in preparation for meeting these people I watched the event log on Flight Stats to see what the estimated time of arrival was. In both cases, I noticed wildly fluctuating estimates.

The flight from Munich, which usually takes about an hour, this time ahd to divert around convective weather, resulting in a much longer flight. Still, as I watched the flight cross the border from Croatia into Hungary, the estimated arrival time jumped an additional 20 minutes. Looking at the distance involved and the aircraft's speed, it was easy to see the fight would reach Budapest in only 15 minutes, not the 40 minutes now estimates, and so I ignored it and left for the airport. Sure enough, the flight arrived when I thought it would, and when I checked again, the event log had been updated back to the original estimate.

Today, the flight from Italy was delayed, but once it got off the ground, the estimated arrival time seemed appropriate. But a couple hours later, the arrival time got bumped an additional 30 minutes. I left for the airport expecting the later arrival time, but as it turned out, the flight arrived about 20 minutes earlier than I expected. AGain, after checking the event log again, I noticed the change to the earlier time had been entered shortly after I left for the airport.

In both instances, the additional flight times were posted relatively close to the destination airport and then shortly afterward updated again closer to the previous estimated arrival times.

I don't recall this volatility in arrival times before. Has something changed recently that causes this, or did I just have bad luck?

[evansb]

Touch down time or time arrived at the gate? Could it be gate availability?

[B2N2]

Depending on Airport design they may have much longer taxi times both departing and arriving if they change the runway being used.
Different runway being used may also mean different departure procedures which may add or reduce the flight time.

[Mark in CA]

Touch down time or time arrived at the gate? Could it be gate availability?
Gate times.

[Mark in CA]

Budapest is not a major airport like LGW. Essentially a single terminal, nothing large than a 777 landing (with passengers), no traffic delays and very short taxi times. Always. Very little to be unpredictable here.

[hoss183]

Dont look at estimates, find the plane on Fr24 and track it in.

[atakacs]

Did you get similar ETA jumps on the airport website?

[WHBM]

Despite all written above it's a uniquely British problem. It used to afflict just Heathrow but has now spread to other airports in the country.

We all experience wherever else we fly to that ground crew are somehow normally awaiting our arrival. In the UK, not so. 15 minutes ahead of schedule, or behind, and that's an excuse. If you chance to be right on the minute, then it's No Stand Available. In my experience it's quite rare to arrive on BA at LHR and all the necessary resources are in place.

I fly BA in and out of far northern, snow-bound airports, where they operate in and out without difficulty. De-icing somehow turns up on time. Runway clearance is maintained. Their ops people must see this. Yet at Heathrow a fraction of these conditions causes everything to fall apart.

The recent Ryanair issues have caused gross difficulties with returning flights delayed to after midnight, and no ground crews available. As Ryanair generally operate out-and-back, the delay is typically predictable the moment it sets off, because the outstations seem to be able to turn the late arrival round in the normal time. But back in the UK - "all gone 'ome, mate".

[wiggy]

I’d suggest for starters answers/reasons:

Local management not able or prepared to pay overtime.

Local management not able or prepared to pay for staffing and equipment to allow a robust operation to continue in the event that the stand plan does not operate to perfection.

I’d agree other places seem on the whole to manage it differently.

[WHBM]

I’d agree other places seem on the whole to manage it differently.

One thing that overseas places manage to avoid is their national aviation infrastructure being seen as a Giant Ponzi Scheme by investors from overseas who don't know one end of an aircraft from the other, an approach which has spread across the UK airport infrastructure. Bad news is always when one of the Australian investors, or in particular Canadian public sector pension plan investors, manage to acquire significant or majority shareholdings.

Heathrow was a classic, where the Spanish investors grossly overpaid for it, and then all management attention was only on making some return on this amount by any means possible. Contractors on the airport, such as handling agents or other commercial operators, are then in turn screwed down for huge fee increases which they can only come up with by minimising their costs all down the line. Minimum equipment, minimum-wage personnel, and operators coming and going as they just can't make a penny out of it, all the funds being diverted out of the operation.

[Peter47]

I've spent a long haul flight at the ETA on the moving map and its amazing how it varies from hour to hour. Some ideas for this:

Changes in altitude, head/tail winds, etc.
Flying off course (I've been in some quite long diversions around thunder storms and you might not get the requested route
Being stuck below a slow flying aircraft (a pilot told me how they couldn't change altitude to pass a slow flying 767 (Mach 0.78?) - cost a lot of time
Being stuck in a stack
Change in landing runway direction - could easily cost 10 minutes, more if it leads to additional holding
Getting a straight in - I saw the eta into HKG magically reduce by 15 minutes at a stroke once
Changing the cruising speed - aircraft often fly below max speed to save fuel. I was once delayed 2 hrs in HKG when the plane went tech (day flight to London). Guess what, the flight time went down by 20 minutes - it may have saved breaking the night jet ban on the return flight.

Of course some of these are less likely to apply to a SH flight like MUC - BUD where the arrival time should be reasonably well known once the plane takes off.

[WHBM]

Don't confuse a simplistic little on-board App to calculate arrival times with the mainstream Ops team information. And long-haul arrivals have always been subject to many variables, quite why terminal arrival resources are then planned to the minute for them shows a lack of understanding.

I wonder if those chaps who do the Heathrow T5 arrival short term plan would be impressed when their wives suddenly call the maternity hospital but are told "no, you can't come in, you are not due for another 2 days. That's your slot".

[DaveReidUK]

Don't confuse a simplistic little on-board App to calculate arrival times with the mainstream Ops team information. And long-haul arrivals have always been subject to many variables, quite why terminal arrival resources are then planned to the minute for them shows a lack of understanding.

I wonder if those chaps who do the Heathrow T5 arrival short term plan would be impressed when their wives suddenly call the maternity hospital but are told "no, you can't come in, you are not due for another 2 days. That's your slot".

That's rather a skewed way of looking at it.

No airline or airport can afford to have twice as many as the budgeted number of engineers or ramp workers sitting around the crewroom as a contingency against a few flights arriving early or late. The planners have to take a reasonable view of what's likely to happen most of the time, and provision manpower accordingly, in the knowledge that there will be some days when everything goes pear-shaped and aircraft have to wait for stands/steps/tugs, etc.

In my experience, maternity hospitals do the same. :O

[WHBM]

No airline or airport can afford to have twice as many as the budgeted number of engineers or ramp workers sitting around the crewroom as a contingency against a few flights arriving early or late.This one always comes up, that the only route seen as an alternative to having every resource planned up to the theoretical hilt is to have DOUBLE the amount. Not a sensibly calculated contingency percentage but it has to be DOUBLE.

BA don't have DOUBLE the crew at their overseas outstation handlers, yet manage far better than back at base. Shouldn't it be the other way round ?

in the knowledge that there will be some days when everything goes pear-shaped and aircraft have to wait for stands/steps/tugs, etc.
I'm not speaking (I believe) about everything pear-shaped days. I am speaking about perfectly straightforward ops days.

In my experience, maternity hospitals do the same.Not mine (admittedly as involved spectator).

[wiggy]

BA don't have DOUBLE the crew at their overseas outstation handlers, yet manage far better than back at base. Shouldn't it be the other way round ?

At most stations they won’t have any “crew” - that’s the subcontractors job, which leads into how different departments at BA work, no doubt service level agreements, certainly silo budgeting and a thousand and one other things that ultimately often leads to T5 being an extremely frustrating place to operate into (not excusing any of it).

[DaveReidUK]

This one always comes up, that the only route seen as an alternative to having every resource planned up to the theoretical hilt is to have DOUBLE the amount. Not a sensibly calculated contingency percentage but it has to be DOUBLE.

Not at all.

It's a fairly simple trade-off: the more slack you have in the form of spare resources, the fewer severely disrupted days you are likely to have. But to ensure no pear-shaped days at all, even doubling resources probably wouldn't be sufficient.

And of course the bean-counters aren't going to countenance that anyway.

Heathrow for example budgets for about a 300:50:15 balance - thats to say in terms of days per year:
300 where things go more-or-less as planned, with only minor hiccups
50 where the schedule is significantly disrupted but recoverable
15 where things get so bad the schedule can't be recovered (snow, severe ATC delays, etc)

[WHBM]

15 where things get so bad the schedule can't be recovered (snow, severe ATC delays, etc)
To anticipate half a month's gross delay days per year is surely way outside the experience of the rest of the air transport world, especially as how such items are measured then gets subject to statistics manipulation.

It's also one of those things where, once such goals are set, people get into the way of accepting them as the norm. We had 15 days complete disorganisation in a year but we have "done our job to target". It's like security/immigration queue times. Once these were set as a "goal" of not more than 45 minutes, the perception comes that 45 minutes is the goal. "People waiting 30 minutes ? We're inside the target, a couple of you can go on an extra break".

[Mark in CA]

Dont look at estimates, find the plane on Fr24 and track it in.
I was monitoring the event log on Flightstats. I believe these arrival time estimates are generated by the airline. But yes, i do track the flights, too, usually on Flightaware to make up my own mind on when it will arrive.

[Mark in CA]

Did you get similar ETA jumps on the airport website?
Wasn't watching that.