Trains can't leave before their departure time right?

[railiner]

Sorry, but wrong.

1) Post Office is just another cargo shipper. They do not have the authority to hold a flight (if they did, we would have had a delay code for them). If they don't have their shipment ready by the cargo cutoff time, it may not go.

2) Plane being full is not a factor. Passengers have both a check-in time cutoff and an at the gate ready to board cutoff. Miss them and your reservation can be cancelled. These days, those times are typically 45 minutes and 15 minutes respectively. So at 15 minutes before scheduled departure, if the plane is loaded and there is no one in the gate room, close the door and send the plane on its way. But note that between closing the door and first movement, there are still several things that have to happen. And scheduled departure time is when the plane is supposed to be moving, not just preparing to move.

3) Basically true. But Ramp ("below the wing") does their thing separately from what Customer Service ("above the wing") is doing and the passenger loading door may be closed even while Ramp is still doing their thing with bags and cargo.

4) Better to leave ATC out of it. Normally, they are not involved until the flight crew calls for push-back or taxi clearance which won't happen until the plane is ready to depart (so from the perspective of a late arriving passenger, the door is closed so the flight has already departed, even if the plane is still there).

Unlike how railroads traditionally operated, the published schedule provides no movement authority (even the public railroad timetable provided no movement authority, only the employee timetable did). If a flight is ready to depart early, there is no one who is going to say "uh, you're early, you can't leave yet"). So once all passengers, bags, and cargo that were ready by the appropriate cut-off times are on-board, there's no reason not to send the plane on its way regardless of what the schedule says.

Like I said, “apples and oranges”.

The only train that operates anything like that, is the Auto Train...

 

[B757Guy]

We do push back from the gate early, if everyone is on-board, and the airplane is buttoned up. Once we push, that starts the clock for the flight and is the time we use to measure if we departed on-time or not.

 

[jis]

We do push back from the gate early, if everyone is on-board, and the airplane is buttoned up. Once we push, that starts the clock for the flight and is the time we use to measure if we departed on-time or not.

Doesn't an early pushback also help in getting an earlier slot in the departure lineup?

 

[B757Guy]

Doesn't an early pushback also help in getting an earlier slot in the departure lineup?

In some cases, yes. It depends on a few different factors, along with the airport.

 

[MARC Rider]

This might become more of an issue with FRA insisting with the host railroads that trains must be on schedule at all en route stops, not just the end points, in their OTP metrics. It should be fun to see how this unfolds as Amtrak the infrastructure company argues with Amtrak the train operating company

I would hope that the performance metrics distinguish between 5 minutes late, 5 hours late, and 5 days late. I'm not sure I would ding a host railroad that much for an Amtrak train being 5 minutes late, the idea is to keep them from being 5 hours late. As far as 5 days late, not sure if that's ever happened with Amtrak, but my brother once took 3 days to fly from Chicago to Washington, due to flights cancelled by thunderstorms.

 

[railiner]

Doesn't an early pushback also help in getting an earlier slot in the departure lineup?

In some cases, yes. It depends on a few different factors, along with the airport.

Interesting...curious about that...does the airport management, or ATC determine that?

 

[AFS1970]

Not sure if this is an FRA thing, that would apply to Amtrak but on paper schedules for Metro North commuter trains there was a footnote that said they considered a train to be on time if it was within 5 minutes of the scheduled time. It did not specify if that was early, late or both.

One bit of fun timekeeping, that wouldn't work for trains, is my local bus service. A few years ago they got rid of quite a few timepoints. They kept all the stops, but only record the time at a few stops now. So they can be late at a lot more stops as long as they are on time to the next timepoint.

 

[railiner]

Not sure if this is an FRA thing, that would apply to Amtrak but on paper schedules for Metro North commuter trains there was a footnote that said they considered a train to be on time if it was within 5 minutes of the scheduled time. It did not specify if that was early, late or both.

One bit of fun timekeeping, that wouldn't work for trains, is my local bus service. A few years ago they got rid of quite a few timepoints. They kept all the stops, but only record the time at a few stops now. So they can be late at a lot more stops as long as they are on time to the next timepoint.

That is essentially what a "flag stop" is...you have to figure out the time yourself, based on the timepoint preceding your stop. Not that big a deal if it is relatively close, but a very big deal if further away, or if you don't like to have to spend extra time waiting....and if you happen to get to the stop and have missed it, you never really know if it has been there yet, unless someone else was waiting for a while...

Airlines generally consider a flight to be "on time" if it arrives within 15 minutes of scheduled time at the gate

 

[Dutchrailnut]

just to be clear the APTA norm is 5 Minutes and 59 seconds as applied by all APTA members for train to be late.

 

[B757Guy]

Interesting...curious about that...does the airport management, or ATC determine that?

ATC, and from an internal perspective, airline dispatch as well.

 

[railiner]

ATC, and from an internal perspective, airline dispatch as well.

I am even more curious now...
I can see airline dispatch deciding between Company flights, but how does ATC determine the 'slotting' otherwise?
If the flight departs, "on time", is it part of a pre-planned sequence? Or is it determined by "the first one to the intersection", or some other protocol.

I can see that at very busy hubs, getting just a slight 'jump; can result in a substantial savings in time and fuel burn, when twenty or so flights are scheduled to leave within minutes of each other....

 

[west point]

ATC does not delay airplanes too much when the weather is good. However there is always bad weather somewhere. I have been delayed by weather 2000 miles away. If too many airplanes can only thru a small space the aircraft will be metered. Another problem is getting a slot southbound from say Albany, Syracuse, Cleveland, Detroit, Chicago as all lanes may be taken by east west traffic. Flow control predicts when you might conflict with 'E-W traffic.
In winter time the effort of all airplane to take advantage of good winds east bound and avoid headwinds west bound. I know of times SFO - to SE destinations some flights will be routed over Mexico. and the Gulf of Mexico.

 

[Palmetto]

"Bad weather" is sometimes a ruse. I've had that excuse put out by an airline, and I've checked with the person picking me up at the destination and the weather was just fine. Speaking of bad weather, if you want to get delayed seriously by a weather hiccup, try flying out of EWR when there's one nearby.

 

[Ryan]

Weather exists and is relevant everywhere in between Point A and Point B.

 

[jebr]

Weather exists and is relevant everywhere in between Point A and Point B.

To add to this, weather between where the airplane is coming from and your Point A also matters, especially if you're not starting at one of that airline's hub airports. (The knock on effect from weather can go even further back, especially with ULCCs and/or carriers without much slack to reposition planes or adjust plane rotations to avoid extended delays.)

 

[PVD]

weather also plays a factor in planes not showing up from somewhere else, making either the plane, and often the flight crew, unavailable...I had that a few years ago, where the arriving aircraft was not going to be at JFK anywhere near departure time. Even though the airline was able to "redirect" another aircraft either released from maint. or from a flight leaving much later, we still had to wait for a crew.

 

[jis]

Indeed. Flight crew and cabin crew not making it to the connecting flight that they are supposed to be crewing can cause all sorts of delays, and frantic rotations and reassignments by operations to get things going. And then there is of course the case of crew making it eventually but then not having enough time left to complete the next flight so for all practical purposes being useless for that flight.

 

[lstone19]

"Bad weather" is sometimes a ruse. I've had that excuse put out by an airline, and I've checked with the person picking me up at the destination and the weather was just fine. Speaking of bad weather, if you want to get delayed seriously by a weather hiccup, try flying out of EWR when there's one nearby.

From someone who spent part of my career doing delay analysis for an airline, destination weather is just one small factor. Enroute severe weather, when it closes off airways and forces flights to longer routes and also increases congestion on the routes bordering the weather issue, can significantly delay flights.

To say that because the destination weather was fine, weather wasn't really the reason for a delay is the same as saying because no traffic congestion was visible at your destination, congestion wasn't the reason you were late driving somewhere.

 

[lstone19]

Airlines generally consider a flight to be "on time" if it arrives within 15 minutes of scheduled time at the gate

Different airlines have different goals. The US Department of Transportation considers a flight on-time for purposes of determining "chronically delayed" flights to be 14 minutes or less late (Arrival 14 as it is called in the industry). A given airline is likely to have both Arrival 00 (exactly on-time or early) and Arrival 14 goals.

Any time you speak of "airlines" as if the industry in one monolithic organization with everyone having the same goals, you are making a mistake. One of my pet peeves is people who would say that I worked for "the airlines". No, I worked for "an airline", one that competed with all those other airlines. Part of my job was knowing how well those competing airlines were performing so that we could see if we were doing better than them. An airline that targets Arrival 00 is likely to perform significantly better than one that thinks Arrival 14 is good enough.

 

[AFS1970]

I had a major delay in a flight once, that lead to one of my worst travel experiences, all due to bad weather in and around the destination. That weather delayed my take off by an hour and had mostly cleared up when we left, but the delays to other flights had lead to congestion that forced us to land at an alternate destination for fuel.

 

[west point]

Absolutely enroute weather delays cause much more late arrivals than origination or destination. It is like a RR has so many trains on say a 4 track main that gets 3 tracks out of service all at once. You cannot squeeze but so many trains or in the case of airlines so many thru one opening. There are if I remember about 5 inbound routes to southern Florida. More than once all routes were weathered and all of us had to deviate around the weather. MIA, FLL, WPB, ORL, TPA and non stops to the islands. The number of diversions in those cases are high .

 

[railiner]

Different airlines have different goals. The US Department of Transportation considers a flight on-time for purposes of determining "chronically delayed" flights to be 14 minutes or less late (Arrival 14 as it is called in the industry). A given airline is likely to have both Arrival 00 (exactly on-time or early) and Arrival 14 goals.

Any time you speak of "airlines" as if the industry in one monolithic organization with everyone having the same goals, you are making a mistake. One of my pet peeves is people who would say that I worked for "the airlines". No, I worked for "an airline", one that competed with all those other airlines. Part of my job was knowing how well those competing airlines were performing so that we could see if we were doing better than them. An airline that targets Arrival 00 is likely to perform significantly better than one that thinks Arrival 14 is good enough.

Was your airline targeting "Arrival 00"?
I am wondering if they did that, were they also competitive in published flight durations? Or did they put in a lot of padding to help achieve that on time performance?

 

[dlagrua]

The only experiences that we had with the train leaving early were the Autotrain (already covered here) and the Cardinal. On the Cardinal going West it will stop in WAS for about 40 minutes for an engine swap from electric to diesel. It leaves on scheduled time. That will probably continue until all diesel/electric engines are used but I see them shortening the stop then. Going East they also change engines at WAS but leave as soon as the swap is completed.

 

[lstone19]

Was your airline targeting "Arrival 00"?
I am wondering if they did that, were they also competitive in published flight durations? Or did they put in a lot of padding to help achieve that on time performance?

Yes, we had an Arrival 00 goal. As for the alleged padding, scheduling the length of flights (block time - time from departing the "blocks" to arriving on the "blocks") is far more complicated than 99% of the customers (or even employees) realize and much of the reason for that can be answered in one word - wind. Airplanes fly at a speed relative to the airmass which is moving. There is an optimal speed and the plane can fly a little faster or a little slower than that but beyond that, pretty much no can do (there is obviously a maximum speed but even slowing down is problematic as long before you get to the minimum speed at which a plane can fly, you get to an uncomfortable deck angle for the plane which makes service difficult and the flight uncomfortable for the passengers). Compare that to your car where you compensate for a head wind by just pushing down on the gas pedal a little bit more (land vehicles chase a desired ground speed, airplanes chase a desired air speed).

So, rather than having one correct block time for a market, you have a distribution of times which is pretty close to a normal distribution (bell curve) and the company decides where on the distribution it wants to target. Block targets are typically in the 65% to 75% area meaning that that percentage of the time, a flight's actual block time is equal to or less than the schedule. If you set the target to 50% and your flights depart exactly on time 100% of the time, you only arrive on-time or early 50% of the time.

So why not go to 100%? Several factors:
- the distribution curve tends to have a long tail. While the difference between 70% and 75% might be three minutes, the difference between 75% and 100% might be three hours.
- Crews at most airlines are paid the greater of scheduled and actual. The higher your block target, the higher your costs
- When there is weather and ATC (in the U.S.) imposes a Ground Delay Program (assigning arrival slots and holding flights at their origins until they can meet their arrival slot), the arrival slots are assigned based on scheduled arrival time. Increase the block time which means a later scheduled arrival (assuming the same scheduled departure) and you get a later arrival slot. You cannot schedule yourself into an on-time arrival when there is a Ground Delay Program.

For about 10 years, I did international block times including many flights scheduled over 12 hours. For those flights, the distribution curves told us that we'd have a few days a year where the actual block time would be an hour under or an hour over. Despite that, we'd have a day where a flight would arrive an hour early followed by some VP trying to tell us our block time was bad and I'd have to find some polite way to say "tell me when it starts happening every day as we know it will happen a few times a year" (if I could tell you which days well in advance, I'd be rich).

Besides enroute wind, there are other factors that go into block times which can vary (although a lot go back to wind). Planned taxi times are built into block times but which runways are active can make a huge difference. 10 minutes might be allowed at an airport but it might be that you normally take off to the north with a five minute taxi yet but some of the time, it's takeoff to the south with a 20 minute taxi. Same thing at the arrival airport. Plus the runways in use (selected by ATC due to wind) affect arrival terminal maneuvering time. Are you landing in the direction you came from allowing a straight-in or do you need to fly 10 miles past the airport and then turn back?

Incredible as it seems, one time I was on an EWR-ORD flight, scheduled at 2:33 where the actual block time was 1:59 - 22% under schedule. Why? Several reasons - it was a Saturday night so less ground congestion at EWR and the runway being used for departure was near the gate so taxi-out time was an insanely short (for EWR) nine minutes. Probably good enroute winds but then ORD winds allowed a straight-in to a runway that then let us exit practically right by the gate (four minute taxi-in time). Yet another day, that taxi-out could be 30 minutes and ORD might be landing to the east meaning fly past the airport and then come back (adding 10 to 15 minutes to the flight time) and then having you exit the runway two miles from the gate and you go over the 2:33 scheduled block.

One more thing - wind is weather. So if a flight departs on-time but arrives late due to enroute winds, at least my airline would say the reason is weather (it might be ATC but weather and ATC reasons both went into the same higher level delay reason category).

 

[RovinMoses]

Yes, we had an Arrival 00 goal. As for the alleged padding, scheduling the length of flights (block time - time from departing the "blocks" to arriving on the "blocks") is far more complicated than 99% of the customers (or even employees) realize and much of the reason for that can be answered in one word - wind. Airplanes fly at a speed relative to the airmass which is moving. There is an optimal speed and the plane can fly a little faster or a little slower than that but beyond that, pretty much no can do (there is obviously a maximum speed but even slowing down is problematic as long before you get to the minimum speed at which a plane can fly, you get to an uncomfortable deck angle for the plane which makes service difficult and the flight uncomfortable for the passengers). Compare that to your car where you compensate for a head wind by just pushing down on the gas pedal a little bit more (land vehicles chase a desired ground speed, airplanes chase a desired air speed).

So, rather than having one correct block time for a market, you have a distribution of times which is pretty close to a normal distribution (bell curve) and the company decides where on the distribution it wants to target. Block targets are typically in the 65% to 75% area meaning that that percentage of the time, a flight's actual block time is equal to or less than the schedule. If you set the target to 50% and your flights depart exactly on time 100% of the time, you only arrive on-time or early 50% of the time.

So why not go to 100%? Several factors:
- the distribution curve tends to have a long tail. While the difference between 70% and 75% might be three minutes, the difference between 75% and 100% might be three hours.
- Crews at most airlines are paid the greater of scheduled and actual. The higher your block target, the higher your costs
- When there is weather and ATC (in the U.S.) imposes a Ground Delay Program (assigning arrival slots and holding flights at their origins until they can meet their arrival slot), the arrival slots are assigned based on scheduled arrival time. Increase the block time which means a later scheduled arrival (assuming the same scheduled departure) and you get a later arrival slot. You cannot schedule yourself into an on-time arrival when there is a Ground Delay Program.

For about 10 years, I did international block times including many flights scheduled over 12 hours. For those flights, the distribution curves told us that we'd have a few days a year where the actual block time would be an hour under or an hour over. Despite that, we'd have a day where a flight would arrive an hour early followed by some VP trying to tell us our block time was bad and I'd have to find some polite way to say "tell me when it starts happening every day as we know it will happen a few times a year" (if I could tell you which days well in advance, I'd be rich).

Besides enroute wind, there are other factors that go into block times which can vary (although a lot go back to wind). Planned taxi times are built into block times but which runways are active can make a huge difference. 10 minutes might be allowed at an airport but it might be that you normally take off to the north with a five minute taxi yet but some of the time, it's takeoff to the south with a 20 minute taxi. Same thing at the arrival airport. Plus the runways in use (selected by ATC due to wind) affect arrival terminal maneuvering time. Are you landing in the direction you came from allowing a straight-in or do you need to fly 10 miles past the airport and then turn back?

Incredible as it seems, one time I was on an EWR-ORD flight, scheduled at 2:33 where the actual block time was 1:59 - 22% under schedule. Why? Several reasons - it was a Saturday night so less ground congestion at EWR and the runway being used for departure was near the gate so taxi-out time was an insanely short (for EWR) nine minutes. Probably good enroute winds but then ORD winds allowed a straight-in to a runway that then let us exit practically right by the gate (four minute taxi-in time). Yet another day, that taxi-out could be 30 minutes and ORD might be landing to the east meaning fly past the airport and then come back (adding 10 to 15 minutes to the flight time) and then having you exit the runway two miles from the gate and you go over the 2:33 scheduled block.

One more thing - wind is weather. So if a flight departs on-time but arrives late due to enroute winds, at least my airline would say the reason is weather (it might be ATC but weather and ATC reasons both went into the same higher level delay reason category).

As a pilot, I appreciate your excellent explanation of winds, airport arrival procedures, and the challenges of predicting block times. Every now and then I would chuckle when the captain would announce, during a late push back, that "we're sorry for the delay, but we'll put the 'pedal to the metal' and try to get you to your destination as soon as possible." That might have been reassuring for some passengers who thought the captain was going to burn more fuel and increase to maximum airspeed, therefore shaving some time off the journey.