Departure Times Varying by a Few Mins Each Hour

[Manutd1999]

Why is that services often have (seemingly) random variations of +/- a few minutes from hour to hour?

Taking an example from my local line - Bradford to Huddersfield for a random day in July. Some departures from Interchange are:
1036, 1137, 1236, 1337, 1436, 1536, 1636, 1736, 1838, 1937, 2038

Why the variation?? For example, why not make all departures at xx36? Similar examples exist all across the network.

Granted this makes little to no difference to the average passenger, but surely it is easier operationally to standardise everything as much as possible? All these little variations must be a nightmare when trying to find new paths through heavily utilised areas.

 

[ABB125]

The variations are presumably due to clashes with freight paths which only run in certain hours. In an ideal world either there would be infrastructure changes to allow a consistent departure time, or the freight paths would be slightly amended. However, both of those are very unlikely to happen, for a variety of reasons (some of which are valid, others less so...)

 

[Manutd1999]

If it is only a 1-2 mins shift why don't they path the passenger services as-if the freight paths exist every hour?

Maybe this would restrict capacity too much in some areas but for a lot of lines it would be fine. It would add flexibility for the freight operators as well....

 

[The exile]

All these little variations must be a nightmare when trying to find new paths through heavily utilised areas.

Presumably they are more the result of the nightmare of trying to find paths. No doubt some are fossils, though.

 

[zwk500]

If it is only a 1-2 mins shift why don't they path the passenger services as-if the freight paths exist every hour?

Maybe this would restrict capacity too much in some areas but for a lot of lines it would be fine. It would add flexibility for the freight operators as well....

There might be different paths in different hours that means sometimes it has to go 2 minutes early.

In an ideal world either there would be infrastructure changes to allow a consistent departure time

Or the more rudimentary strategy of putting the advertised times back a minute or two could be used. Only problem is that then in some hours the train will always be intentionally 2 minutes late starting...

 

[mangyiscute]

There might be different paths in different hours that means sometimes it has to go 2 minutes early.

Or the more rudimentary strategy of putting the advertised times back a minute or two could be used. Only problem is that then in some hours the train will always be intentionally 2 minutes late starting...

Often the train could be being pathed around "runs as required" freight paths so the freight train using that path doesn't run too frequently, hence you could schedule it as usually, most days it could run at the right time and occasionally it may be delayed a couple mins

 

[greenline712]

As a retired bus scheduler, I've often wondered why this "odd-minute" timing is scheduled. For Joe Public, who have no real interest in the "why", it must be bemusing!1

I do get the need to slightly flex paths for freight services, but the North London Line has many such instances, not all of which are repeated hourly. Having a time which is marked "arrives 1 minute earlier on School Wednesdays" or some such is daft . . . just show the journey as a standard time, and ignore the variation . . . much simpler!

Is this because passenger services are monitored as "right time plus three", and a four minute delay triggers some compensation payment?

{In the bus world, we are monitored (generally) as time plus five, and we accept that running times are only ever an average}.

 

[zwk500]

As a retired bus scheduler, I've often wondered why this "odd-minute" timing is scheduled. For Joe Public, who have no real interest in the "why", it must be bemusing!1

I do get the need to slightly flex paths for freight services, but the North London Line has many such instances, not all of which are repeated hourly. Having a time which is marked "arrives 1 minute earlier on School Wednesdays" or some such is daft . . . just show the journey as a standard time, and ignore the variation . . . much simpler!

Is this because passenger services are monitored as "right time plus three", and a four minute delay triggers some compensation payment?

{In the bus world, we are monitored (generally) as time plus five, and we accept that running times are only ever an average}.

The primary reason is time - going through each schedule to establish the pattern and offset the public times would take a very long time. There are many reasons why the working time might vary by a minute or two, so it would be done on advertised times.
However the delay stats are probably part of the thinking.

 

[43074]

Doesn't tend to help that Network Rail planning don't "get" the concept of a repeating standard pattern timetable; when they check each train in the timetable works with one another, the process is done on a train by train basis, in doing so losing sight of the bigger picture and causing variations to the timetable like this from hour to hour.

Chances are Northern will have submit the paths to Network Rail with all of them departing Bradford at 36 past the hour, and that's what Network Rail have offered back and it's been accepted by Northern...

 

[jammmmmsssssss]

Doesn't tend to help that Network Rail planning don't "get" the concept of a repeating standard pattern timetable; when they check each train in the timetable works with one another, the process is done on a train by train basis, in doing so losing sight of the bigger picture and causing variations to the timetable like this from hour to hour.

Chances are Northern will have submit the paths to Network Rail with all of them departing Bradford at 36 past the hour, and that's what Network Rail have offered back and it's been accepted by Northern...

But why does it matter if a train is scheduled one minute later one hour?

 

[BrianW]

Why is that services often have (seemingly) random variations of +/- a few minutes from hour to hour?

Taking an example from my local line - Bradford to Huddersfield for a random day in July. Some departures from Interchange are:
1036, 1137, 1236, 1337, 1436, 1536, 1636, 1736, 1838, 1937, 2038

Why the variation?? For example, why not make all departures at xx36? Similar examples exist all across the network.

Granted this makes little to no difference to the average passenger, but surely it is easier operationally to standardise everything as much as possible? All these little variations must be a nightmare when trying to find new paths through heavily utilised areas.

I imagine the planners would be delighted to time everything in your example at xx36 of they could. I imagine passengers also in that example remember the trains are about xx36 and are therefore there 'in good time' for the trains which are scheduled to depart at xx37 or2038, hoping that the train is on time. The reasons for the odd times will be as others have stated- trains are scheduled to fractions of minutes. Irregular stopping patterns, skipping some, may also be relevant. I think the measure of, and penalties for, lateness are also pertinent.

 

[Manutd1999]

Doesn't tend to help that Network Rail planning don't "get" the concept of a repeating standard pattern timetable; when they check each train in the timetable works with one another, the process is done on a train by train basis, in doing so losing sight of the bigger picture and causing variations to the timetable like this from hour to hour.

I agree - this is what we should work towards IMO and it is achievable in many areas, but it doesn't even seem to be a consideration at present, with major recasts still ending up with these variations.

It would actually save Network Rail time long-term as the future timetabling edits would become much simpler.

 

[mangyiscute]

I imagine the planners would be delighted to time everything in your example at xx36 of they could. I imagine passengers also in that example remember the trains are about xx36 and are therefore there 'in good time' for the trains which are scheduled to depart at xx37 or2038, hoping that the train is on time. The reasons for the odd times will be as others have stated- trains are scheduled to fractions of minutes. Irregular stopping patterns, skipping some, may also be relevant. I think the measure of, and penalties for, lateness are also pertinent.

The main issue is when the trains are sometimes slightly earlier - for example anyone who often takes it would tell you that the cross country to Manchester leaves Reading at xx:15, but it occasionally leaves as early as xx:12, which definitely could catch some people out.

 

[The Planner]

Doesn't tend to help that Network Rail planning don't "get" the concept of a repeating standard pattern timetable; when they check each train in the timetable works with one another, the process is done on a train by train basis, in doing so losing sight of the bigger picture and causing variations to the timetable like this from hour to hour.

Chances are Northern will have submit the paths to Network Rail with all of them departing Bradford at 36 past the hour, and that's what Network Rail have offered back and it's been accepted by Northern...

Of course we "get" it. If Northern bid the paths like that, and they are compliant, why would we alter them? Northern also don't have to accept what NR offer them.

 

[WAB]

{In the bus world, we are monitored (generally) as time plus five, and we accept that running times are only ever an average}.

Then again, the railway would fall to bits if bus practices were adhered to... I seem to recall this caught out a lot of the bus companies who bidded for franchises in the early days of privatisation...

 

[Llandudno]

Merseyrail pride themselves, and quite rightly so on their clockface every 15 minute service on all lines between 7am-7pm, the only exception are the trains from Chester which depart at 01, 15, 31 and 45 minutes past each hour.

I know that the 01 and 31 departures omit Capenhurst but wouldn’t it be simpler to advertise in the public timetable that trains depart Chester on the hour and every 15 minutes even though the trains omitting Capenhurst will actually depart 1 minute ‘late’ in the working timetable?

 

[Manutd1999]

Of course we "get" it. If Northern bid the paths like that, and they are compliant, why would we alter them? Northern also don't have to accept what NR offer them

I highly doubt Northern requested that their departure times vary throughout the day?

More likely NR told them that they had to if they wanted to operate those services. Now there may be good reasons for that (ie, freight paths) but, from the outside, it does seem like there is little effort put into obtaining a "standard hour".

In this example, I assume the conflicts are in the Brighouse - Hudds section. So why couldn't all the departures from Bradford be set to xx36, with select trains waiting an extra minute somewhere between Halifax and Brighouse if necessary? They would still have different arrival times but the Bradford - Halifax section would be standardised.

It may look like tiny differences but if you did this across the network then all of these little changes could add up to a big improvement in simplicity/efficiency.

 

[The Planner]

I highly doubt Northern requested that their departure times vary throughout the day?

More likely NR told them that they had to if they wanted to operate those services. Now there may be good reasons for that (ie, freight paths) but, from the outside, it does seem like there is little effort put into obtaining a "standard hour".

How do you know that? Some of the stuff we get is far from fantastic in bids.

 

[Basil Jet]

It's worth mentioning that no train in the public timetables departs or arrives at 0000, so if a train normally departs on the hour, it will be varied by a minute at midnight to prevent it leaving at 0000.

 

[Wilts Wanderer]

There could be different reasons for minor variations from hour to hour:
- rolling stock with different sectional timings or reversal times etc
- variations in platforming with bespoke timing allowances
- different stock working - particularly in the shoulder-peaks
- variations in running time
- specific crew diagramming requirements
- interactions with other services (each with their own variations)

Most/all of these can be tidied up with ‘GBTT/WTT differentials but this does take a lot of time/effort and in a way is essentially misleading the travelling public slightly as to what time the train will run.

== Doublepost prevention - post automatically merged: 14 May 2024 ==

How do you know that? Some of the stuff we get is far from fantastic in bids.

From my experience on the operator side, timetable bids are usually evolutions of previous NR timetable offers. Very rare to start from a blank piece of (graph) paper.

 

[Snow1964]

From my experience on the operator side, timetable bids are usually evolutions of previous NR timetable offers. Very rare to start from a blank piece of (graph) paper.

The primary reason is time - going through each schedule to establish the pattern and offset the public times would take a very long time.

Fair enough in old days, but in an era where computers can scan millions of entries in seconds for patterns, is this really still valid. Feels like an old fashioned excuse from slide rule and using graph paper era, rather than modern solution getting a computer to optimise paths and clock face intervals.
A 1950s answer to avoid using AI era computer analysis

However the delay stats are probably part of the thinking.

A one minute variation is not going to affect right time within 3 minutes, or delay repay. If the thinking is along lines of ok to be 2 minutes late, but not 3 minutes, then have to question why worried about first minute, if talking 2-3 minutes for stats, unless being one minute late is somehow seen as dishonourable

 

[The Planner]

Fair enough in old days, but in an era where computers can scan millions of entries in seconds for patterns, is this really still valid. Feels like an old fashioned excuse from slide rule and using graph paper era, rather than modern solution getting a computer to optimise paths and clock face intervals.
A 1950s answer to avoid using AI era computer analysis

If that was possible, it would have been done by now. Considering the last major software change in 2010 was meant to reduce headcount and we now have more, you would assume it would have been done already.

 

[Wilts Wanderer]

Fair enough in old days, but in an era where computers can scan millions of entries in seconds for patterns, is this really still valid. Feels like an old fashioned excuse from slide rule and using graph paper era, rather than modern solution getting a computer to optimise paths and clock face intervals.
A 1950s answer to avoid using AI era computer analysis

As yet there is no automated timetable planning software capable of assimilating several hundred thousand timetable planning rules and applying them accurately. If you worked in the industry you’d be surprised how much human thought is required to optimally plan a timetable, and maintain it.

A one minute variation is not going to affect right time within 3 minutes, or delay repay. If the thinking is along lines of ok to be 2 minutes late, but not 3 minutes, then have to question why worried about first minute, if talking 2-3 minutes for stats, unless being one minute late is somehow seen as dishonourable

With connectional allowances as tight as 3 minutes in places, cheerfully planning a train to be advertised 2 minutes earlier than it actually runs is not putting the customer first. Right Time Railway is based on WTT not GBTT.

 

[WAB]

Fair enough in old days, but in an era where computers can scan millions of entries in seconds for patterns, is this really still valid. Feels like an old fashioned excuse from slide rule and using graph paper era, rather than modern solution getting a computer to optimise paths and clock face intervals.
A 1950s answer to avoid using AI era computer analysis

A lot of time and money has gone into trying to upgrade train planning systems, but not a lot of progress has been made. Large parts of the industry still work on what is effectively a souped-up version of BR's PROTIM. Let's just say that the upgrade progress hasn't matched the general advance of computing technology...

 

[nw1]

Another reason, admittedly not so prevalent these days, is that certain lines could in theory have a two-or-three-times-a-day long distance service which has to be inserted into the regular clockface somehow.

Say your standard pattern of a local stopping service is an xx18 departure, for example, to match an xx48 on the opposite half-hour. Say the xx18 immediately follows an HST service at xx15.

The occasional long-distance might, for pathing reasons, also have to go out immediately behind the HST. For example, it might be a classic loco-hauled service which has to keep ahead of a further HST at xx45 and cannot be delayed too much. But it's faster than the regular xx18 stopper, and therefore has to go out in front of it. So in those small number of hours the occasional long distance might go out at xx18 and you would then need to retime the stopper to xx20 or xx21.

 

[Bletchleyite]

Another reason, admittedly not so prevalent these days, is that certain lines could in theory have a two-or-three-times-a-day long distance service which has to be inserted into the regular clockface somehow.

The Avanti services on the North Wales Coast do contribute to that timetable being an absolute mess, certainly.

 

[Llandudno]

The Avanti services on the North Wales Coast do contribute to that timetable being an absolute mess, certainly.

I think TfW are the main contributor to the messy timetable on the north Wales coast though..!

 

[Bletchleyite]

I think TfW are the main contributor to the messy timetable on the north Wales coast though..!

Probably so. It shouldn't in theory be a hard route to timetable fully clockface with just 2tph plus the Conwy Valley and the shuttles (and the odd freight), but they do nonetheless seem to make a right mess of it.

Sundays are particularly bad (and work to a totally different, more 1990s-like, pattern than a weekday).

 

[185]

Passengers do expect a clockface timetable, and proper planning should be able to produce this. Freight should be fitted in around them - but in the small number of times this would be impossible, the working time should be up to 2 minutes later, whilst keeping the public advertised time clock face.

Passenger trains should never be worked around freight... I'm a big believer in freight on rails, however poor path planning and unreliable/faulty loco operators which lose time & block the job do need to be addressed. Some weeks ago, a signaller letting a light loco out held up five class 1 passenger trains - compound delays of 40 minutes. Another signaller had two 100mph+ trains nose to tail, seemingly forgot there were two, and brought the second to a stand to allow a somewhat unhealthy 20-30mph freight in front for 20 miles. 26L when it got to overtake.

 

[zwk500]

Some weeks ago, a signaller letting a light loco out held up five class 1 passenger trains - compound delays of 40 minutes. Another signaller had two 100mph+ trains nose to tail, seemingly forgot there were two, and brought the second to a stand to allow a somewhat unhealthy 20-30mph freight in front for 20 miles. 26L when it got to overtake.

Was this a planning error or was the signaller regulating trains running out of booked paths?