MSE TicketySplit Advanced - beta

[IanD]

Hi Deerfold,

Was this your search? http://www.moneysavingexpert.com/split-cheap-train-tickets/results/52b1b646806f00.52766415

I see that your request was for the 16:44 train, which is £41.80 as an advance ticket.

We won't pick a later train if the price is cheaper - we simply pick the first train that satisfies the user's request.

Adam

That can't be true - I asked for MKC to Thurso at or after 0600 and was offered a service at 0813 towards Glasgow (with a change at Lancaster, Haymarket and Inverness) whereas the first train that would have satisfied my request would have been the 0622 to Glasgow Central (then cross to Queen Street for a service to Inverness).

=== New bit

A simple split on the 0622 at Glasgow would have brought the cost down to £39.70 which is much better than the split at Aviemore I was offered. (Ignoring the £36 through Advance you can get from RSH). Does it not like cross-city transfers?

 

[MSEAdam]

Hi Ian,

From looking on thetrainline and national rail, I only see trains departing at 7:13 or 8:13 - both arriving at the same time - which is why you see the 8:13 selected...

Adam

 

[Deerfold]

Hi Deerfold,

Was this your search? http://www.moneysavingexpert.com/split-cheap-train-tickets/results/52b1b646806f00.52766415

I see that your request was for the 16:44 train, which is £41.80 as an advance ticket.

We won't pick a later train if the price is cheaper - we simply pick the first train that satisfies the user's request.

Adam

That seems to be a shame - many people would be prepared to travel 20 minutes later to save a third on the price.

However even if I search a few minutes later so the next departure would take me via Appleby it waits until the train afterwards (presumably as it has the same arrival time).

Also, please note, they're not Advance tickets - they're all Anytime Singles.

 

[MSEAdam]

Thanks Deerfold - this is the first iteration of our development, so enhancements like the bits you are suggesting here will likely appear in our roadmap if this tool is useful to users.

And noted about advance - thanks!

Adam

 

[Deerfold]

Thanks Deerfold - this is the first iteration of our development, so enhancements like the bits you are suggesting here will likely appear in our roadmap if this tool is useful to users.

And noted about advance - thanks!

Adam

Happy to be of help.

 

[IanD]

Hi Ian,

From looking on thetrainline and national rail, I only see trains departing at 7:13 or 8:13 - both arriving at the same time - which is why you see the 8:13 selected...

Adam

That's not so good - put in via Glasgow on NRE and it suggests a £36 ticket at 0622 as well. However, it sends you to Virgin Trains by default who then try to sell you the much the much more expensive ticket on the 0713/0813 unless you re-enter the via Glasgow bit.

The trainline also offers the £36 ticket if you specify via Glasgow.

To me, via Glasgow would be the most obvious route but I suppose the journey planners think it would be preferable to change at Haymarket (even though this requires one or 2 additional changes before you get that far).

 

[bb21]

That can't be true - I asked for MKC to Thurso at or after 0600 and was offered a service at 0813 towards Glasgow (with a change at Lancaster, Haymarket and Inverness) whereas the first train that would have satisfied my request would have been the 0622 to Glasgow Central (then cross to Queen Street for a service to Inverness).

You won't have got there earlier.

The way journey planners work is that if a later departure gives the same arrival time, the earlier departure is not always shown. Of course some booking engines may return these journey opportunities on earlier departures, but not always.

 

[rww100]

Any chance of being able to specify 'via' points, as well as alternative routes at some point?

One example I've just tried is Liverpool Lime Street to London Euston.

Firstly I notice that using the correct three-character station code 'LIV', the system searches for tickets from London Liverpool Street instead.

The tool does seem to work on the direct Virgin service bringing the cost down slightly by a couple of pounds to around £60 on a sample journey tomorrow. If I were able to specify a slower route of 'London Midland Trains only' though, I would hopefully be shown a £28.90 Super Off-Peak Return, or an even cheaper London Midland Advance, something which the tool doesn't currently allow.

For people who are keen to save money on train tickets, sometimes a slightly longer journey time is something they are happy to endure. It would be nice to see this option in the tool rather than only showing splits on the fastest services which are often many times more expensive.

 

[IanD]

You won't have got there earlier.

The way journey planners work is that if a later departure gives the same arrival time, the earlier departure is not always shown. Of course some booking engines may return these journey opportunities on earlier departures, but not always.

Not earlier, no, but £108 cheaper - the point of ticketysplit is to save money.

 

[bb21]

MSEAdam has made it very clear that currently the tool only looks at the next one journey opportunity, in which case it is understandable that the option with the shortest journey time is analyzed if several arrive at the same time.

It would be a good tool to incorporate for future releases, but for now that is what the tool does. MSEAdam's claim in this post is therefore not false as you alluded to.

 

[IanD]

Thought he was asking for feedback on how well the thing performs.

 

[bb21]

That's absolutely fine. It is something that can be developed in the future. I was merely pointing out that MSEAdam was not making false statements and explaining possible reasons behind the seemingly strange results to any reader who may not be aware of the intricacies of the booking engines.

I was not having a go at you, so apologies if I came across as being too abrupt or aggressive.

 

[Lockwood]

I put in Havant to Bristol, and it came back with nothing.
I know that on a return, you can do some epic splitting and trim off a fiver.

Havant - Romsey - Salisbury - Trowbridge - Bristol

I think even with one of those splits in it works out cheaper.
I put in Salisbury - Bristol, and that too omits the Trowbridge split

(It may be Warminster, not Trowbridge, point remains)

 

[MSEAdam]

I put in Havant to Bristol, and it came back with nothing.
I know that on a return, you can do some epic splitting and trim off a fiver.

Havant - Romsey - Salisbury - Trowbridge - Bristol

I think even with one of those splits in it works out cheaper.
I put in Salisbury - Bristol, and that too omits the Trowbridge split

(It may be Warminster, not Trowbridge, point remains)

Hi Lockwood,

For the journey you searched, walk-on fare provides a best split @ Basingstoke.

When checking advance fares, splitting at Basingstoke causes the fare to rise from £39 to £40.50.

Adam

 

[soil]

It's worth noting that this is Money Saving Expert, not Money Supermarket!

Not particularly. Money Saving Expert, which is owned by Money Supermarket, makes its money from commission. The Trainline advertise a 1.5% commission rate, http://www.thetrainline.com/about-us/affiliates/ though whether they will have negotiated a higher rate with Money Supermarket is not clear to me. My guess would be yes, as it appears that RSH are offering a higher commission rate - 2%. http://www.redspottedhanky.com/about-us/affiliates/join-the-program/

Money Saving Expert originally started with a 'eyeballs first, figure out how money later' business model, and gradually over time moved firstly towards listing commission-paying links and then non-commission-paying links side-by-side asking people to click the commission-paying links to 'support the site', then subsequently moving the non-commission-paying links to the bottom of the page, finally removing the non-commission-paying links entirely.

This is not the first case where the quality of the advice has been compromised by a focus on commission.

I can't recommend this tool at all, because the premise on which it is built and has worked from the beginning is wrong, specifically 'We don't yet cover returns, they often aren't cheaper'

"We don't yet provide a specific split ticket search for returns. Firstly, this is because two singles are often (not always) cheaper, and secondly, well - it's a big job, and next on our list.

Yet if you want a return, simply do a single ticket search, then see if splitting at the same stop works on your return. It's not perfect, but often works."

This is really terrible advice, the fact is that returns are pretty much ALWAYS cheaper than two singles, assuming of course that you are making a return journey and that a return ticket is available.

Pushing people to buy split singles that aren't fool-proof (e.g., if the trains don't always stop at that station) and are more expensive than returns really isn't great advice.

I couldn't get it to return any results, but the ones in this thread:
http://www.moneysavingexpert.com/split-cheap-train-tickets/results/52b19307b62c71.69760931 are wrong:

£5 + £18.90 = £23.90 single, in standard

but the cheapest advance fare for the through journey is £44.60, and the advance first is £68, not £73.50.

Also the results it returns are potentially very much worse than using a normal booking site.

Example, if you search on National Rail Enquiries for Reading - Leeds on Christmas eve at 10:15, you can easily see that there are single fares from £50 up to around £110

http://ojp.nationalrail.co.uk/service/timesandfares/RDG/LDS/241213/1015/dep
and maybe if you check, you can find a better price.

Searching Tickety Split, however, offers this:

http://www.moneysavingexpert.com/split-cheap-train-tickets/results/52b1d44962f851.49143000

a split reducing £109 to £104.80 (or £106.30 when you add on the booking fee), which is still double what the cheapest fares on NRE are going for, and a very poor deal compared to the flexible super-off-peak return for £110.

Also some of the claimed 'successes' are very dubious, given that split advances have the potential to go very expensively wrong if you accidentally buy the wrong ticket, the buy-through system whereby the user has to select the right train is clearly going to result in some people buying tickets for the wrong train, e.g., this:

http://www.moneysavingexpert.com/split-cheap-train-tickets/purchase/52b1d55a6de382.81008583

where the promised saving is only 50p - not worth it given the potential for error (and again, a flexible return for £110 is much better than buying two advance singles at ~£55 each).

The Ipswich - Swindon example is a nice saving, one that works because it breaks down from an AGA Advance to a FGW Advance by splitting at Advance, but this is far from the fool-proof automated bargain-hunting tool that it implies, it's actually quite a specialised function, and in some aspects it works against the consumer's interest, because it encourages him to buy tickets in advance on a website charging booking fees when for most train journeys buying in advance is a bad idea (since your plans might change, and in for most journeys the cheapest tickets are walk-up).

It still encourages, via emotional blackmail, "buying tickets this way helps secure the tool's future" people to buy walk-up tickets in advance, even where no saving is offered.

Until this tool at least lists the return fare (which is relatively trivial to show), it's misleading at best.

 

[MSEAdam]

Hi soil,

Thanks for your comments. I'll leave the MSE-specific points to be discussed in another forum.

In terms of the tool itself:

We search specific tickets for the time set - so your request was for a Reading->Leeds ticket for 10:17.

The next time train available was 10:26 which is why we selected this train. Currently picking and/or showing alternative times is out of scope of our initial release.

For the IPS->SWI route for that specific date & time, the first class ticket is cheaper than standard, so maybe not the best example for people to discuss:

Of course we want to include returns, but we're aware it's a complicated proposition, because we'd need specific times for both journeys, and to therefore determine which calling stations are common.

Thanks for the feedback all, keep it coming!

Adam

 

[tractakid]

I think that there are 2 main approaches to split ticketing.

The first is 'cheapest possible' and the second is 'anything that makes this specific itinerary cheaper'.

Of course, like many things, the best solution is a compromise. Which is where tools will always struggle to excel.

ticketysplit seems to favour the latter option. In which case, I think the user should be able to view itineraries and select the one(s) they wish to find a split for. At the moment the tool leaves the user blind as to which itinerary will be split if they have not done prior research into potential itineraries, and then the tool may turn up a different one than they expect anyway.

 

[MSEAdam]

Thanks tractakid - we'll be feeding this back into our development process.

Adam

 

[sheff1]

I agree with those who highlight the lack of return ticket information a major drawback. Even a comparison list of return ticket pricess would be helpful.

Most 'journeys' involve someone leaving their home address and then, at some point, returning home. The tool as it stands almost encourages people to book two (or more) singles when a straightforward return would be both cheaper and more flexible.

Example Searching for Sheffield to Bristol, out 1000 on 20 Dec & back 1000 on 27 Dec offers

Southbound - £56.50 (split at Derby)
Northbound - £77.50 (split at Cheltenham)

So a total of £134, when an Off Peak Return on the same trains is only £87.60. The latter ticket would accommodate a change of travel times, whereas the MSE offer would tie you to specific trains.

Searching again from Derby (which I think is what the text on Ticketysplit is suggesting I do for a return) gives a two way total of £78.50*. Add Sheffield to Derby and back onto that and again the total cost will be more than the through Off Peak Return.

* again, more than the Off Peak Return for that section

 

[soil]

Of course we want to include returns, but we're aware it's a complicated proposition, because we'd need specific times for both journeys, and to therefore determine which calling stations are common.

Not really, you can just list the through fare for comparison purposes, without attempting to split.

Also the system attempts to choose specific times at the moment, and in many cases the calling points would be similar. The work flow I would go for is:

* user selects outbound and (optionally) return date, and times
* system returns a list of train times and tickets, identical to National Rail Enquiries, example:

http://ojp.nationalrail.co.uk/service/timesandfares/IPS/SWI/150114/0900/dep/220114/1115/dep

The user can then view and chooses the most convenient trains from the list.

Since you then have a specific outbound and return journey, you can perform a split check separately for each leg (the functionality for this is already built), and display the costs alongside the cheapest flexible return fare for the through journey (a simple lookup).

This is adding no additional splitting logic, it's just showing the fares.

Also I would tend in any case if a user only selects a return journey to list the cheapest return fare valid on that train, because he may not realise that the return ticket is about the same price.

So the final output for say Reading - Leeds, single, would be (the advance fares will obviously vary by date/time):

* Walk-up single fare: £109
* Walk-up return fare: £110
* Advance single fare: £54.50
* Single advance split fare £52.50

For the return, I would list the following options:

* Walk-up return fare: £110
* Advance single fare outbound and return: £54.50 + £43.50 = £98
* Advance split fares = £52.50 + £41.50 = £94

I don't like the results for Reading - Coventry:

http://www.moneysavingexpert.com/split-cheap-train-tickets/results/52b1ea09678b94.97428684

£41.50 Advance Single is compared with £11.60 CDS + £9 Advance = £20.60

I would prefer to see this output:

'Non-split'
* Walk-up Single £48.30
* Advance Through Single £41.50

'Split':
* Split walk-up-only singles: £11.60 + £11.60 = £23.20 (this is the actual walk-up fare, split at Banbury)
* Split singles with Advance: £20.60

The current output implies that you need to get an Advance to get a big saving on this route - you don't, the big saving comes from splitting at Banbury.

And again the lack of a return is a problem here, if we are coming back the same day, the cost is only:

£11.70 + £13 = £24.70, which is far cheaper than buying two sets of two singles at £20.60 each

In terms of return splits, I don't see it as hugely complicated - once the user has selected his trains you just find the common set of calling points, and try splits at each. This approach shouldn't increase complexity much, because the number of potential splits would be no higher than the number of splits on the single journey (since the common calling points cannot exceed the number of calling points on any one journey).

There is I suppose a slight additional complexity, in that you do have to compare both advances + returns and advances and singles:

E.g.,
Swindon - Ipswich, we can potentially buy tickets

Swindon - Didcot, Didcot - Ipswich, Swindon -Reading, Reading - Ipswich, Swindon - Paddington, Paddington - Ipswich, Swindon - Colchester, Colchester - Ipswich, Colchester - Ipswich, Swindon - Manningtree, Manningtree - Ipswich, and so on.

For the single journey, it's just the cheapest sum of any two such pairs.

But for the return, we can do:

(Swindon - Manningtree Advance, Manningtree - Ipswich single) * 2
or we can also consider

(Swindon - Colchester advance) * 2 + Colchester - Ipswich return

Given the constraint that we won't split into more than two tickets, given

A - B via splitting point X

A-X single (cheapest fare either advance/walk-up) + X-B single (ditto)
A-X single + X-B return
A-X return + X-B return
A-X return + X-B single

So there are four times as many comparisons I guess.

 

[maniacmartin]

OK, so I know first hand how nasty rail fare data is to work with, and how ATOC/RSP are not happy with too many searches for AP quota availability, so I do empathise here, but:

First, the site should be able to add the tickets to the trainline basket automatically. The user interface is nasty, and doesn't work at all for me - I can't figure out how to buy the tickets at all (see screenshot attached - same in Firefox and Chrome)

The homepage information about returns is deceptive. A walk-up through return ticket is often cheaper than two batches of singles. Saying they "often aren't cheaper" is a bit misleading. At the very minimum, it should show the walk-up return price to compare with.

The names of the railcards are confusing. "GOLD" doesn't obviously sound like a Network Gold card.

The grammar is bad. There are spurious apostrophes and missing words everywhere.

Statements like "you'll need to use thetrainline.com to book them" (emphasis mine) when Trainline levies a booking fee that others don't is misleading and goes against everything I thought MSE stood for, namely "editorial line (the things we write) is NEVER impacted by these links"

I disagree with your blurb about delays. As all tickets relate to the same journey, the passenger should be covered on later trains if earlier ones are delayed, as per the NRCoC.

 

[tannedfrog]

Well, for a start it's worth understanding the difference between advance and advanced. The split ticketing may well be advanced (above average; clever) but the tickets themselves are Advance tickets (before the event). Trivial but important if you want it to look right.

Advanced can also mean "before the event". Minicabs often say "advanced bookings only".

 

[yorkie]

I disagree with your blurb about delays. As all tickets relate to the same journey, the passenger should be covered on later trains if earlier ones are delayed, as per the NRCoC.

Absolutely. The NRCoC makes it very clear you can use "two or more" tickets for "one journey", and in the case of Advance tickets the conditions are clear that "if delays occur while travelling" you can take the next train. (You don't need to get a ticket 'endorsed' though if you have time to do so, without missing a train, then it can save hassle).

This is also clarified in the Advance Fares FAQ, which is a page in The Manual. The public don't have access to The Manual but we have been given permission to reproduce the relevant page as a PDF attachment in our Fares & Ticketing Guide (1.1.3 Advance tickets).

Advanced can also mean "before the event". Minicabs often say "advanced bookings only".

According to some minicabs, there are places like "St Pancreas", so anything they say should be taken with a pinch of salt.

Advance and advanced have very different meanings.

Advance tickets should definitely be called by their correct name, and I would recommend completely avoiding the term "advanced". If walk-up fares are purchased "in advance" I would suggest using alternative wording such as "prior to the date of travel" to avoid any confusion.

But there's no way the word 'Advanced' should ever be used in this context.

Here's some sources:

• http://forum.wordreference.com/showthread.php?t=2385581
• http://public.wsu.edu/~brians/errors/advance.html
• http://www.accu-assist.com/grammar-tips-archive/GrammarTip_homophone-potpourri-8.htm

On the topic of avoiding confusion, the term "open return" should be avoided, as confusion can lead to possible prosecution.

 

[Mojo]

Advanced can also mean "before the event". Minicabs often say "advanced bookings only".

Maybe they only accept bookings through their website or a Smartphone App, and are trying to state that they do not accept bookings by ye olde telephone?

 

[IanD]

Absolutely. The NRCoC makes it very clear you can use "two or more" tickets for "one journey", and in the case of Advance tickets the conditions are clear that "if delays occur while travelling" you can take the next train. (You don't need to get a ticket 'endorsed' though if you have time to do so, without missing a train, then it can save hassle).

This is also clarified in the Advance Fares FAQ, which is a page in The Manual. The public don't have access to The Manual but we have been given permission to reproduce the relevant page as a PDF attachment in our Fares & Ticketing Guide (1.1.3 Advance tickets).

To be fair, the ticketysplit site does say that with split Advance (and Anytime) tickets you should have no problem if delayed. It just points out there might be issues if you have a string of time restricted tickets and a delay early on causes you to be on a time barred train.

Personally, I agree with maniacmartin and Yorkie. I don't see the difference between having a string of Advances and a string of any other ticket types - you should be covered if the delay is not your fault (i.e. your first train started out early enough to complete the journey within the valid times of the tickets held and delays occurred after that). Whether on-train or gateline staff will agree with that is an entirely different matter. You'd at least have to show all your tickets rather than just the one for the leg you're on.

 

[Lockwood]

Hi Lockwood,

For the journey you searched, walk-on fare provides a best split @ Basingstoke.

When checking advance fares, splitting at Basingstoke causes the fare to rise from £39 to £40.50.

Adam

That's a completely different route than I'd normally use (using the Portsmouth-Cardiff train for the bulk of it)
Walk on with the 3 splits comes down to ~£35.

Would we ever see a basic multisplit on MSE?
If I go from A to B, and there's a split at C, have a check of AC and CB to see if they have a cheaper split.
If A-D-E-B is cheaper than A-C-B, this won't be picked up using this dissection method, but the trade-off is that the dissection method is a lot simpler to code than testing everything, and more efficient for the server

 

[MSEAdam]

soil - thanks for your extensive feedback, really useful. I'll ensure this goes back to our team to discuss.

Re delays, we've consulted with ATOC directly, so have added in related text to ensure the user is adequately informed.

Apologies for the use of "advanced" instead of "advance" - amazing what a difference a single letter can make.

In terms of multiple splits, that's not our current intention as we have a trade-off between finding great savings, but also keeping the system simple for all of our users.

Great feedback all - please keep it coming!

Adam

 

[maniacmartin]

It doesn't look like you've changed any of the text to me.

 

[MSEAdam]

Hi maniacmartin,

We'll be having a review here to discuss all the feedback made.

Were there any other particular sentences or issues that you'd like to raise?

Thanks,
Adam