Noticed this on a consuming app and confirmed with the API - a train arriving at Northfields apparently arrives at both platform 3 and 4 going eastbound, though this is obviously not possible
Hello and welcome to the forum!
Where our API does not know which platform a train will call at, a prediction will be generated for each possible platform.
So in this case, it doesn’t know whether Platform 3 or Platform 4 will be used, therefore there is a prediction for both.
Quantum mechanics in action! Is Schrodinger’s cat on board?
I’m reminded of the announcement “Will passengers please use all available doors…”. What am I? A photon?
@LeonByford thanks for your quick reply! Just to appease my curiosity, why is the API data ‘worse’ than the station boards?
Station boards typically get their data directly from the signalling system – the quality of data varies across the network, but the predictive capability can sometimes be quite advanced.
The API uses a different method to generate predictions, and does not have access to all the data that the station boards do.
That said, we do have a couple of initiatives to try to improve the quality of predictions we provide as open data (as well as through our website and app).
That’s not too bad for adjacent platforms but a real problem for many passengers where multiple stairs and/or a long walk are involved in a last minute platform swop. Much better to show the time with the platform as something like “TBA” so one can wait at an optimum point until the correct platform appears.
That’s a good point. I think that’s something open data users can already do in their applications (by seeing if there are multiple arrivals with the same ID, for example), but we can think about doing that on our website and app.