Comment by chimeracoder | Hacker News Reader

Comment by chimeracoder | Hacker News Reader

I live in NYC, and I follow transit news very rigorously.

This article underplays the extent of the problem, which is that the information they give is just not actually correct.

First of all, the MTA made a big deal about rolling out the subway countdown clocks across (almost) all the stations. Except, for the old BMT lines[0], the clocks are just... not actually accurate! They're literally just publishing the scheduled times, and almost never actually update it to correspond to a delay[1][2]. Delays are dramatically worse across every single subway line than they were ten years ago[3], so by this point, those scheduled times are almost completely useless. Thankfully, the BMT countdown clocks are a different style from the IRT ones, so as a rule of thumb: if it's black text on white background, then it's completely useless, and if it's green LEDs on black background, then it's probably-somewhat-useful, but still not necessarily up-to-date.

Even for planned service changes, they publish information in unbelievably obtuse ways[4]. Yes, they're technically correct. But the point is to communicate things clearly, not to publish statements that are defensibly not-wrong in a court of law.

Just last month, they decided to stop service for the express trains midtown (or run it in two segments - I never figured out which). They didn't tell anyone, so an entire train full of passengers went two stops downtown and two stops back uptown. Because they also held the train before one of the stations, we literally took half an hour to end up right back where we started[5].

[0] For the most part, any line with a letter instead of a number, like the A/C/E instead of the 1/2/3

[1] https://twitter.com/2AvSagas/status/950863996323794944

[2] https://twitter.com/2AvSagas/status/948727423444639745

[3] https://www.nytimes.com/interactive/2017/06/28/nyregion/subw...

[4] https://twitter.com/2AvSagas/status/933826118364299266

[5] https://twitter.com/chimeracoder/status/1000422530736377857

I totally agree. I was stuck at the Broadway Junction station this weekend waiting for the L. Every 5 min or so, both the screen and the announcer would claim it is now arriving. We were there for 35 min. I was supposed to meet some friends and initially I was updating them that it's about to get there. Eventually, they were almost pissed because they thought I was lying. Had to send pics.

I really really get frustrated by the inaccuracies. I'd rather they just let me know straight up that it's 30 min away and I make alternative arrangements.

> Except, for the old BMT lines[0], the clocks are just... not actually accurate!

I can't say I agree. I've been a daily commuter on the ACE, 1, and NQR/BD trains since the introduction of the clocks and found them all to be wonderfully accurate.

Yeah, it's almost surely a case of remembering the handful of times when the clocks were inaccurate because they caused you trouble, while forgetting the hundreds of times when they worked correctly. They do work correctly most of the time, and the A division works even better than the B division, as you'd expect. (the opposite of what OP claimed.)

> Yeah, it's almost surely a case of remembering the handful of times when the clocks were inaccurate because they caused you trouble, while forgetting the hundreds of times when they worked correctly.

It absolutely is not availability bias, as demonstrated by the concrete data included in one of those links.

But even if it were - the whole point of having those signs is to alert you about the actual state of the subway, factoring in delays or service changes. Otherwise I could just download the PDF of the scheduled train times from the MTA's website and look at that. (Yes, those exist. There's almost no reason any person would ever need to look at them).

> They do work correctly most of the time, and the A division works even better than the B division, as you'd expect. (the opposite of what OP claimed.)

No, that's not the opposite of what I said. The A division is the IRT lines. Those are the ones that use the ATS system (which was outdated even at the time it was rolled out), and that's why the IRT lines (the ones with the LEDs on black background) are at least somewhat useful. Even if the train is delayed from its normal schedule, the times posted there are supposed to be somewhat accurate because it's using the ATS signal data.

For the B division (the BMT lines), you might as well be using the static PDFs that post the idealized train schedules[0]. I am not exaggerating; that's literally the data they are displaying. Once in a blue moon, for very serious delays such as a train accident, it will tell you that a train is "delayed" instead of telling you the updated ETA, but that doesn't help, because you want to know how long the train will take, rather than just the fact that it's "delayed"[1]. Most of the time, though, it won't actually update the time at all, and it'll keep displaying the wrong estimate until the train is literally about to pull into the station (at which point it'll skip straight to "0 minutes away").

The reason they don't tell you more information is because that information doesn't exist. They don't know where the trains are, beacuse the BMT lines use outdated signaling technology that's decades old, and the MTA has stonewalled all efforts to upgrade them[3].

So, in short, the BMT countdown clocks are useless because:

* They only publish the scheduled times, not the actual estimated arrival times

* Only about 50-60% of BMT trains arrive according to the times posted on those schedules

* For the remainder, it usually displays the wrong information, because it doesn't have any way of knowing that the information is out-of-date

[0] http://web.mta.info/nyct/service/pdf/tacur.pdf

[1] Saying a train is "delayed" doesn't give me any useful information, because I already know that about half the trains are delayed. And since it doesn't tell you what the scheduled ETA was, telling me that it's "delayed" from the scheduled ETA is also pretty useless[2].

[2] If the train is scheduled to come in 20 minutes, but it's been delayed by five, that's different from if it's scheduled to come in 3 minutes, but has been delayed by 10. Those two situations are displayed identically on the BMT countdown clocks, with no way to distinguish between them.

[3] The official MTA party line is that they want to wait and do the really expensive, comprehensive, state-of-the-art overhaul. In reality, that means that they've been saying this for nearly thirty years, and have actively fought all attempts to provide quicker and cheaper ways of achieving the same end results.

> * They only publish the scheduled times, not the actual estimated arrival times

Do you have a source for this? I'm positively certain it's not true. I time my commute every morning with the online subway clock that matches exactly with the one in the station -- I leave my apartment when the Q is ~8m away and it always arrives shortly after I get to the platform. I've only encountered one situation where the clock was unacceptably wrong: a downtown A train at 59th that was 1m away for over 5m. That happened over a year ago and I still remember it.

Similarly, when delays happen the clock does get updated accordingly. Today was a good example -- 24m for the next N train during commuting hours :/

I think you may be confusing the countdown clocks with the kiosks. Once upon a time they did display the scheduled time. I'm not sure if they still do but the above head displays are without a doubt more accurate. Anecdotally speaking, the margin of error seems to be roughly +/- one minute. I'm aware that's anecdotal evidence but I simply don't believe that somebody could take the subway on a daily basis and think the countdown clocks are that inaccurate.

> I think you may be confusing the countdown clocks with the kiosks

No, I'm not. I know the difference between the kiosks and the overhead clocks.

> I simply don't believe that somebody could take the subway on a daily basis and think the countdown clocks are that inaccurate.

Benjamin Kabak, whom I linked above, has been covering the MTA and transit in New York for over a decade, in addition to being a daily subway rider. Offhand, I can't think of a single independent journalist who has more comprehensive knowledge of the minutiae of the NYC transit systems than he does, let alone a more established track record of documenting not only the visible problems, but the factors that create those problems. If that, combined with the other data provided, isn't convincing, I don't know what else could be.

You haven't really provided any data. A few random tweets about incidents, only one about the countdown clocks where the next response offers a probable explanation.

You're trying to sell me on a different version of reality, here. You're telling me a system I rely on every single day is completely unreliable.

Do you live in NYC? Do you take the trains? Go put your theory to the test. Go sit at at a busy train station for 20m and watch the countdown clocks tick down until the train comes. I promise you'll be pleasantly surprised. The countdown clocks are just about the only part of the MTA system that works.

> You're trying to sell me on a different version of reality, here.... Give the MTA their single, well-deserved victory. Put your theory for the test. Go sit at at a busy train station for 20m and watch the countdown clocks tick down until the train comes. I promise you'll be pleasantly surprised.

I'm not trying to sell you on anything. In fact, I'm not even interested in continuing this discussion. I've provided you with a reference to the most well-respected journalist covering the technical details of the MTA and NYC transit. If you're really interested in learning more, it's not hard to do some basic Googling and find their extensive coverage and analysis of this problem.

> Do you live in NYC? Do you take the trains? I'm so confused -- are you astroturfing for some reason?

Just because someone presents information that contradicts your anecdotal experience, that doesn't mean they're astroturfing. Though at this point, if you're going to start slinging bad-faith accusations with no basis, it's pretty clear this conversation is going nowhere.

> it's pretty clear this conversation is going nowhere.

Yes, obviously, because you keep saying "look at the data" while providing none of relevance. Nobody's arguing that the MTA is rife with delays and poor planning, we're talking about the countdown clocks specifically.

The countdown clocks do not show the scheduled time. They use Bluetooth receivers to physically track the trains [1]. Your fundamental premise is wrong.

I'm honestly curious -- do you Uber everywhere? How often do you take the train?

1. http://www.thetransitwire.com/2017/09/01/mta-uses-bluetooth-...

I understand your frustration. Precision without accuracy can be absolutely infuriating.

Imagine a friend saying "I'll be there in 5 minutes" and showing up 2 hours later. Just maddening.

> Imagine a friend saying "I'll be there in 5 minutes" and showing up 2 hours later. Just maddening.

Sounds like L.A.!

Reply on news.ycombinator.com