Replying to Green/Yellow/Red

You're assuming that past data is stored in a database accessible to the program that makes the service predictions.

Given that the website only gives access to a few days worth of data, that's likely a bad assumption.

Regardless of how easy you may think it is, it doesn't come free, and you may have noticed that Amtrak isn't exactly swimming in cash.

Let's say it is in a different database. It still isn't hard to copy that data back onto the database used for service predictions.

Honestly, I don't think that the reason they do optimistic/best-case predictions is because it's too hard to program. More than likely, it's because they don't want to predict the train to come in (or more importantly, leave) later than it actually does. If Amtrak stated that the train is estimated to leave at 10:45 AM on the status page, and it actually leaves at 10:30 AM, someone would complain because they missed their train because the Amtrak website said that the train won't leave until 10:45 AM. Even if the schedule (and their ticket/reservation) says that the train leaves at 9:45 AM and Amtrak posted that this is an estimated time and that trains often make up time en route, so it may be there (and leave) sooner than shown here, etc. etc. Someone will complain.

And that's why we can't have as many nice things.

You're assuming that past data is stored in a database accessible to the program that makes the service predictions.

Given that the website only gives access to a few days worth of data, that's likely a bad assumption.

Regardless of how easy you may think it is, it doesn't come free, and you may have noticed that Amtrak isn't exactly swimming in cash.

Easy to say. Not so easy to spend the money to write a computer program to do what you're asking.

You are aware that these predictions are automatically generated, not by a person waiting to tell you how late they think your train is going to be, right?

I would think with as few trains they have a couple of people could monitor it and update it correctly...

That'd be nearly $100,000 a year to hire two people to man it and make even somewhat correct predictions. (They may only get paid $35,000-$40,000 a year, but Amtrak benefits aren't cheap either.)

However, almost every (object-oriented) programming language has a math class that would have an average function in there. SQL also has an average function, so it could be automatically calculated by the database. It's nearly trivial to call a variable from a database in an object-oriented programming language.

Even if they don't have the time traveled between stations already calculated for each day (or at least have a program calculating the difference), I think that can be calculated in most database programs/languages if they already have the arrival/departure times in a database (which they do for at least five days). I can't imagine it would take terribly long for even a single programmer to make that change...a month or two max, including testing?

Easy to say. Not so easy to spend the money to write a computer program to do what you're asking.

You are aware that these predictions are automatically generated, not by a person waiting to tell you how late they think your train is going to be, right?

I would think with as few trains they have a couple of people could monitor it and update it correctly...

That's far more complicated than the current system. Spending lots of money for very little gain. Amtrak would need a solid business case for making such of a change, and I don't see one.

Maybe tempering expectations of train service? I'm not sure if there'd be a strong business case, per se, but there could be a "truth in advertising" sort of claim for doing it.

Furthermore, it wouldn't seem to take terribly long to program an algorithm to do that. The average would be easiest, but even a percentile wouldn't be terribly difficult, even if we're starting from scratch. My guess is that there's already some math going into it, so it'd be simply replacing those values with a calculated average or percentile. A computer could be easily programmed to do that (unless they're using a weird programming language currently.)

What is funny is I checked today's Cardinal which was already an hour plus late leaving Indianapolis this morning and yet according to the Amtrak web site itself, says it's "estimated" to be into NYC 9 minutes EARLY. Are you kidding me, I doubt that train has ever been early at any of it's stop and today's train hasn't even hit it's normal hour plus delay in the WVA/VA area. What is Amtrak thinking...

First off, Amtrak's ETA is based on the idea that the train won't encounter any more delays. So if the train is 43 minutes late right now, it assumes it will remain 43 minutes late. I suppose Amtrak could just assume the train was going to get even later (by some random amount), but that would be just as likely to be inaccurate.

EDIT: Clarity.

This is what I also find kind of funny. Anyone who can look at data will see that the chances of the train not encountering any more delays are if not 0%, then darn close. I doubt there has been a trip east via the Cardinal that hasn't had issues with delays at the same exact place each trip. This is a "known" factor so why not "factor" it in.

I am kinda surprisded how the federal gov't is all over the airline industry over delays, planes sitting at gates and such to the point where fines are being handed out, but I guess the same concern isn't there for Amtrak. Maybe it's train folk don't complain as loud or maybe Congress is just happy the trains run at all. But I doubt any airline could offer a route that has the lack of on time performance that the east bound Cardinal offers without feeling the wrath of consumers/congress.

My guess is that airlines are more scrutinized for these reasons:
1. Many, many more people use them.
2. There's a higher expectation of OTP. If my flight is only supposed to take three hours, then a 3 hour delay doubles my time in transit. However, if I have a 24 hour trip, a 3 hour delay, while annoying, could be "explained away" more easily.
3. Planes just aren't set up to handle delays as well as trains are. If my train is delayed, I can still have my electronics on (granted, this may also be true with airlines, but it'd be 100% certain with Amtrak), I can walk around to different cars, I can buy a snack (unless the train runs out of food), etc. Basically, instead of being stuck for three hours buckled up on a tarmac, likely without food being provided/available for purchase (and certainly without food-on-demand being provided or available for purchase) and possibly without electronic devices (since they may need to be kept off in case a quick slot opens up and there's not enough time to make a "shut your electronics off" announcement), I'm on a train where I can walk around, usually buy food (assuming the train hasn't run out of food), and I can text/call the people I'm meeting at my destination that I'll be late. I'm still late on the train, but at least I'll stay sane throughout. =]

Easy to say. Not so easy to spend the money to write a computer program to do what you're asking.

You are aware that these predictions are automatically generated, not by a person waiting to tell you how late they think your train is going to be, right?

What is funny is I checked today's Cardinal which was already an hour plus late leaving Indianapolis this morning and yet according to the Amtrak web site itself, says it's "estimated" to be into NYC 9 minutes EARLY. Are you kidding me, I doubt that train has ever been early at any of it's stop and today's train hasn't even hit it's normal hour plus delay in the WVA/VA area. What is Amtrak thinking...

First off, Amtrak's ETA is based on the idea that the train won't encounter any more delays. So if the train is 43 minutes late right now, it assumes it will remain 43 minutes late. I suppose Amtrak could just assume the train was going to get even later (by some random amount), but that would be just as likely to be inaccurate.

EDIT: Clarity.

This is what I also find kind of funny. Anyone who can look at data will see that the chances of the train not encountering any more delays are if not 0%, then darn close. I doubt there has been a trip east via the Cardinal that hasn't had issues with delays at the same exact place each trip. This is a "known" factor so why not "factor" it in.

I am kinda surprisded how the federal gov't is all over the airline industry over delays, planes sitting at gates and such to the point where fines are being handed out, but I guess the same concern isn't there for Amtrak. Maybe it's train folk don't complain as loud or maybe Congress is just happy the trains run at all. But I doubt any airline could offer a route that has the lack of on time performance that the east bound Cardinal offers without feeling the wrath of consumers/congress.

That's far more complicated than the current system. Spending lots of money for very little gain. Amtrak would need a solid business case for making such of a change, and I don't see one.

What is funny is I checked today's Cardinal which was already an hour plus late leaving Indianapolis this morning and yet according to the Amtrak web site itself, says it's "estimated" to be into NYC 9 minutes EARLY. Are you kidding me, I doubt that train has ever been early at any of it's stop and today's train hasn't even hit it's normal hour plus delay in the WVA/VA area. What is Amtrak thinking...

First off, Amtrak's ETA is based on the idea that the train won't encounter any more delays. So if the train is 43 minutes late right now, it assumes it will remain 43 minutes late. I suppose Amtrak could just assume the train was going to get even later (by some random amount), but that would be just as likely to be inaccurate.

I would wager that a more accurate estimate could be given by using some sort of average time that it takes to get there, based on past performance. A lot of times they seem overly optimistic about the train times, and have almost no basis in reality. Heck, even just using a 75th percentile (as in, they assume that the train will run faster than 75% of trains in the past along that route) would make it more realistic.

What is funny is I checked today's Cardinal which was already an hour plus late leaving Indianapolis this morning and yet according to the Amtrak web site itself, says it's "estimated" to be into NYC 9 minutes EARLY. Are you kidding me, I doubt that train has ever been early at any of it's stop and today's train hasn't even hit it's normal hour plus delay in the WVA/VA area. What is Amtrak thinking...

First off, Amtrak's ETA is based on the idea that the train won't encounter any more delays. So if the train is 43 minutes late right now, it assumes it will remain 43 minutes late. I suppose Amtrak could just assume the train was going to get even later (by some random amount), but that would be just as likely to be inaccurate.

As for how the Cardinal could be an hour late at Indianapolis and still be early into New York, that's because of schedule padding. That happens on every route. The Capitol Limited, for example, is scheduled to take 60 minutes to get from Rockville to Washington. But it's only scheduled to take 24 minutes to get from Washington to Rockville. The MARC afternoon train #877 (the Martinsburg Express) gets from Washington to Rockville in 26 minutes, with an intermediate stop at Silver Spring.

So, what you can see is that the actual travel time from Rockville to Washington is only about 25 minutes. If the Capitol Limited arrives in Rockville on time, it will get into Washington (without further delay) 35 minutes early. If it arrives at Rocville 30 minutes late, it will still get into Washington 5 minutes early. The same can be true of the Cardinal. When I took it last year, we lost an hour in South Chicago due to a medical emergency, but we still ended up getting into Washington only 5 minutes or so late.

But you are right that the Card has awful on-time performance.
#50 was not on time once in May 2012, and has been on time only 20.7% of the time in the last 12 months.
#51 was on time 38.5% of the time in May 2012, and has been on time 64.1% of the time in the last 12 months.

EDIT: Clarity.