Then I went into ways to fix delays: they can either build in measures to make the delay less hard felt once it gets started (i.e. schedule padding) or stop the delay from happening in the first place. We ruled out massive schedule padding as a solution to the problem, as that can often do more harm then good if things get too out of hand. Therefore, we have to stop the original delay from happening in the first place.
But how do we do that? Well, lots of times the mess of delays usually stems from one single train that is almost guaranteed to be off schedule and therefore guaranteed to mess things up for the rest of the day.
After the Part 2 post appeared on this site, Sunny Zheng sent me his collection of East End OTP records that he has been compiling for the last three summers. Below is a break-down of every train that runs on the east end (including the Greenport Scoots) and how late they have been on each occasion. As usual, a train is considered "on time" if it arrives at its final destination within 5 minutes and 59 seconds of its scheduled arrival time.
I took the three Excel files he sent me and compiled them into this one master spreadsheet with data dating back to 2011. I uploaded the file onto Google Drive and embedded it in this post below. You can scroll up and down or to the right to see more data (from top to bottom it goes through Friday's trains, then Saturday's, then Sunday's). You can compare the stats to years past by changing the year at the very bottom of the embedded box:
The rows at the very bottom of each train tell you three things: the average delay, the average lateness, and then there is a third value that I added: the total delay minutes (colored in yellow). This figure, in my opinion, is the most telling. This figure tells you which trains were the most late the most often.
First, going through the statistics for Fridays, in 2011 and 2012 the Cannonball (train 2798) was the worst performing train for those two years, racking up 305 and 368 delay minutes in those two years, respectively. This year's Cannonball is performing much better than the Cannonball's of 2011 and 2012. Through the first 12 Friday's of 2013, the train has only drawn up 106 delay minutes. Comparatively, in the first 12 weeks of 2011 and 2012, the Cannonball had 280 and 271 delay minutes, respectively.
So going from 280 and 271 minutes to only 106 in one year is a big improvement. Through the first 12 weeks of this year, the Cannonball has accumulated less than half the amount of delay minutes than in years past. That's a big improvement!
And the overall delay minute numbers have improved as well because of this. Through the first 12 weeks of 2011 and 2012 the grand total amount of delay minutes accumulated by all trains on the Montauk Branch on Friday Evenings was 1,120 in 2011 and 918 in 2012. So far in 2013, we've only accumulated 477 delay minnutes! Nearly HALF of year's past!
All things considered, changing one thing about one train seems to have cured the entire branch on Fridays. The amount of delay minutes experienced by passengers have been practically cut in half thanks to the changes to the Cannonball made this year.
So they've found the solution to the Friday Montauk Branch Delay Mess, what about Sunday's?
Again, we turn to the total delay minutes figure to find the culprit. For the most part, the 1:30pm Montauk westbound departure (train number 8703 in 2012 and 2011, train 8705 in 2013) has been one of the worst performing Sunday trains. In the last three years, the 1:30pm departure has been the leader in the westbound delay minutes contest.
In 2011 the train caused 322 minutes of delays, in 2012 it contributed 307 delay minutes, and so far in 2013 it has caused 109 minutes of delays (not including the US Women's Open weekend for which it got an OTP waiver) and it has been canceled at Speonk once.
Looking at the schedule for 8705, everything seems fine until Southampton. At Southampton, the train is scheduled to meet train 8710, the 12:10pm train from Jamaica due Montauk at 3:00pm. And do you want to take a guess as to which train is the second worst performing on Sunday's? Yup, train 8710.
So we have a meet between the day's worst and second worst performing trains in Manual Block territory, and they have less than 15 minutes to do it! That doesn't seem like a recipie for success now does it?
And what's more, 8705 is frequently plagued by equipment trouble. On one of the first weekends of this summer 8705 had such bad equipment trouble that it was canceled at Speonk. It seems that 8705 is frequently stuck with a troublesome set of equipment. I'm not sure if that is either just rotten luck or something that could be avoided.
Going further, 8710, which is frequently delayed, is due to arrive Montauk at 3:00pm. The next westbound train, train 8709, is scheduled to leave Montauk at 3:30pm, just 30 minutes later. There have been a number occasions this year alone where 8710 has been so late that it has prevented the net westbound train, 8709, from being able to depart Montauk on-time. There have even been two occasions where 8710 was so late that it even delayed the next train after 8709, 8711, the 4:00pm Montauk departure, from leaving on-time.
So we might have found our culprit(s). But how can this get fixed? Well, in my opinion, meets in Manual Block territory should be avoided at all costs. Therefore, the meet should either be moved someplace else or eliminated all together.
That is a tricky proposition, however. In order to move the meet out of MBS territory they have to go all the way to MS (JJD), but that is a very long way away. But moving the meet elsewhere would involve adjusting the timing of one or both of the trians. 8710 is practically the last eastbound train of the day on weekends (that leaves at 12:10pm--the next train doesn't leave until 9:41pm). You want to avoid widening that gap by moving 8710 earlier, but then again you can't move 8705 too much later because there are a lot more trains going westbound and you can't have them running too close together.
But is it possible to eliminate the meet all together? Ehh, probably not.
This predicament is going to need more thought on my part. If you have an idea, leave a comment, but for now, we'll have to wait for the next installment of this series!
For more information on this topic, see Mending the Montauk Branch Delay Mess Part 1 and Part 2.