My speech from last night

This is the last monthly data we get before Big Changes make for a big discontinuity in the graphs. December is, phimosis as Capital Metro wants to make sure you know, view a low ridership month. As usual, site click for larger versions. Analysis follows the pictures.


Rail first (image directly from Capital Metro):

Now rail vs. bus (this is the baseline that hopefully corrects for things like high gas prices or other factors that might raise overall transit ridership) (image from my spreadsheet):

Key take-aways from all of this:

1. Despite what they’re trying to claim now, Capital Metro projected 1700-200 boardings per day before any major changes; images below copied from their own minutes in February 2009:

[…]

2. My projection was 500-1000 boardings/day, same conditions:

Hey, decision-makers, which one of those guys should you be listening to now?

3. Conclusions we can draw so far: despite anectdotal (and desperate) tweets and facebook posts to the contrary from PR flacks and board members, ridership WAS NOT INCREASING AT ALL and was stuck in the 800-ish-boardings-per-day doldrums. This service was not winning hearts and minds; any claims to the contrary are completely contradicted by the actual boarding numbers.

Important notes going forward:

1. In January, Capital Metro eliminated some very popular express bus runs to drive more people on to the Red Line. Early indications were that this had limited success – with people complaining about crowded conditions on the one new route (#985) that replaced the two popular ones (#984 and #986), but I bet some people switched to the Red Line to avoid said crowding. This #985 bus was crowded despite the fact that it’s slower than the previous direct-from-park-and-ride services it replaced. (In other words, the 2 old buses were much faster than the Red Line while this new replacement service is only a little faster).

2. Capital Metro also consolidated a couple of peak runs – meaning that you’re more likely to see a ‘full’ train, because fewer trains are being run during rush hour.

3. Ridership is being goosed artificially. Any gains we see now are most likely not the result of people trying the service and loving it; or being convinced to ride during rush hour because of new midday service; they’re most likely the result of people who liked a competing service being driven here because Capital Metro made the competing service less competitive.
I ripped up one of the copies I had and gave a short excoriation of the lack of meaningful public input, this as this KUT story indicates. Here’s the outline of the speech I was going to give (4 people had donated me time; I’m not sure I could have fleshed this out to 15 minutes if I tried).

Unlike some people who spoke with most or all of their time, prostate I thought it more important to indicate that we didn’t agree with the decision to limit testimony (at the only real public hearing this thing was ever going to get).

1. Background
a. Member of AURA (founding member of the new version; supporter of the old)
b. UTC 2000-2005
i. Mention PG, modern UTC opposition in JD, MDG
c. Writing on transportation since 2003
d. On corner in 2000 supporting LRT with Eric Anderson (LAB)
e. Opposed Red Line in 2004 due to high operating subsidy and low benefit to Austinites
i. (mention this has borne out – operating subsidy ‘down’ from 35 to 18 after cancelling buses; census from rider at Lakeline showing 80% Cedar Park)
2. PC Process
a. In it since beginning.
b. Assured LG on table. No obstacles. (Also assured of this years ago when Rapid Bus was pushed).
c. Process clearly designed with thumb on scales
i. Subcorridors instead of routes
ii. West Campus into Core
iii. I35 ridership into Highland but neither I35 nor 183 ridership into Lamar
iv. Various ‘errors’ all of which hurt GL
v. Bad flier – canvassed at my house with flier designed to fool old people into dropping opposition to plan. No real plans for rail on Guadalupe!
d. At end, people still didn’t know what was best for them!
i. Repeated, strong, unbending preference for Lamar ‘subcorridor’
e. So we brought up the FTA out of the blue
i. Disputed by the guy in charge of Rapid Bus!
ii. Either lying now or lying at beginning.
iii. Getting mixed messages – we’ll do LG right after election but LG can’t ever be done because of traffic but we’ll do it next anyways.
f. Nobody in Austin should trust the output of a process this corrupt. You’re being fed a line about transparency that doesn’t hold up. None of our local transit activists who aren’t connected to the machine believe this.
i. National commentators:
1. Christof Spieler: "It's amazing: Austin, the self-proclaimed progressive city, could have had the best rail system in Texas but has the dumbest."
2. Steven Smith: "Austin light rail is becoming more of a joke by the minute. Textbook example of politics getting in the way of good transit planning."
3. Jeff Wood, Reconnecting America: "I'm going to use this as a bad transit planning example forever"
4. Others at the time ranging from “What A Sham” to “What A Shame”.
5. Honestly have not seen a single national transit person approve this plan.
3. The output
a. High operating subsidies even WITH assumed out of reality growth at Highland Mall
b. No way to tell whether new residents around Highland will work along rail line
i. Mention Mueller – people work all over the city
ii. Birds in hand on a good bus line are worth more than ten birds in bush (working all over city)
c. Theory pushed by Chamber of Commerce that people will hop off I-35, go to park and ride, look for space, walk to station platform, wait for train, ride slow meandering train downtown instead of riding
i. Park and rides DO work but only at far end of quicker, straighter, lines.
ii. Or like in Houston where parking is very very expensive.
d. Urban rail should be urban.
i. Walk to stations from dense residential areas, not apartments in a sea of parking
ii. Entire Airport Blvd segment a waste – only one side can ever be developed; no good crossings to other side and low-density over there
iii. Hancock area – residential only, not as walkable as we need; no opportunity for redevelopment more urban.
e. Respond to density instead of create it
i. Christof Spieler – density wants to be near other density (fill in gaps rather than greenfield)
ii. Most of our supposed TODs underperform compared to background conditions
iii. Remember the TOD up in Leander that was going to help the Red Line?
iv. Crestview Station <<< The Triangle v. Not going to get high quality development in the planning straitjacket around Highland Mall (also remember birds in hand argument) f. Even with their bogus assumptions i. 18,000 boardings/day would be a bad light rail line. BAD. ii. Houston around 35,000/day. Phoenix above 40,000. g. Precludes expansions ANYWHERE else if line isn’t packed i. Operating subsidy argument ii. Horrible spine – slow, windy makes bad backbone. h. Precludes expansions on GL forever even if line is good i. We don’t trust you now after Project Connect Phase 1 ii. FTA reluctant to fund two early lines in ‘same’ area iii. Local politics makes funding 3rd line apparently in NCentralAustin a nonstarter iv. Are they promising Guadalupe or “Lamar subcorridor”? 1. Ridiculous longrange map proposes Guadalupe served after MLK but we suspect grade too high on MLK; doesn’t go south into core of downtown. Why not just stay on Guadalupe/Lavaca? 2. We don’t believe you anyways. 4. Conclusion a. Bad rail line can end system rather than start it b. Don’t mischaracterize our arguments. Highland is not just not our favorite line; it is a BAD line. Never get a chance to build system if you use up all your capital on a second high-operating-subsidy line.

  • Novacek

    (I’ll try to confine my objections to purely, objectively factual disagreements)

    >>i. 18,000 boardings/day would be a bad light rail line. BAD.>>ii. Houston around 35,000/day. Phoenix above 40,000.

    Houston is 3 times our size (both city and metro) and is 30% denser to boot. Phoenix is twice our size, AND their line is 20 miles (twice the size of the proposed Austin line). Both are _significantly_ denser than Austin by weighted density.

    http://www.austincontrarian.com/austincontrarian/2012/09/the-50-densest-american-metropolitan-areas-by-weighted-density.html

    It’s unsurprising the projected ridership is less on the Austin line.

    • http://m1ek.dahmus.org/ M1EK

      It’s ironic you would make the claim that this is an objective argument when the definition of ‘success’ is primarily subjective.

      It’s not appropriate to use density or even weighted density to excuse low ridership. If you get low ridership, you should not have built that rail line. It just so happens that in Austin we have better corridors than Highland where we would get higher ridership (which is almost a perfect relationship to lower operating subsidies, which is the real thing we should be measuring – although that’s harder to get precise agreement on).

      Austin’s 2000 light rail line projected 35-40,000 boardings/day by the 2020s. The Feds agreed with this projection. We’d see a higher projection today. Your implication that we could not achieve ‘success’ levels of Houston and Phoenix here is false. But I’m sure you knew that, and yet you made the implication anyways in the hopes some others would be fooled.

      • Novacek

        1. Why in the world do you think 18,000 would be _low_ ridership? That’s a very successful number. By total ridership, it would be top-25 in the US (which given that Austin isn’t even a top 30 metro, would be impressive). In ridership per-mile, it would be top-15. Again, amazing given Austin’s size and relative density.

        2. Austin’s light rail line had projected ridership of 37k by 2025. http://www.fta.dot.gov/12304_2915.html That’s after 25 years of growth and infill. How can you consider 18k a failure, when the 2000 plan would have had similar ridership in the same time period (say 8 years after opening)?.

        3. That 37k figure was also pre-McMansion and pre-duplex/stealth dorm limits. Given those new limits on density increases, I’d expect the (25 year) projected ridership on a G/L line would be lower today.

        • http://m1ek.dahmus.org/ M1EK

          1. 18,000 boardings is very low; there aren’t enough light rail lines in the USA to consider top 25 impressive. Metro area is not a meaningful metric here. Neither is size and relative density. A line is a success if it carries lots of people; it is a failure if it does not. Density tells you where you should put the line; its lack is not an excuse for poor performance – it is a reason you never should have put the line there in the first place.

          2. 2000 plan would have had far more than 18,000 boardings after 8 years IMO. They did NOT count on massive, huge, disproportionate and unbelievable growth and infill (certainly nothing along the lines of what Project Connect claims will happen at Highland, for instance). Moderate growth as typically seen along good light rail lines was what was projected.

          3. McMansion and duplex/stealth dorm have far less effect on # of people within walking distance of Guadalupe than do the Triangle, VMU, etc.

          • Novacek

            1. And 18k is a _lot_ of people. It’s a very successful starter line, and on the same order as the expectation for the 2000 G/L line (see below)

            2. How many do you think? What’s your definition of “moderate growth”? What was the expected growth rate? 3%, 4%? Anything that low (basically just Austin’s population growth) would have to be a “failure”, wouldn’t it? CapMetro is already seeing that sort of ridership growth system-wide. Certainly the expectation/hope is that fixed rail encourages additional growth and infill. Let’s look at this analytically:

            After a 2000 vote, the system would have finished construction in 2007, 2008? 8 Year ridership in ~2015, so ~10 years of growth to get to the 2025 ridership number.

            With 5% annual growth rate:
            X * 1.05^10 = 37K
            X = 22k

            With 6% annual growth rate:
            X = 20.5k

            Let’s try just 4% (even though that’s just barely faster than CapMetro is already growing)
            X = 25K

            And all of those are with a ~15 mile line. Per-mile, the 18k figure is actually _higher_. 18,000 is _not_ low ridership. It’s high-ridership.

            3. The triangle development _should_ have already been included in the 37k figure (as expected green-field infill). The VMU corridors, though better than nothing, are really not the sort of density you want/need for mass transit. They’re linear, while what you want for transit is _nodal_.
            Basically, draw a .25 mile radius around each station. That’s where you want super density. Draw a .5 mile radius around each station, that’s where you want high density. Do that for the for the stations of the 2000 plan, and you include a huge number of SFH that is now limited on the density it can increase due to mcmansion and other limitations.

          • http://m1ek.dahmus.org/ M1EK

            No, it’s not the same order of people. You’re engaging in gymnastics here to justify a really crappy rail line. 18K is boardings; it’s 9K people; that is NOT A LOT. By national standards that would be a failure.

            If you have arguments with the perception by Austin’s serious transit advocates that the Highland growth projections are out of line, I urge you to attempt to have that discussion with Julio at his blog. I’m accepting it as a given here that it is incredibly unreasonable, because Julio’s logic and methodology are sound, and he is more experienced in this realm than you and I put together.

            As for the Triangle, I am telling you it was not included (as housing) in the 2000 projections. I was here, you were not. Again you are treading on the line of calling me a liar. I urge you to reconsider this practice of yours. For background, the original Triangle proposal was quite different from what ended up being built – it was more of a large retail center originally (and this would have been what factored into the 2000 projections): http://www.austinchronicle.com/news/1997-02-14/527386/

          • Novacek

            Yes, 18K is boardings. The 37K from the G/L was _also_ boardings. They’re about the same number (if you compare similar timeframes). You can’t call one an abject failure and the other a raging success.

            http://www.fta.dot.gov/12304_2915.html

            >>If you have arguments with the perception by Austin’s serious transit advocates that the Highland growth projections are out of line,

            But that’s not what we’re talking about. We’re talking about the 18K projection vs. the projection of the 2000 G/L line. Given the assumption of 18K boardings, as you noted above.

            The validity of the 18K projection is a _separate_ issue. One which will take care of itself. If the Feds don’t like our projection, they won’t fund us.

          • http://m1ek.dahmus.org/ M1EK

            They are not the same number. Not even close. You have to go through more gymnastics than you see in the Olympics to justify this crap.

            The 2000 MOS projected 37,400 boardings/day by 2025, with minimal growth assumptions compared to the Project Connect plan.

            The 2000 plan projected less population growth than we have actually seen. I was here at the time. It underestimated growth. The Project Connect plan projects incredibly higher growth than we have EVER seen. It overestimates growth.

            Your intent here appears to be to mislead rather than inform. Only somebody whose goal was to obfuscate would conflate those two lines and declare that they were about the same.

          • Novacek

            Minimal growth? It projected population growth of 120% within .5 mile (from 48k to 109k).

            http://www.fta.dot.gov/12304_2915.html

            Annualized over 25 years, and that’s 3.4% /year. That’s about the same as the 3.6% you all are claiming as totally unrealistic for the highland corridor.

          • http://m1ek.dahmus.org/ M1EK

            The 2000 line included some completely undeveloped areas up north. You’re cherrypicking now. All that had to happen to fill in those areas was development of similar character to nearby areas; nothing out of the ordinary. Suburban single-family and MF-3 development would and did add a lot already.

            “Densities are highest around the eight stations in the CBD and UT area, while the northernmost two station areas are largely undeveloped. Currently, only minimal development exists along the northern portion of the MOS between McNeil Road and Howard Lane.”

            Note this doesn’t go as far as Leander or Lakeline; the MOS just went to McNeil. I’d be curious to see what the population is today; I suspect Scott Morris has this information by now. Even just to Crestview, he found 54K population today (2013).

            There is no similar largely undeveloped area near a projected station on Highland. Redeveloping the mall is a lot harder than greenfield development.

          • Novacek

            Again, do the math:

            Say those two northern stations were _completely_ undeveloped and unpopulated in 2000. Say they get developed with SF and MF-3 to a density of 4300 (the average density around the station areas in 2000, from the fta report). Bear in mind this is significantly _above_ the average austin density, as that 4300 average included CBD stations.

            That provides a population increase of 6751 of that 61500 increase (.5^2 x 3.14 x 2 x 4300).

            It’s not the greenfield development in the northern two stations that provided the population increase. Its the densification and infill in the lower stations.

            The 2000 plan predicted (and relied on) significant population growth (and ridership growth) over the 25 years. That’s completely right and proper, as we would expect it to do so. But because it relies on such growth, you can’t pretend that the 2025 ridership and the ridership at 8 years in (~2015) would have been the same.

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            The 2000 plan did not contemplate the Triangle’s residential; nor did it contemplate the VMU ordinance. All anticipated growth along that corridor was due to greenfield development and some modest infill on previous, smart growth, grounds.

            It boggles the mind that you consider this projection, which is actually well on the way to being EXCEEDED, with the aspirational scenarios around Highland.

          • Novacek

            Again, the VMU corridor concept is _bad_ for transit. It doesn’t put the density where it needs to go.

            >>It boggles the mind that you consider this projection, which is actually well on the way to being EXCEEDED, with the aspirational scenarios around Highland.

            That’s just it, _if_ 3.4% growth is being _exceeded_ along that corridor (most of it not greenfield, as shown), why is 3.6% so impossible elsewhere?

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            The VMU concept is not bad for transit. An unbroken string of 5 or 6 story buildings along Lamar and Guadalupe is great for transit. Granted, 5 or 6 story buildings with 10 story buildings at the stations would be even better, but the 2000 plan did not anticipate that happening either.

            The Lamar/Guadalupe corridor is growing as Christof Spieler says, because density works best between existing density; not in the middle of nowhere. The Highland proposal is more like the Red Line’s anticipated (still not there) TOD at Leander.

          • Novacek

            VMU is good for transit.

            VMU _corridors_ are bad for transit.

            Like I said, it’s linear vs. nodal. Draw a radius around each station. That’s where you want to have density. You don’t want (given the choice) to have density in the no-man’s-land _between_ stations.
            You want 30 VMU buildings clustered around each station. You don’t want (at most) 4 buildings next to the station, then 4 more further way, then 4 more far away, then 4 more _really_ far away.
            Transit can’t effectively serve linear density. You can’t stop in front of each building, or your transit is going to be as slow as SanFran, where you can walk faster.

            You draw that radius around each of the 2000 rail plan stations, and it encompasses huge swaths of SFH-zoned properties. Properties that now can’t even get as dense as they could in 2000.

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            You can make this claim all you want. I submit that in the central area, where there is projected to be a station every mile or less, your half-mile circle will encompass basically all of the VMU corridor. Some of the ‘linear density’, as you call it, will be as much as a half-mile away from the rail station, it’s true. But again, we are better off with a corridor of 5 or 6 story buildings than what the real-world alternate would be in the world where Steven Zettner gets more time than you or I do: 5 or 6 story buildings within one block of the station, but still ONLY on the road itself – nowhere into the neighborhood, and the same old single story strip malls in between the nodes.

          • Novacek

            >> But again, we are better off with a corridor of 5 or 6 story buildings
            If you even get that. More likely one or two 5/6 story buildings, then a few plots the NA “opted out”. Then some plots the NA association will fight tooth and nail against any business that increases traffic or plays music or is open late or serves alcohol.

            But those aren’t the only options, and that’s the beauty of the development at highland. Nodal density outside of the control of the Nimbys/NAs.

          • Novacek

            >> Redeveloping the mall is a lot harder than greenfield development.
            Come on, it’s not like Highland is a toxic brown field requiring remediation. It’s a bunch of parking lots. It’s probably actually easier to redevelop, since you don’t have to clear or grade the land. Especially once you factor in the fact that it’s 5 miles from downtown instead of 15 miles from downtown. There’s a lot more economic incentive to develop it, as compared to McNeil.

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            Highland may be geographically close to downtown, but it’s not well connected, and the urban rail proposal will not change this much. It’s an odd island separated from everything by ugly freeways or ugly stroads – neither of which is ever likely to change. It’s not ever going to be as desirable a place as the Triangle was to develop.

            Greenfield development is much easier than redeveloping a mall. Since I have a paid day job and am not being paid to do this, I don’t have time to cite dozens of supporting pieces here, but if you were an honest actor your response would begin with “Granted, but” rather than “Come on”.

          • Novacek

            >>Highland may be geographically close to downtown, but it’s not well connected, and the urban rail proposal will not change this much.

            So a direct rail connection to Campus and downtown would be “not well connected”? Do you really hear yourself? (that’s not even mentioning the connection to I35 and 290 and the 290 toll and the red line).

            Also, you do realize that the triangle is surrounded by stroads (Lamar Guadalupe and 45th are all stroads in those sections).

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            The ‘direct rail connection’ to Campus and downtown will be slow, meandering, and indirect. Look at the route.

            I35, 290 are barriers, not connections.

            And if you are claiming that 45th, Lamar, and Guadalupe are not somewhat qualitatively different from Airport Boulevard in this section of town, just go ahead and say so directly. I’ve earned a good laugh for putting up with you this long. The Triangle IS more isolated from its neighbors than I would like, to be fair; but that’s not the example I was using. The Triangle is still better connected _to_ downtown than Highland ever will be; Triangle just isn’t well connected to things _across_ Lamar and Guadalupe. Highland isn’t well connected to _anything_.

          • Novacek

            >>The ‘direct rail connection’ to Campus and downtown will be slow, meandering, and indirect. Look at the route.

            That “slow, meandering, indirect” route is like 10% longer than the direct straight-line distance from highland to downtown (which while ideal, is basically never achievable in real life. Even the G/L 2000 route followed existing roads/RoWs instead of going as the crow flies). That’s like the difference between a 15 minute ride and a 17 minute ride. But you think a 40 minute ride from 3 times as far away (McNeil) is a preferable location for development?

          • http://m1ek.dahmus.org/ M1EK

            Again, I wonder why you are never this hard on the people with paid staff to generate this kind of backing information.

            Compare to the Guadalupe/Lamar route from Highland to downtown (straight down Lamar and Guadalupe, one 30-degree-or-so turn, done).

            McNeil is a preferable location for development because it is a greenfield, which is easier to develop on.

            Highland is the definition of an isolated enclave. It’s trapped on 2 sides by freeway frontage roads, one more side by an old state highway with nothing across the street to get to because it’s got freight rail on it; and on the north by crappy strip development (which is at least eventually, possibly, redeemable). If you won’t admit this, you are nothing but a troll.

          • Novacek

            >>Compare to the Guadalupe/Lamar route from Highland to downtown (straight down Lamar and Guadalupe, one 30-degree-or-so turn, done).
            The full G/L route had way more turns than that
            http://www.fta.dot.gov/12304_2915.html

            I’m not saying that G/L had more or less turns than the proposed highland route. Just that simply having turns doesn’t make the route bad (just like they didn’t make G/L bad), as long as it doesn’t add huge amounts of delay (10% is not huge). No one has a perfectly straight route as the crow flies.

            >>one more side by an old state highway with nothing across the street to get to because it’s got freight rail on it

            This is completely not true, development can cross the rail. There’s already a crossing at Denson, Plus, CapMetro owns the rail, and can add additional pedestrian crossings.
            Edit: Not that it’ll be an issue for a long time. Airport Blvd is a half a mile from the proposed transit stop. The rest of highland will be full up before we start worrying about the rail line being a barrier to development.

          • http://m1ek.dahmus.org/ M1EK

            Got it. You’re a troll. Bye.

          • Patrick Goetz

            > This is completely not true, development can cross the rail. There’s already a crossing at Denson

            It sounds like you’re making this argument theoretically, with no practical experience dealing with these issues. I live in the neighborhood planning area that includes Denson/51st. When the Red Line went in, Cap Metro sealed off a number of walking paths that have been used for decades to travel east/west to and from Airport Blvd.. We’ve asked Cap Metro REPEATEDLY to re-open at least one these paths, perhaps using switchback walkways, but made it clear we were open to any solution that would provide pedestrian access. Nothing whatsoever has been done about this other than JMVC making empty promises that they would look at the problem. In nearby Crestview, the Red Line bisected a public park; you can no longer get from one side of the park to th other. If you don’t think the Red Line is a major barrier to especially pedestrians then you simply have no idea what you’re talking about and need to get out in the real world more.

          • Novacek

            My practical experience is living immediately next to a signalized pedestrian-only crossing of the red-line, that remains open to this day, which I use daily.

            >>When the Red Line went in, Cap Metro sealed off a number of walking paths that have been used for decades to travel east/west to and from Airport Blvd.
            Were these official signalized crossings, or dirt tracks where people had been illegally trespassing for years?

            >> In nearby Crestview, the Red Line bisected a public park; you can no longer get from one side of the park to the other.

            Umm, there are _no_ public parks in crestview. That’s why the NA is trying to get one built. What land are you talking about?

          • http://m1ek.dahmus.org/ M1EK

            Bye again a second time. When I see you being this hard on Capital Metro, who is paid by our taxes to do these plans, I’ll reconsider.

          • George Purcell

            “The full G/L route had way more turns than that
            http://www.fta.dot.gov/12304_2…”

            WTF?

            It has as many turns going all the way out past 183 as this BS line has to get to Hancock!

            Are you unable to read a map?

          • Patrick Goetz

            > Come on, it’s not like Highland is a toxic brown field requiring remediation. It’s a bunch of parking lots.

            It’s a bunch of parking lots with an impossibly complicated ownership contract involving multiple parties. I was part of the Highland Mall redevelopment discussions and can assure you that the situation is far from straightforward.

        • Steven Yarak

          Of course, it’s not really 18k. It’s 15.5k. And even getting that requires a huge amount of handwaving under the guise of “analysis.” http://keepaustinwonky.wordpress.com/2014/06/18/a-day-late-and-a-few-thousand-riders-short/