Comments on: Driving, How To

Driving, How To

wittgenstein — Wed, 16 Jan 2008 11:00:01 -0800

Some unusual thoughts on how to drive. Wait and See. Trucks. Bubbles and Barriers.

There is a lot more, though you may have better luck at figuring out exactly how to navigate the site than I did.

By: mrnutty

mrnutty — Wed, 16 Jan 2008 11:11:16 -0800

My favorite article on "traffic theory" (a name I just made up, but surely it exists?) is this one about traffic waves.

By: vorfeed

vorfeed — Wed, 16 Jan 2008 11:23:50 -0800

Wow. I don't think I've ever seen "drive defensively" put less succinctly. We ought to invite this guy over for a nice plate of beans.

By: Blazecock Pileon

Blazecock Pileon — Wed, 16 Jan 2008 11:30:25 -0800

Interesting. There's a Java-based 4D maze application there, too, which made me look around and find this gem about a 9D maze.

By: uncleozzy

uncleozzy — Wed, 16 Jan 2008 11:31:28 -0800

Unusual, in that they are thoughts at all? I commute in sometimes-heavy traffic, and the complete lack of understanding about what will get us all home fastest is mind-blowing. (Although I will admit to occasionally fucking with people who I can see are assholes, rather than just clueless; gotta keep myself entertained somehow.)

By: DU

DU — Wed, 16 Jan 2008 11:31:38 -0800

I (over)think this stuff all the time when I'm driving. The problem is that all of the strategies that work towards a global optimum (rather than just getting me myself home faster) require me to be able to program other people's brains. Which is what the webpage is intended to do, but I don't think he's getting much market penetration.

By: mr_crash_davis

mr_crash_davis — Wed, 16 Jan 2008 11:32:17 -0800

If you're the driver of car A in the linked "wait and see" scenario, you shouldn't be in the passing lane in the first place unless you're pulling out to pass, in which case "slowly catching up to car B" should be "speeding up to pass car B in order to clear the passing lane so car C can go by".

By: Kirth Gerson

Kirth Gerson — Wed, 16 Jan 2008 11:39:38 -0800

About daytime running lights:

The problem is, if and when a substantial number of cars have DRLs, a car with DRLs will no longer be relatively more visible. It will still be absolutely more visible, but for most people, driving in traffic, that's not so important. So, the benefit of having DRLs will be lost, or at least reduced, and the collective effort spent on them will have been wasted.

This person obviously has not spent much time driving a motorcycle. Being absolutely more visible most certainly is important, no matter what you're driving. DRLs at the very least reduce the impact of people driving their fog-colored cars with lights off in foggy conditions.

By: uncleozzy

uncleozzy — Wed, 16 Jan 2008 11:40:36 -0800

(On the subject of fucking with people, I unwittingly fucked with a state trooper a few weeks ago. All I knew was there was a car in my peripheral vision cutting in and out of lanes, so I boxed him in. Note to staties: if I'm in heavy, 70mph traffic, and you're so far up my ass that I can see neither your headlights nor your red-and-blues in the rearview, I'm probably not going to move out of the way until you turn on the siren. And if you're going to weave like a maniac, turn the lights on.)

By: chinston

chinston — Wed, 16 Jan 2008 11:44:19 -0800

Neat site. I hope this guy has read some Montaigne, because he would find a kindred spirit.

By: Ambrosia Voyeur

Ambrosia Voyeur — Wed, 16 Jan 2008 11:46:34 -0800

The combination of common sense and applicability to dense traffic situations conjures an ideal reader who is a disaster waiting to happen.

By: geoff.

geoff. — Wed, 16 Jan 2008 11:53:35 -0800

The mathematics seem rather doable. Why aren't cars automated yet? At least on major high ways? You program your exit and let the cars do the rest. You wouldn't need to know where every car is, just the ones in your local area and create probability structures based on that. It would be the most efficient for you (local randomness), but most efficient for the structure as a whole (global determinism). I mean cars driving themselves is a rather simple matter, correct? Add some redundant sonar, some basic communication between cars and a virtual line so the car can orient itself ...

By: Cool Papa Bell

Cool Papa Bell — Wed, 16 Jan 2008 11:54:03 -0800

From one of his essays... If I had to choose any single rule to replace the existing speed limits, it would probably be a rule based on relative rather than absolute speed. Which is why most states have a basic speed law that says no matter what is posted, you can't go faster than what is safe and appropriate for prevailing conditions. Just because the freeway speed limit is 70 mph, you can't go 70 mph in a snowstorm at night while towing a boat. But apparently that's not obvious to Captain Obvious the essay writer.

By: CheeseburgerBrown

CheeseburgerBrown — Wed, 16 Jan 2008 12:06:49 -0800

Idle-traffic-thoughts-self-whoring-filter: Aggregate Traffic Animals.

By: CKmtl

CKmtl — Wed, 16 Jan 2008 12:10:47 -0800

On Trucks - In stop-and-go traffic, trucks need to maintain larger gaps than cars. You can use these gaps to move around; on the other hand, you should also be nice. I also like to think of trucks as rocks in a stream. "Feel free to cut off trucks, but be nice about it!" I'm pretty sure the truckers I know would be greatly annoyed.

By: MrMoonPie

MrMoonPie — Wed, 16 Jan 2008 12:17:09 -0800

Reminds me of an earlier AskMe about merging.

By: Skorgu

Skorgu — Wed, 16 Jan 2008 12:24:30 -0800

All of driving can be summed up in two excellent sentences: If you're having to use your reflexes to avoid a collision, you've already screwed up. Don't get surprised; don't surprise other people. I usually go into a little more detail when I'm teaching a new driver. Your brain is a man riding a dog riding a lizard. They're good at entirely different things, and you have to know where to apply each layer. Your lizard brain is completely reflexive, it deals with things that are here (the guy who swerved into your lane unexpectedly and braked hard). ~250ms reaction time * 65mph = ~25 feet. Inside that semicircle you're running on pure instinct and reflex, something that only gets built up by practice and experience. So don't try to actively think about it, you won't react in time anyway. And that's where the dog brain comes in. You can read the minds of your fellow drivers. Look at how that guy is drifting over, how this guy brakes randomly, see how the other guy's wheels are angling a little bit towards your lane. Using those cues you can learn to predict dangerous situations before they get inside that circle of instinct. The only place your real conscious train of thought comes in is looking down the road, seeing situations long in the distance and planning for them. Is there stopped traffic waaay over that hill? Are people 100' in front of you braking? You can also run through situations during idle moments, what would I do if the guy in front of me's tire blew? If my tire blew? If there's a stopped truck around that blind corner?

By: cowbellemoo

cowbellemoo — Wed, 16 Jan 2008 12:45:19 -0800

Cool Papa Bell, the writer is more concerned with keeping different lanes travelling at different speeds. With absolute speed limits, most people would end up travelling the same speed. If limits were relative, say...the left lane can't go more than (but is allowed to go as much as) 30 miles faster than the right lane...barriers wouldn't form and (in theory) traffic would ease.

By: vorfeed

vorfeed — Wed, 16 Jan 2008 13:17:09 -0800

The mathematics seem rather doable. Why aren't cars automated yet? They've done a lot of work on this, and it's bound to happen eventually, but it's not an easy problem. Real-time detection of unpredictable moving obstacles & real-world signs, with close to zero tolerance for error, is not straightforward. In fact, it's so complicated that DARPA is willing to give millions to anyone who gets close. In their 2007 automated-vehicle contest, only 6 out of 11 vehicles could complete the course, and even the best of the best could only manage to move through the simplified urban environment at an average of about 14 miles an hour. So, unless you want to spend 4 to 6 hours to go about 60 miles, and maybe freeze up or hit something along the way, automated cars are not ready for prime-time yet.

By: mmrtnt

mmrtnt — Wed, 16 Jan 2008 13:59:25 -0800

This, is exactly how I drive. With the additions: On surface streets, I try to time the lights, keeping in mind the lag that the cars ahead of me stopped at the red will induce and On highways, trying to leave enough distance between me and the cars ahead so that we can all keep moving. And yes, on highways, I drive in the slow lane unless passing.

By: Cool Papa Bell

Cool Papa Bell — Wed, 16 Jan 2008 15:01:52 -0800

If limits were relative, say...the left lane can't go more than (but is allowed to go as much as) 30 miles faster than the right lane...barriers wouldn't form and (in theory) traffic would ease. Ahem. That's what we have right now. Slow traffic move right, with "slow" and "fast" being relative to the flow of traffic, the prevailing conditions and the absolute maximum speed allowable. The engineer-focused essay-writer is blissfully unaware of the real world, thinking that if he can just reach into the code and change a few ones to zeroes, everything will move faster. And that's so, so not the case, as any competent traffic engineer will tell you.

By: Kirth Gerson

Kirth Gerson — Wed, 16 Jan 2008 15:33:14 -0800

Ahem. That's what we have right now. Slow traffic move right, ... I don't know where you drive, Cool, but in MA, it's often just the opposite. So many people rush over to the left lane as soon as they can that it gets clogged and slow, while the right-hand lane flows smoothly at a much faster rate. It would be funny if it weren't such a display of utter stupidity.

By: Cool Papa Bell

Cool Papa Bell — Wed, 16 Jan 2008 15:55:53 -0800

Cool, but in MA, it's often just the opposite. Don't confuse the rules with what people actually do. This guy is arguing to change the rules. I'm saying the current rules already specify exactly what he wants and people still don't behave the way he wants them to.

By: Kirth Gerson

Kirth Gerson — Wed, 16 Jan 2008 16:07:59 -0800

Well, again, there is no such rule here. There used to be sign that said "KEEP RIGHT EXCEPT TO PASS," but they took them all down when passing on the right was legalized.

By: TrialByMedia

TrialByMedia — Wed, 16 Jan 2008 16:43:43 -0800

I wonder if this guy drives a 5-speed. I drove automatic for years without paying much attention to traffic other than what the car immediately in front of me was doing. I found myself thinking about driving a lot more after buying my first stick shift. My goal in driving is to get where I'm going while staying in gear as much as possible. This involves lots of observation, anticipation, quick thinking and a variety of techniques I had to teach myself. I think it's made me a much better driver.

By: Mitheral

Mitheral — Wed, 16 Jan 2008 20:18:30 -0800

Kirth Gerson quotes "The problem is, if and when a substantial number of cars have DRLs, a car with DRLs will no longer be relatively more visible. It will still be absolutely more visible, but for most people, driving in traffic, that's not so important. So, the benefit of having DRLs will be lost, or at least reduced, and the collective effort spent on them will have been wasted." This is a fundamental misunderstanding of the purpose of DRLs. They are to help during time of high contrast (driving into the sun, mirage/heat waves, glare off water, etc.). This is why they are so much more effective in northern latitudes like Sweden and Canada where periods of twilight at dusk and dawn are long.

By: Cods

Cods — Wed, 16 Jan 2008 21:00:40 -0800

As a motorcycle rider who also drives cars and heavy vehicles on occasion, I'm simply chuffed that this many people clicked on a link about driving, and then decided to comment. I'd contend that the vast majority of car drivers aren't thinking about driving whilst they're driving, let alone thinking about the most efficient manner of driving. Present readership excluded, of course... Further, I'd say that most motorbike riders' heads are filled with little else other than the task at hand and what's going on around them (yes, yes, huge generalisations, I know). Why? Motorbike riders who place some value on the continued operation of their squishy & breakable bits tend to be quite aware that most other road users are either a) out to kill them through inattention or b) see a). On a bike, good Situational Awareness (SA) = a better chance of being able to avoid having some car squish ones squishable bits or break ones breakable bits. Sure, the same is true for car drivers, but as would be obvious, car drivers tend to have have a tad more than a thin layer of leather and plastic armour between them and J. Random Commuter's somewhat cluelessly driven 2.5 ton 4WD. This extra buffer of automotive steel, glass, soundproofing and cupholders tends to blunt the realisation that one is punting a rather large machine along at quite respectable speeds, and that it doesn't take that much for it to all go pear-shaped. Of course, all these MeFi readers who are driving along contemplating the most efficient manner of driving rather than keeping their SA bubble alive probably aren't helping my riding brethren stay alive... ;) And I got all the way to the end of my slightly off-topic comment before mentioning OODA loops... but of course the OODA model assumes the other bloke is O, O, D and A'ing.

By: dhartung

dhartung — Wed, 16 Jan 2008 23:41:12 -0800

The other day, I was driving my family home on a four-lane divided highway at night [about here]. Naturally, even with no traffic around in either direction, I was in the right-hand lane. Suddenly, a car whooshed past in the opposite direction. My mind registered this as Oh, I didn't know I was back on the two-lane highway already. And about two seconds later, Hey, I'm NOT on the two-lane! WTF? Confirmation came as more cars appeared going that direction, but across the median. He'd probably just turned left, was unfamiliar with the area, and in the dark didn't get all the way over to the eastbound lanes. I didn't know what to do -- my cell battery was dead, so I couldn't call 911. I just said a little prayer for those behind me and kept on going. And the rest of the way home, even the part on the Interstate, I was a little wary of the left-hand lane.

By: Skorgu

Skorgu — Thu, 17 Jan 2008 07:23:26 -0800

Cods there's a truly great exploration of OODA at EjectEjectEject: Part 1, 2. Try to ignore the politics it trends into, but the discussion of dogfighting and being inside the decision loop is excellent.

By: Cods

Cods — Wed, 13 Feb 2008 17:12:14 -0800

Thanks Skorgu. Yeah, agree, that's pretty good (ignoring all the political bollocks, of course). Of note is that OODA tends to be explained using adversarial examples (fighter pilot A vs. fighter pilot B; Company X vs. Company Y), which is of no surprise given its source. However I believe that it works well in a non-adversarial context too - say, on the highway. If your goals are a mixture of defensive (avoiding getting squished by other road users), and self-interest (getting home as efficiently as possible), having a swift Observe / Orient / Decide / Act loop of your own going is a pretty good start. Of course, this is all a bit advanced for most drivers - paying some small amount of attention to what's going on around them might be a good start...

By: Skorgu

Skorgu — Wed, 13 Feb 2008 18:12:01 -0800

"I just assume I'm not invisible. I assume I'm wearing fluorescent clothes, and there's a million-dollar bounty going to the first driver who manages to hit me. And I ride on that assumption." —Neal Stephenson, Zodiac