Yet More Prius iPhone Blogging!
How does one rack up a high score playing Prius: The Videogame?
The instinctive strategy is to try to minimize the percentage of the time the gasoline engine is on.
But that cannot be right.
« Proof of Concept iPhone Blogging | Main | Larry Summers on Stupidest Man Alive Emeritus Donald Luskin »
How does one rack up a high score playing Prius: The Videogame?
The instinctive strategy is to try to minimize the percentage of the time the gasoline engine is on.
But that cannot be right.
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00e551f08003883400e55238f2628834
Listed below are links to weblogs that reference Yet More Prius iPhone Blogging!:
You can follow this conversation by subscribing to the comment feed for this post.
The comments to this entry are closed.
"I now know it is a rising, not a setting, sun" --Benjamin Franklin, 1787
J. Bradford DeLong, Professor of Economics at U.C Berkeley, a Research Associate of the NBER, a Visiting Scholar at the Federal Reserve Bank of San Francisco, and Chair of Berkeley's Political Economy major.
Among his best works are: "Is Increased Price Flexibility Stabilizing?" "Productivity Growth, Convergence, and Welfare," "Noise Trader Risk in Financial Markets," "Equipment Investment and Economic Growth," "Princes and Merchants: European City Growth Before the Industrial Revolution," "Why Does the Stock Market Fluctuate?" "Keynesianism, Pennsylvania-Avenue Style," "America's Peacetime Inflation: The 1970s," "American Fiscal Policy in the Shadow of the Great Depression," "Review of Robert Skidelsky (2000), John Maynard Keynes, volume 3, Fighting for Britain," "Between Meltdown and Moral Hazard: Clinton Administration International Monetary and Financial Policy," "Productivity Growth in the 2000s," "Asset Returns and Economic Growth."
The Eighteen-Year-Old is going to college next year, which means that I need to think about making more money. (The idea that one might write checks to rather than receive checks from universities is now strange to me.) So I have signed up with the Leigh Speakers' Bureau which also handles, among many others: Chris Anderson; Suzanne Berger; Michael Boskin; Kenneth Courtis; Clive Crook; Bill Emmott; Robert H. Frank; William Goetzmann; Douglas J. Holtz-Eakin; Paul Krugman; Bill McKibben; Paul Romer; Jeffrey Sachs; Robert Shiller;James Surowiecki; Martin Wolf; Adrian Wooldridge.
Work is force times distance; minimizing the time you are pushing down the gas pedal is a mistake. Given that total distance is usually a constraint, get there by pushing down the gas pedal the minimal amount. (The Prius video game is giving rise to a generation of drivers who coast for as far as they can.)
Posted by: Sean Carroll | August 31, 2007 at 03:23 PM
Indeed, there is extensive web activity on efficient driving techniques, for the Prius in particular - exceeding 100mpg is far from impossible.
A quick search turned up this:
http://en.wikibooks.org/wiki/Toyota_Prius_driving_tricks
A popular aftermarket mod is the "EV" button, a switch which reduces the usage of the gas engine, useful for short hauls. This button is standard for non-US models.
Posted by: Lab Rat | August 31, 2007 at 04:00 PM
The conventional wisdom is pulse and glide. That is strong acceleration to the speed limit then coast until you are going slow enough to be a hazard then repeat. It must be true because I read it on the internets.
google pulse and glide prius and the first entry is http://www.toyota.com/html/hybridsynergyview/2005/fall/marathon.html
http://www.google.com/search?q=pulse+and+glide+prius&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a
Posted by: marc sobel | August 31, 2007 at 04:00 PM
Prius: The Video Game is pretty straightforward.
First, taking energy out or putting energy in to the battery drops your milage.
Second, you cannot be a "stomp and go" driver. (Even though the Prius has great low end torque, don't do it. Full Disclosure: I love beating high HP diesel pickups off the line with my Prius.)
Third, be a "pulse and glide" driver. A major hypermiler skill is "feathering" the engine - running neither the gas engine nor electric motors. Pulse the speed up, say 5 MPH over the speed limit, and feather down to 5 MPH below the speed limit.
Fourth, exploit gravity. Coast downhill, use gas uphill.
Fifth, look ahead and coast to stopped lights. Regenerative braking, while useful, violates rule #1 above and ignores physics. Don't be unsafe but don't be in a hurry to get to that red light.
The Prius' milage is remarkably sensitive to how you drive it. It will reward skillful application of the accelerator.
Andrew
Posted by: Andrew Donoho | August 31, 2007 at 04:02 PM
please tell me you are not watching the videogame while iphone blogging while driving.
Posted by: marc sobel | August 31, 2007 at 04:04 PM
A strange game. The only winning move is not to play. How about a nice bicycle ride?
Posted by: joshua | August 31, 2007 at 04:46 PM
This is actually an excellent illustration of why a better version of the CAFE standards--and one far, far more preferable to automakers--is to mandate immediate feedback of the instantaneous gas mileage of cars. People drive much differently--and much more efficiently--when they understand just how much energy is involved in, say, an acceleration to 30 only to brake at a red light at the next block. Prius drivers get that feedback and alter their behavior based on their preferences--and not just their environmental preferences. They choose whether to drive in a manner that gets them 60 mpg or 40 or 25 based on their needs at the moment. Indeed, I wonder if there's not a good academic paper in it on marginal work (eg, an extra minute driving to the store to save 25 cents' worth of gas) and home production.
Posted by: mario | August 31, 2007 at 10:57 PM
A small set of useful heuristics:
- Flat terrain mileage sucks when kat and engine are cold, usually during the first 5 minutes. During this time, if you have the choice between two routes, one going uphill and then downhill, the other going downhill and then uphill, choose the former. You'll warm up the car going uphill, and then get almost double the mileage in flat sections, and a silent coast going downhill.
- Go 30 in 25 zones, go 35 in 30 zones, go 35 in 35 zones. Efficiency seems optimized between 35 and 55 mph.
- Prefer routes with longer but gradual declines over routes which mix shorter, steeper declines with flat sections when going downhill, but choose the opposite when going uphill.
- Unless in stop-and-go traffic, use the cruise control. This has a dramatic effect, as the board computer will be able to help find the optimal mix. However, vary speed manually with the cc lever only downwards. Use the foot to accelerate.
- Coast, but also break lightly early when coming up on a red light, particularly going downhill. Regeneration is likely less effective with sharp breaking. Slowing down early also increases the chance that you won't have to come to a complete stop, as the light might change before you get there. You'll also reap an increased share of middle fingers.
- Don't go in the right lane on freeways. Incoming mergers might make you lose ticks on the yellow bar as you need to re-accelerate after breaking, raising blood cortisole levels.
- Keep distance from other cars when stopped at a light. It will allow you to start moving earlier and accelerate slowly using only the battery when the light changes to green.
Posted by: Thor's Hammer | August 31, 2007 at 11:36 PM
Genuine question: is there any control of the battery recharging? Extra green points could then be won by reaching home with a flat battery and recharging it at night from renewable electricity.
Posted by: James Wimberley | September 01, 2007 at 02:28 AM
Meh, doesn't sound as if could get me through the burbclaves fast enough. And I would be worried about viruses attacking through the bluetooth.
Posted by: Hiro Protagonist | September 01, 2007 at 07:41 AM
andrew, mario, and thor's all have good things to say. I've come to similar conclusions myself, although I don't have a real way to test for optimality.
I've always been suspicious of cruise control. On non hybrids, they seem likely to do dumb things, like brake on small dwonhills, and downshift on small uphills. I suspect the advice to use them is because they are less wasteful than the average (ie. terrible) driver. Is it possible that the cruise control might be programmed for mini-pulses?
It would be nice to have a bias toward/away from battery drawdown switch. Approaching a longish downhill you'd like your battery to have been drawn down. During situations where EV-glide is unlikely you might want to be biased towards charging. Note battery life is maximized by keeping the charge level near the center of its range, that is probably why they have been reluctant to endorse plug-ins. The proposed Chevy-Volt is to use A123 Lithium Ion batteries. LI batteries have twice the energy density. The A123 has a different electrode design which is purported to solve the battery-fire hazard, which has been such a problem in laptops & cellphones.
So despite the fact that Toyota is well ahead of US auto manufacturers good research/development is being done here.
I tend to accellerate fairly fast to cruising speed. Pulsed driving is only really sensible when you have a slow driver behind you.
Posted by: bigTom | September 01, 2007 at 08:48 AM
"The proposed Chevy-Volt is to use A123 Lithium Ion batteries."
Assuming the Volt is delivered in the serial-hybrid configuration currently being hyped, much of this nonsense about driving habits goes away. The electric drive train has much the same efficiency across its entire range of speed/acceleration; the gasoline engine (in configurations that include one) will run at a constant speed that maximizes its efficiency as a generator; slow gradual braking or hard abrupt braking should recover about the same amount of energy.
Increased federal funding for battery research to make serial-hybrids affordable, please. Aluminum-sulfur looks very nice, but probably requires nanotech engineering to deal with the current problems.
Posted by: Michael Cain | September 01, 2007 at 09:25 AM
Michael, thanks for the heads on on the Volt. I only know what I've seen on Wikipedia -or read in insidegreentech.com (without a paid subscription).
My guess is if they don't hit a major snag it will be out in 2010. The Prius was originally out in 2000 (or was it maybe a year earlier), but conservative production ramp-up meant it has only very recently been available in sufficient numbers to make a difference. I suspect that the Volt (if it comes to be), will have a long slow ramp as well.
Posted by: bigTom | September 01, 2007 at 12:08 PM
mario:
"when they understand just how much energy is involved in, say, an acceleration to 30 only to brake at a red light at the next block."
My one experience, as I mentioned on the previous thread, is that an instantaneous gas mileage does not teach that lesson, although it may teach other relevant lessons.
I doubt that there are many situations today in which someone would trade off time for gas. You'd have to value time very little to want to drive more slowly on a highway. But in urban traffic, there is much room for improving efficiency without time cost, as in the red light you mention.
Posted by: Douglas Knight | September 01, 2007 at 01:14 PM
Doug Knight is likely right. I consider instantaneous gas consumption measurement as essentially useless -i.e. it is too noisy. The trip average on the Prius is useful, but it requires a bit of intelectual effort to correlate average milage with specific driving habits.
Few would waste even a second. Even in the red light situation, most people are using something very close to what us computer programmers would call a "greedy algorithm", go as fast as possible for the next very short period of time. Clearly in the red-light situation, there is a very real chance that one or more impatient drivers will zoom past you to get ahead in the que, that is enough negative reinforcement to prevent most from coasting.
Posted by: bigTom | September 01, 2007 at 03:46 PM
Douglas Knight:
In classical economic theory, you are indeed correct. Or you should be--for if you make 30 or 60 or 100 dollars an hour, spending an extra minute for 15 cents' worth of gas doesn't make sense.
I have two responses to this. First, as Deirdre McCloskey and others have pointed out, all economic actions depend at heart on perfect information. Indeed, *preferences* depend on perfect information. The above action--i.e., spending a minute for 15 cents' worth of gas--is only inefficient if we consider driving to be solely a cost; that is, that it's only something we do because we have to, and that the act of driving itself gives no utility. This is obviously wrong to me, as one only has to think of a 16-year-old new driver or anyone infatuated with car culture to see that driving itself, even going nowhere, has value.
What I find in practice with a Prius is that the relationship between speed, mileage and utility changes based on the information that the car presents its driver. Say I'm enjoying driving the section of 580 between El Cerrito and Pinole, and that my enjoyment will go down once I go past Pinole. And let's say I could have taken some alternate route that would take 10 seconds less time door to door. Intuitively, it's clear that I'm going to take 580 every time--because it's more fun to drive.
Now say my alternate route takes 10 minutes less, and that that's my marginal point. We obviously can establish a drive-minute valuation of 580; it's not clear, however, that that value is the same as that of 10 minutes of my time at work.
Once I am driving a Prius (or any other car with a similar instaneous mpg indicator) I can better adapt my behavior to my preferences for that day. My drive-minute preference is not constant, but instead changes based on the day. If I'm late to a job interview, I'm going to go fast no matter what. But at the margins--meeting my partner's parents, say--the mileage is going to help me decide what's best for me.
*
And in my experience, people often decide to go much more slowly than they would otherwise. Driving isn't purely about getting from points A to B, though that's the majority of its value.
Another population on whom we can see this same effect is among serious bicycle commuters, the ones with panniers and racks and perhaps the frame extender known as the Extracycle. Cycling is a feedback-heavy experience, and it's very clear which speeds are most efficient--they're the ones that are most sustainable. While I could go 28 mph on level ground, it uses a lot of energy and makes me tired. I modulate my speed based on how fast I need to be at my destination, whether I need to arrive un-sweaty, my daily preferences, and others. Prius drivers similarly modulate the work the car is doing based on their costs (time and gas, as opposed to time and calories on a bike), which is information they know better than drivers of any other vehicle I am aware of.
*
BigTom:
I use instantaneous gas mileage in a loose sense, to mean any metric similar to that in the Prius. While it's indeed messy, there are a number of good compromises, mostly based on the average wattage of the engine for the past [time interval]. This is actually such an important piece of information that even noisy data is very, very useful.
Posted by: mario | September 03, 2007 at 12:33 AM
It might be worth noting here that the thermodynamic efficiency of the typical internal combustion engine (both Diesel and Otto cycle) is optimal at steady-state loads. Stomping on the gas, then coasting, lather, rinse and repeat is unlikely to achieve optimal results.
I haven't played the videogame, but I would suspect that a winning strategy is (a) to make the most of regenerative braking, (b) to accelerate slowly, and (c) to use the cruise control as much as possible during long, flat, highway segments.
Posted by: s9 | September 03, 2007 at 01:08 PM