The MPG Illusion (why we should maybe advocate for kwh/mi)

[lostpacket]

Slow Rivian news day today so I thought I might post this...

I was just learning about the "MPG illusion". Many of y'all may already know this but I think it would be good thing to advocate for as we adopt EVs, there is a chance to make this right.

The TLDR is that our brains have trouble calculating improvements in efficiency because of how we represent it. Our brains have a much easier time reasoning when the numerator and denominator are flipped (like they do in most of the rest of the world. (kwh/100km))

http://www.mpgillusion.com/p/what-is-mpg-illusion.html

Sponsored

 

[crashmtb]

I use L/100km already.
kWh/100km isn’t a stretch.

dunno what all this “miles” is

 

[OrthoBlock]

I find this a funny topic: I have historically completely agreed with how silly the US way (MPG) is for looking at fuel efficiency compared to the rest of the world (L/100km).

In terms of overall estimate of costs/consumption for a given time period/trip, the L/100 is much more insightful and useful.

However, in the EV world, where the point of concern for a trip is much less about the cost of the trip (at least for many), but rather on the range provided by the energy available then suddenly the range per unit of energy (e.g., Miles per KWh) is suddenly much more helpful in thinking about just how far can I make it.

...so perhaps the US way of thinking has just been ahead of the curve all along

(and yes, I realize that for the figuring the full-year costs of driving the EV, the energy per distance [e.g., KWh / Mile], is a more helpful metric for comparing vehicles)

 

[lostpacket]

I find this a funny topic: I have historically completely agreed with how silly the US way (MPG) is for looking at fuel efficiency compared to the rest of the world (L/100km).

In terms of overall estimate of costs/consumption for a given time period/trip, the L/100 is much more insightful and useful.

However, in the EV world, where the point of concern for a trip is much less about the cost of the trip (at least for many), but rather on the range provided by the energy available then suddenly the range per unit of energy (e.g., Miles per KWh) is suddenly much more helpful in thinking about just how far can I make it.

...so perhaps the US way of thinking has just been ahead of the curve all along

(and yes, I realize that for the figuring the full-year costs of driving the EV, the energy per distance [e.g., KWh / Mile], is a more helpful metric for comparing vehicles)

Good point but do I wonder would you be doing that "how far you can make it" math in your head? Or just using the range estimate displayed by the vehicle?

It feels like either kwh/mi or mi/kwh can be something people adapt to and get an intuitive sense of "how efficient am I driving lately" as it would just be deviation from a norm

 

[OrthoBlock]

Good point but do I wonder would you be doing that "how far you can make it" math in your head? Or just using the range estimate displayed by the vehicle?

If feels like either kwh/mi or mi/kwh can be something people adapt to and get an intuitive sense of "how efficient am I driving lately" as it would just be deviation from a norm

For me the application is more along the lines of: "...if I roll-up on that 6kw charger, how many miles should I expect my vehicle to add while I take my 1.5 hr hike... [hint: sadly not that many!]"

 

[zefram47]

Diehard mi/kWh fan here. And yes, I do the mental math constantly while driving my 110 mi rated range car. Until I get down below 25% I don't trust the car's estimated range. I live near/in mountainous areas and my consumption changes dramatically depending on where I drive and at many points along a given trip. I can look at current %SoC and know that for a given type of road I'll get around 4 mi/kWh or going up a canyon I'll get 2-2.5 mi/kWh and adjust accordingly. I spose with a big enough battery I wouldn't care as much, but that's not been my EV journey to date with small battery vehicles.

 

[ajdelange]

In terms of overall estimate of costs/consumption for a given time period/trip, the L/100 is much more insightful and useful.

Of course it is. Liters (or in the BEV case, KwH) is the dependent variable. It depends on how far you drive, how long the car has sat (phantom drain etc.). Most analyses involve wanting to combine use by rolling resistance, use for altitude change, drag... or determine average use per trip or per season... Such calculations are made much easier with the dependent variable because normal arithmetic can be used rather than harmonic arithmetic as the (rather convoluted) examples in the video illustrate. One of the worst examples of using the independent variable is in the "efficiency" display where it completely obscures the effects of regenerative braking,

However, in the EV world, where the point of concern for a trip is much less about the cost of the trip (at least for many), but rather on the range provided by the energy available then suddenly the range per unit of energy (e.g., Miles per KWh) is suddenly much more helpful in thinking about just how far can I make it.

The question of how far one can go on a charge or how much charge one must add to pick up a certain amount of range are most easily handled by knowing nominallly how far one can go on 1% of his battery (about 3 miles in an R1T) and how much that varies under adverse conditios (e.g 2.8 miles in rain?). Then it is a simple matter to realize that if one's next leg is 100 miles he'll need about 33% SoC or if he wishes to add 100 miles he'll need to add 33%. Some (the mathematically inclined) pick up on this pretty quicly. Otheres must rely on the miles estimates offered by the vehicle.

 

[lostpacket]

Diehard mi/kWh fan here. And yes, I do the mental math constantly while driving my 110 mi rated range car. Until I get down below 25% I don't trust the car's estimated range. I live near/in mountainous areas and my consumption changes dramatically depending on where I drive and at many points along a given trip. I can look at current %SoC and know that for a given type of road I'll get around 4 mi/kWh or going up a canyon I'll get 2-2.5 mi/kWh and adjust accordingly. I spose with a big enough battery I wouldn't care as much, but that's not been my EV journey to date with small battery vehicles.

If the range estimate is inaccurate, shouldn't the solution be better range estimates? By all accounts so far it seems the Rivian is good at this.

Do you think you might have just developed an intuitive sense of what is "good" and "bad" efficiency for your vehicle based on deviation from a norm? The way you describe it sounds like that is what you're doing.

know that for a given type of road I'll get around 4 mi/kWh or going up a canyon I'll get 2-2.5 mi/kWh and adjust accordingly.

Can't you get this sense if the equation is flipped? if it said 200 wh/mi normally, then going up a canyon it says 300 wh/mi -- is that really so much harder to reason about?

Edit: I flipped my example numbers in my head and now I'm begging to see where you are coming from...

I am new to EVs so I am happy to learn more about this, and update my thinking, and happy to hear from folks who have actually driven them. So please feel free to inform me here.

Have you driven other EVs that showed you the efficiency in another format? Some folks on the reddit discord mentioned Teslas have it displayed as wh/mi. And they also had the problem of inaccurate range estimates. So I can see in these cases, there might be some benefit. But I am trying to understand it a bit more.

The concern for me is power/distance makes it easier to compare vehicles to each other and to past version of the same vehicle to understand improvements. If we all stay with distance/power we may lose that intuitive sense for comparing.

Anyways thanks for the haha reaction I guess? Not sure what was funny about my original post but don't want to read too much into that.

 

[ajdelange]

Can't you get this sense if the equation is flipped? if it said 270 wh/mi normally, then going up a canyon it says 135 wh/mi -- is that really so much harder to reason about?

The place where its value is seen if you start to go up a hill and you know this adds X Wh.mi AND it starts to rain and you know that adds Y Wh/mi.

 

[OverZealous]

I think the real issue here, as most things, is fighting for a "single" right solution. Each has their place. If you are trying to estimate how far you can go, knowing mi/kWh is helpful.

If you are trying to compare the efficiency of two vehicles, then only kWh/mi (or kWh/100miles) makes any sense at all. Otherwise you end up in the same boat we're in now with gas vehicles.

But these things are massive computers. So why worry about which is shown?? Show both! Or let the driver toggle between them, at the least. But for window stickers, we should really focus on kWh/100miles and gal/100miles.

 

[zefram47]

Can't you get this sense if the equation is flipped? if it said 200 wh/mi normally, then going up a canyon it says 300 wh/mi -- is that really so much harder to reason about?

Edit: I flipped my example numbers in my head and now I'm begging to see where you are coming from...

I don't have a "tank" of miles...I have a tank of kWh and given a certain amount in the tank I know how far I can go. Wh/mi doesn't really do that without the extra step of taking the reciprocal. The math is just simpler.

The concern for me is power/distance makes it easier to compare vehicles to each other and to past version of the same vehicle to understand improvements. If we all stay with distance/power we may lose that intuitive sense for comparing.

I also don't care about comparing against other vehicles. I care about how my current drive compares with the previous or next. And it still works out...a vehicle getting 4 mi/kWh is still twice as efficient as one getting 2 mi/kWh.

 

[MXA121]

Distance over Energy works just fine to understand efficiency and compare vehicles. (How much more efficient is 20mpg than 10mpg? 100% more. Not difficult.)
No need to change the system. Especially when the proposed system doesn't roll of the tongue at all.

MPGe is the best we have right now and does a good job comparing ICE and BEV energy efficiency.
Remember your car never tells you how many gallons or Kwh are in there...you only know what you put in at the time.

 

[ajdelange]

Remember your car never tells you how many gallons or Kwh are in there...you only know what you put in at the time.

It does indeed. The BEV estimate is doubtless quite a bit more accurate than the ICE vehicle's estimate. As I indicates in an earlier post setting your gas gauge for percent and your consumption to wH/mi (if you can and you can't with a Rivian to their disgrace) allows you do do a whole bunch of simple check calculations in your head which give excellent indications as to true fuel condition at any point along a journey - if you want to. Most people don't want to and that's fine. Mi/kwh is the "for dummies" setting available in many BEV but of those many (but not all of them) give the cognoscenti the choice of the much more meaningful Wh/mi.

 

[MXA121]

It does indeed. The BEV estimate is doubtless quite a bit more accurate than the ICE vehicle's estimate. As I indicates in an earlier post setting your gas gauge for percent and your consumption to wH/mi (if you can and you can't with a Rivian to their disgrace) allows you do do a whole bunch of simple check calculations in your head which give excellent indications as to true fuel condition at any point along a journey - if you want to. Most people don't want to and that's fine. Mi/kwh is the "for dummies" setting available in many BEV but of those many (but not all of them) give the cognoscenti the choice of the much more meaningful Wh/mi.

...So you multiply your % remaining, times vehicle 'tank' size, then divide by energy/mile to get estimated range?

Example: 20 mpg (5 gal/100mi) half tank (20 gallon total)
1. 0.5 x 20 gal = 10 gal, 10 / 5 = 2, 2 x 100 = 200 miles range
2. 0.5 x 20 gal = 10 gal, 10 x 20 = 200 miles range

If you can do option 1 faster in your head than option 2 you must be smart.

 

[ajdelange]

...So you multiply your % remaining, times vehicle 'tank' size, then divide by energy/mile to get estimated range?

Let's not be too silly. In the R1T Large context a percent is 3.14 (pi!) miles EPA. If you have 15% battery left you have a range of 45 miles. If your next leg is 190 miles you know you have to charge 63% and if you know your charger does the nominal 1C charge that it's going to take 63% of an hour (38 min) to pick it up.

The EPA consumption is simply 135000/314 = 430 Wh/mi. If you know it is going to rain on this next leg and have learned from experience that this usually adds 30 Wh/mi to rolling resistance then it's apparent that CONSUMPTION is going to be about 8% higher than EPA and so you had better add 8% to your charging time and make it 41 minutes.

Can you figure this out using harmonic arithmetic? Sure but why do it the hard way? The rest of the world has figured this out. Why can't we?

Sponsored