Fraud By Algorithm?

Fast Eddie B

Fast Eddie B

Well-Known Member
OK. I'll stipulate fraud is a strong word. But consider this...

The car knows exactly how far you've traveled, and I assume it knows exactly how much fuel its burned over any given distance. So calculating average mpg exactly over any given trip should be trivially easy.

And yet, we just took a trip to Florida and back and kept careful records of the 7 tanks of gas burned. More on the trip in a dedicated post to follow, but it revealed that the car consistently overestimates gas mileage:

32357308337_7747ae43a2_o.png


The left column is the actual mpg from dividing distance traveled by gals consumed. The "Est" number is what the display in front of the driver showed before resetting Trip A. As an aside, over the entire trip the EV range dropped 24 miles (from 44 to 20) in spite of being in HV mode the entire trip, but I'm not clear how to factor that in.

I'm generally not a conspiracy theorist, but can anyone think of an explanation other than Honda made a conscious decision to pad the mpg to make the Clarity seem more fuel efficient than it actually is? It really seems like they wrote an algorithm to calculate gas mileage, and then added about a +12% multiplier to it.

Thoughts or other theories/explanations? And has anyone else run a similar trial? In any case, it seems like Honda and/or some consumer advocacy group should be made aware of this.

My working assumption is that Honda is not alone on doing this, but it still seems wrong.
 
Last edited:
I have found exactly the same as you and have written it before. No different with my previous 16 Honda Civic and 13 Accord FWIW.

Honda’s mpg dashboard display is notoriously exaggerated on every Honda I’ve ever owned. Always. Clarity is the worst of all of them.
 
Interesting data! What device would the car have (or need) to measure exactly how much fuel was used? The estimate seems especially tricky in a hybrid that occasionally shuts off the ICE or uses it for battery charging. I wonder if in other travel circumstances the estimate would be too low? Anyway, just as some have called “fraud” because they didn’t get exactly 47 miles EV per charge all the time may not mean Honda purposely rigged the numbers.

I know both Toyota and Subaru over-estimate mpg on the vehicle’s computer.
 
To be clear, I'm over-the-moon ecstatic about the fuel mileage we got! 42+ mpg in a full-sized, well-equipped sedan is excellent. It's about twice the mileage we got when we took a similar trip in a turbo EcoBoost V6 Ford Flex. Saving roughly $100 in fuel at a similar level of comfort (if somewhat less raw performance).

I did not mean to imply I was unhappy with the mileage and I'm sorry if my post read that way. I just don't like purposeful misrepresentation of data, in an apparently systematic way.

Again, my working assumption is the car has a fuel-flow measuring device somewhere - how else would it give the instantaneous mpg readings it purports to show? And if its just a matter of imprecision, I would expect the car's estimates of mpg to cluster both below and above the real number with no consistent trend.
 
Fast Eddie B said:
OK. I'll stipulate fraud is a strong word. But consider this...

The car knows exactly how far you've traveled, and I assume it knows exactly how much fuel its burned over any given distance. So calculating average mpg exactly over any given trip should be trivially easy.

And yet, we just took a trip to Florida and back and kept careful records of the 7 tanks of gas burned. More on the trip in a dedicated post to follow, but it revealed that the car consistently overestimates gas mileage:

32357308337_7747ae43a2_o.png


The left column is the actual mpg from dividing distance traveled by gals consumed. The "Est" number is what the display in front of the driver showed before resetting Trip A. As an aside, over the entire trip the EV range also dropped 24 miles (from 44 to 20) in spite of being in HV mode the entire trip, but I'm not clear how to factor that in.

I'm generally not a conspiracy theorist, but can anyone think of an explanation other than Honda made a conscious decision to pad the mpg to make the Clarity seem more fuel efficient than it actually is? It really seems like they wrote an algorithm to calculate gas mileage, and then added about a +12% multiplier to it.

Thoughts or other theories/explanations? And has anyone else run a similar trial? In any case, it seems like Honda and/or some consumer advocacy group should be made aware of this.

My working assumption is that Honda is not alone on doing this, but it still seems wrong.
Click to expand...

I've had this frustration with every vehicle I've owned over the last 20 years. In theory, you are correct in thinking that the car should know exactly how much fuel is used. In a fuel-injected system, this would seem to be easy by doing something like counting injector pulses. However, they do not do that, and instead rely on indirect methods to estimate consumption like measuring vacuum, RPM, etc. I'm not sure exactly how Honda does it, but they do not measure flow directly. Before I bought my Clarity last November, I owned a series of Prii that consistently overestimated MPG by anywhere from 5%-15% like clockwork. Drove me nuts! So, no, I don't think Honda is padding the MPG estimates, its just that they use a poor estimation algorithm, just like every other automaker!
 
I just meant that by the end of the trip, the meat computer between my ears was just subtracting roughly 5 mpg from whatever the display said, and coming up with a number more reflecting reality. If Honda wanted, and assuming their engineers are smarter than me, and that accuracy is their goal, they could easily plug in roughly the same correction factor. That they choose not to seems revelatory.
 
I am not sure I would say it is just Honda. I have had a Chevy, Dodge, Nissan, Kia and BMW. They have all overstated their MPG by 5-15%. Hasn't anyone else noticed this? Of course with my luck it could be just me. Given that with fuel injected engines the computer know how much gas is consumed to the gram it must be a conscience decision on their part to increase the number. Why they would do it is beyond me, why give a value when someone with 5th grade math training can disprove it. Maybe it is one of those alternate facts that I have been hearing so much about.
 
My recent experience:
2016 Honda Fit EX Car: 38.1mpg, Actual: 36.3 mpg, 5% difference (11,000 miles)
2017 Honda Civic Si Car: 36.8, Actual: 36.9, just scary accurate! (25,000 miles)
2018 Honda Clarity Car: 103.1, Actual: 86.9, 18% difference (thru 2,500 miles so it hasn't reached a stable statistical place yet.)
 
I really think the EPA reported MPG in all vehicles is assumed in conditions quite different than most real life scenarios. Payload of the vehicle (150 pound driver going solo vs a full hefty family with a trunk full of suitcases), wind, outside temperatures and climate control use, tire size/air pressure, traffic conditions, gasoline quality and many more, vary a heck of a lot more than some computerized, hangar assessed value.
Even if you drive mostly at night (bcs of headlights using more juice) I would expect it would affect the mpgs from a day driver.

I'm happy with anything over 40mpg and I've seen it every day since I commute 70miles daily in HV. While on topic, using HV charge mode is a WHOLE other proposition to the expected mpgs.
 
My question is "Why the battery loss?"

Most gurus on this forum claim that in HV mode Clarity (thru its ICE) maintains the battery charge at the set point of switching from EV to HV. So if you departed in EV with 44miles charge, and switched to HV at that point, the assumption it is your ICE would work to maintain a charge of ~44 or so for the duration of HV drive. Maybe small drops here and there that would be recovered by ICE, but not a 20+ (halving of range) without never switching to EV during the trip?

What say you?
 
Thanks all for bringing up that many other cars have the same problem. My Clarity is the first car I've owned that makes this calculation and I didn't know that "everybody does it". dp
Dante said:
My question is "Why the battery loss?"

Most gurus on this forum claim that in HV mode Clarity (thru its ICE) maintains the battery charge at the set point of switching from EV to HV. So if you departed in EV with 44miles charge, and switched to HV at that point, the assumption it is your ICE would work to maintain a charge of ~44 or so for the duration of HV drive. Maybe small drops here and there that would be recovered by ICE, but not a 20+ (halving of range) without never switching to EV during the trip?

What say you?
Click to expand...
I average using about 30% of full charge per each tank of gas when in 100% HV mode. Not a big deal to me.
 
hi David T - my point is not about big deal or not... it was about the fact that such loss as described by OP and (looks like) second by you as well, goes against a lot of technical and convincing discussions (in my opinion and understanding) of how the EV/HV and battery and ICE work together in this vehicle.

Add to that the observation from my humble experience (70 miles HV daily) where I loose no EV capacity whatsoever - if I Switch to EV at 44mark, it stays there +/- 1-2miles, and same for switching in the 30s, 20s, teens and single digits.

What gives with you guys? Is there something using the battery that cannot be compensated by ICE? Smart people here would say you should hear the "bees" before the significant drop as the ICE will do all in its might to maintain that point level...

The answers may assist us all in better management of some behavioral changes - or nothing happens and we're all good as is.
Cheers
 
I think you will find many more on this site with the same experience as me, I think I recall Insightman for one saying the same thing on other threads.
 
I have found on my trips back and forth across the mountains that the EV range does stay pretty constant once I go to HV mode. So this gradual, persistent loss on a longer trip surprised me a little bit.

As an aside I had three opportunities to charge overnight on my trip. That certainly would have taken care of the issue. But I declined as I wanted to see how their car performed as a straight hybrid.
 
Fast Eddie B said:
It really seems like they wrote an algorithm to calculate gas mileage, and then added about a +12% multiplier to it.
Click to expand...

Have you noticed that the MPG is 199.99 when driving in the EV mode and the ICE has not ignited?
This 199.99 mpg is averaged in whenever the ICE is not igniting. IMHO.

The true mpg will be obtained when you drive without plugging in. IMHO.

I will try this in my next 45 mile trip on March 9, 2019.
 
Similar to the EV (only) Estimate, HV estimates are in MILES units. However, 100 miles with no AC/seat heater/anything else using POWER will not consume the same POWER as 100 miles with AC/seat heater/anything else on full blast. While on HV (HYBRID vehicle) mode, you're using your gasoline not only to move your car but also to generate energy back into the battery, that's Also being used (hence, Hybrid) along with your gas engine. For all those tank fill ups, were everything else in the car the same, in terms of power usage? And since this was a long distance trip, did all fill ups encounter the same road/driving conditions? Similar to power usage, 100 miles uphill is not the same 100 miles downhill or flat roads.
 
You must log in or register to reply here.
Back
Top