Fraud By Algorithm?

Discussion in 'Clarity' started by Fast Eddie B, Mar 6, 2019.

  1. Fast Eddie B

    Fast Eddie B Well-Known Member

    Original tires with maybe 7,000 miles on them. Tire pressure checked regularly, and right before this trip. But wouldn’t any “tire circumference error” equally effect both the car’s display and calculated mileage?

    Sure, lots of variables can creep in. But again, if the amount of fuel actually added varied, one would expect errors in both directions tank-to-tank. It’s the consistency in the error in one direction that is suspicious, not that the car’s calculation isn’t perfect.
  2. ClarityBill

    ClarityBill Active Member

    I am guessing the manufacturers want the 'casual observer' to feel the best they can...

    If the manufacturer reports the most accurate number possible (some up and some down), the person that never checks the accuracy (my definition of a casual observer), would be more disappointed with the mpg of their car.

    With the system that is installed, the casual observer can report to their friends that they are never less than 42 mpg, rather than the true 'sometimes less than 40 mpg'. It would be even worse if the reported number was lower than the true number.

    It is interesting how consistent it actually is: It is consistent enough to give an indication when changing the operating method has actually improved mpg.
  3. Tim66

    Tim66 Active Member Subscriber

    I still don't get what all the consternation is about. The OP got .6 MPG better than the EPA MPG posted on the window sticker. Everyone I know would be very happy if their car got the EPA predicted MPG. Computer estimates are just that and cannot take into account every variable. Honda can tweek the computers to give a better than expected MPG but when the expected MPG is printed on the window sticker supplied by Honda what's the point? The OP's real world results are factual and all that really matter.
  4. Fast Eddie B

    Fast Eddie B Well-Known Member

    For the nth time...I AM VERY HAPPY with my actual mpg!

    For the nth time...We’re not talking about estimates or projections. We’re talking about a modern computer that’s taking an input of 200 miles traveled and 5 gals consumed and outputting 45 mpg instead of 40 mpg. This “post facto” calculation is not rocket science, and the best explanation I can come up with is Honda is “stroking” its buyers by exaggerating fuel efficiency. As ClarityBill so eloquently put above.

    And I don’t appreciate being “stroked”. At least not in this context!
    Walt R, David Towle and sniwallof like this.
  5. ClarityBill

    ClarityBill Active Member

    I am really fascinated by your trip data... When do we get to see more info on the trip?

    Have you come up with reasons for the variation in mpg? Any idea why your mpg was 46 for a time, then 37.5 later?
    (I produce ethanol, so I am really curious about experiences with ethanol. Do you know the % ethanol of the fuel?)
  6. Fast Eddie B

    Fast Eddie B Well-Known Member

    Our N GA home is at about 1,800’, so there’s an overall elevation loss of about that much driving south. And of course an equal gain driving home.

    And the last two legs had a 15 to 20 mph nearly direct headwind.

    Beyond that, maybe 85% Interstate, cruise control set to 65 to 75 mph. The rest mostly 2 and 4 lane state roads at 55 to 65 mph, plus some city traffic thrown in in Ft. Lauderdale, Jacksonville, Tallahassee and the slog both ways through Atlanta traffic.

    I did not track what percentage of ethanol was involved, though around here it’s usually 10%.

    Try to consolidate all this into a trip report later today.
  7. Dante

    Dante Member

    And I don’t appreciate being “stroked”. At least not in this context![/QUOTE]

    I take stroking any way I can... lol
    and I stand with Fast Eddie - that calculation (after the fact as he states) should be accurate. It's a flaw!
  8. ClarityBill

    ClarityBill Active Member

    Looks like that headwind had a big impact. It would be nice to have a way to measure that - relative wind speed vs. road speed.

    Any idea how much driving was in gear mode? Did the headwind push you out of gear mode?
  9. Thomas Mitchell

    Thomas Mitchell Active Member

    The manufacturer comment about ethanol got me to thinking about variances in energy density for non-ethanol gasoline. Not all “gasoline” has the same energy density. The energy density of the gas you put in your car will vary every time you fill the tank. Even if you use the same octane rating every time you fill, the blends will vary and so will the energy you get from that tank. I’m not certain the cars computer can determine this factor and the MPG algorithms are probably determined using a constant value for energy density of gas. This value is probably “optimistic”, or a best case value and may rarely be seen in actuality. I don’t disagree that the practice of overestimating MPGs is intentional, but I would say the manufacturers can probably make a decent case for the values they choose. Ultimately, it is still an “estimate”, regardless of the precision with which certain inputs can be measured, and, no doubt, they would prefer to overestimate than underestimate. Here is another link, one that shows the variances in energy for different fuels. Also, labeling of ethanol content varies from state to state and many of us may be pumping ethanol without even knowing it.

    Sent from my iPhone using Tapatalk
  10. ClarityBill

    ClarityBill Active Member

    Thomas Mitchell:

    The display should not use any estimate of energy value, just the actual gallons consumed, and the miles traveled.

    The energy content varies, so the actual mpg will vary, but the discussion here relates to whether the display should correctly show the actual mpg achieved. The level of stroking seems to be the point of contention.
    Walt R and David Towle like this.
  11. Fast Eddie B

    Fast Eddie B Well-Known Member

    I was about to say the same thing, and reinforce it with a “reductio ad absurdum” thought experiment. To wit...

    Imagine one tank of gas ended up having only 50% of normal gasoline energy, due to massive amounts of ethanol or whatever. For that tank, due to the lower energy, the car only went 100 miles on 5 gals instead of its normal 200 miles or so.

    In that case, I would expect the mileage display to dutifully show 20 mpg, which would be accurate and reflect the crummy fuel. It would still just take distance traveled and divide by gallons consumed. And, in the case of the Clarity, probably show about 22 mpg on its display.
  12. Jordan

    Jordan Member

    Oh my, I'm so glad this topic is here. I've always been bugged by this car and every car I've owned being so far off on their MPG. It just doesn't seem that hard to me. I'm glad I'm not the only one bugged by it. Again, I love my clarity and love my MPG. I just wish it told me what it actually was getting.
  13. insightman

    insightman Well-Known Member

    I've been keeping a notebook in my Insights for 20 years. Every month when I go to the gas station I write down the mileage and gallons. It would be nice if our gen-1 Insight, our gen-2 Insight, or Clarity was as smart as my notebook, but I've accepted the discrepancies.

    We keep track of the gas we put into our Clarity, too, but it's such a rare occurrence that I never bother to calculate the numbers. Our Bosch 40-amp EVSE isn't one of those smarty-pants units that tracks how much electricity we use, so we drive on in blissful ignorance, pretending we're driving around for free. Eliminating trips to the gas station for a stretch of 10 months was a dream come true for us. The monthly gas-station visits for the Insights seem so prehistorically bothersome now (although it does give us an excuse to go to Costco).
    sniwallof and KentuckyKen like this.
  14. 4sallypat

    4sallypat Active Member

    Coming from a car that got 22MPG (2L 4 cylinder), the Clarity far exceeded my expectations for gas mileage.
  15. The Gadgeteer

    The Gadgeteer Active Member

    Just 12%. Better or the same compared to most cars. It could be a conspiracy among all the car manufacturers.

    Sent from my iPhone using Inside EVs
  16. ukon

    ukon Member

    What a great data, Fast Eddie. Thanks for sharing. And I agree with the assessment that Honda and all car manufacturers do not want to show the actual data to consumer even though they have access to it post-facto.

    I have very similar data for city driving post battery discharge if interested. The car display always shows 45-55mpg for stop and go traffic but the reality is around 38-42. I need to hit atleast 3 miles before it goes above 40 per single trip. I can do these experiments as my work is 2 miles, grocery - 1.5 miles, daycare/school - 4 miles, weekly once grocery 5.5 miles. So I did have only 2 empty tanks(with empty EV range) to experiment this.

    I do have some more thoughts on this. With EV depleted I often see car dynamic mpg meter in display showing 20-30 while accelerating from stop, settles at 30 and the switches to EV on next stop sign. Then EV depletes again for another half miles and goes back to 20-30 mpg. I have only seen it stay at 40mpg when I am steady running the car at 35mpg for long distance.

    BTW I think gasoline engine is far worse in my use case. Elantra rated at 25 but gets 21. Infiniti rated at 19 gets 14. Those 2-3 mile trip are killers.
  17. ukon

    ukon Member

    Also, I think most cars I have driven in recent years display close to rated mpg and get worse(rated 30, shows 29-32, gets 27). Clarity may be an exception in getting close to EPA rating and in that sense at least better than others. But inflated display for sure.
  18. LegoZ

    LegoZ Active Member

    I concur with both and tested this last trip home from Illinois to Texas. One tank I made sure not to shut the car off at ALL at any stops rotating people in and out to have someone in the car. The car nearly perfectly held range from when I first pressed the HV button. The problem is when you run in to the bathroom and have to shut the car off. It does reset that HV hold setpoint. It would be nice to be able to tell the car what you want for EV range and it get it there and maintains it.
  19. Edward Dries

    Edward Dries New Member

    Here is a novel idea for auto manufacturers.
    Provide a MPG "offset" in their software settings so the consumer could tweak (up or down) the computed MPG results. Easy Peasy.
  20. ClarityBill

    ClarityBill Active Member

    Or just provide a sticker that goes over that part of the dash that says 44 mpg....

Share This Page