Erroneous Average (simple math) in Efficiency History

[geraldmw]

The graph of efficiency history is wrong. I teach 4-8th grade math and my students could pick the ERROR!
The SIMPLE average of a series of averages is NOT the same as the overall average. The correct answer is a WEIGHTED average.



Here is the explanation:
Say for each of the 5 mile ranges I get 5 miles per kWh, but for one I get 250 miles per kWh (the picture attached shows this!)

The total distance traveled is 50miles.
The total energy consumed is (5/5*9 + 5/250*1) = 9.02 kWh.
The CORRECT average is 50/9.02 or about 5.54 miles per kWh.

The graph displays the simple average, specifically (9*5 +250*1)/10 = 29.5 miles per kWh.

This should be fixed! Even a fourth grader can spot the error

 

[geraldmw]

I wonder if anyone from Chevy actually sees this?

 

[GetOffYourGas]

:lol:

This is great. As an engineer, I found this both amusing and bewildering that they made such an obvious mistake. Was that code written by an intern?

 

[ScooterCT]

I'm currently a software engineer (30 years now), but my college major was English. You'd be surprised how many programmers have poor or atrophied math skills.

I have a red face saying this, but I could have easily made that mistake. And I've written aircraft navigation software using spherical trig. Somebody tells you to put a line on the graph for an average. You're in a hurry. You type "AVG" or something, and the code editor inserts some standard function for average. It puts a line on the graph. You move on to the next of the 100 things you have to implement today. QA sees a line on the graph. They move on to the next of the 100 things they have to do today.

The code that runs your life is filled with stuff like this. Programmers are in a hurry these days. Lots of us are under-qualified. Sometimes I'm amazed any of it works. I'm not excusing it. Just explaining why it happens.

s

 

[geraldmw]

I'm currently a software engineer (30 years now), but my college major was English. You'd be surprised how many programmers have poor or atrophied math skills.

I have a red face saying this, but I could have easily made that mistake. And I've written aircraft navigation software using spherical trig. Somebody tells you to put a line on the graph for an average. You're in a hurry. You type "AVG" or something, and the code editor inserts some standard function for average. It puts a line on the graph. You move on to the next of the 100 things you have to implement today. QA sees a line on the graph. They move on to the next of the 100 things they have to do today.

The code that runs your life is filled with stuff like this. Programmers are in a hurry these days. Lots of us are under-qualified. Sometimes I'm amazed any of it works. I'm not excusing it. Just explaining why it happens.

s

Now the bug has been reported I hope they actually fix it.

 

[phil0909]

Now the bug has been reported I hope they actually fix it.

Hope springs eternal in the human breast.
:roll:

 

[user 1340]

I understand the averaging error, but I'd like to know what the hell happened that the car thought you went 250 miles and used only 1kwhr! Did you coast 5 miles downhill?

 

[user 1034]

I understand the averaging error, but I'd like to know what the hell happened that the car thought you went 250 miles and used only 1kwhr! Did you coast 5 miles downhill?

According to that graph, it was for 20 miles coasting downhill.

 

[GetOffYourGas]

I understand the averaging error, but I'd like to know what the hell happened that the car thought you went 250 miles and used only 1kwhr! Did you coast 5 miles downhill?

According to that graph, it was for 20 miles coasting downhill.

Actually, it's a 5-mile increment, at miles 15-20.

 

[davidjmel]

While my Bolt does not have this error, this error also occurs in the Mychevrolet.com online web-site for the Bolt (I have a Bolt since April 26) and also on the Mycadillac.com website (for the Cadillac ELR I own since 12/15). I assume it is true for ALL GM electric car owners, since the software is IDENTICAL. Perhaps if enough people on this forum complain they will fix it. The software update done in March to these sites makes the data total nonsense (hundreds of KWH per hundred miles). I have called onstar, and emailed Mary Barra. I have SEVERAL times had the issue escalated but never fixed. Note in the screen shots below, the ELR has ~12,800 miles and the Bolt has ~3,300, neither shows lifetime data either.

Suggestion: connect your car to your home wifi and check for a software update. I have the most recent version and the in-car software (and the android apps for my chevrolet and my cadillac both display correct data)

 

[user 1034]

Suggestion: connect your car to your home wifi and check for a software update. I have the most recent version and the in-car software (and the android apps for my chevrolet and my cadillac both display correct data)

This does not work at this time. Updates can only be performed by a Chevy rep at this time. They have not enabled OTA updates for anything.

 

[davidjmel]

Interestingly, it does allow you to check for the version, to SEE if an update is available. I received the car with up-to-date software, so I cannot confirm if it allows an update or not. I think if you see an update available you could show the error to the dealer's service dept and they would install the update for you.

 

[joe]

I understand the averaging error, but I'd like to know what the hell happened that the car thought you went 250 miles and used only 1kwhr! Did you coast 5 miles downhill?

I may have a small insight into this question. I did a fast charge for about 1 hour after I got my battery down to maybe around 20-25%. After that I drove home from the fast charge station (about 1-2 miles) and did a slow charge with the stock EVSE for a long time, like 12 hours or more.

The next day, I was looking at the efficiency history graph and perplexed to find the "present" bar at almost zero (like maybe 0.2, it's hard to tell on that graph). I thought maybe something was wrong, so I reset the efficiency history screen. The "present", i.e. last 5 miles, average bar was still stuck close to zero as before. I had hit the reset button at a stop sign and after accelerating away, the upper limit on the y-axis of the graph immediately shot up to 252 mi/kWh, and the present bar went all the way to the top, and started going down quickly and in a few seconds, the upper limit went to 151 mi/kWh. Shortly after the stop sign, I came to a red light, and the regen made the bar start moving up again, finally to 252 and then over, after which point it went back to almost zero on a graph with upper limit 8 (or 4) or whatever the minimum is.

So apparently, 252 mi/kWh is the maximum efficiency that the bar graph will record. Over that, it goes to zero. I'm guessing the actual limit is 400 km/kWh.

The question is how did this person get it to go exactly to 252 and not over.

Another question is if my long story about charging is relevant to this matter -- maybe I left the car on while it was charging and it took the charging into account in the efficiency history as something close to infinity mi/kWh.

 

[geraldmw]

I have most recent version of the software (vrs 13.4.0) as of 5 minutes ago.
I believe the issue still exists!

Would be great if Chevy behaved like a real company and issued release notes with software upgrades (and made it possible to see a release history!) I guess they are a car company at heart. The idea of updating a product after selling it may be beyond their comprehension!?

 

[geraldmw]

So apparently, 252 mi/kWh is the maximum efficiency that the bar graph will record. Over that, it goes to zero. I'm guessing the actual limit is 400 km/kWh.

The question is how did this person get it to go exactly to 252 and not over.

I am "this person" --- The 252 mi/kWh is what happen on a long coast of downhill. The actual place where I do this is about 5.3 miles and I am there once every two weeks, so have to be 'lucky' to hit it just right. It is also 30 miles from home (where I charge) so I do not believe it is related to the charge. I should point out I have never seen the efficiency "go over the top" of the range and return to zero.

 

[gpsman]

Mine always maxes out at 252 on long & steep downgrades. But why not 255 (0-256 bytes)

My useage screen has been stuck for a long while at 1272.6 miles, while kWh continue to climb.

1272.6 miles = 2048 km.

Exactly 2 kb of memory.

It is like some memory register gets full.

I always get 4.0 miles per kWh (+/- 0.2).

 

[jeffhqb]

My Bolt also max'ed out at 1272.6 miles while the energy usage continues to go up. So I can't get a reliable energy usage per mile from that screen. And since the histogram gives an incorrect figure for the average Miles/kWh I can't rely on that. I have to believe Chevrolet will fix these issues at some point because I really would like to know the car's real efficiency.

 

[EastWestBrothers]

The graph of efficiency history is wrong. I teach 4-8th grade math and my students could pick the ERROR!
The SIMPLE average of a series of averages is NOT the same as the overall average. The correct answer is a WEIGHTED average.



Here is the explanation:
Say for each of the 5 mile ranges I get 5 miles per kWh, but for one I get 250 miles per kWh (the picture attached shows this!)

The total distance traveled is 50miles.
The total energy consumed is (5/5*9 + 5/250*1) = 9.02 kWh.
The CORRECT average is 50/9.02 or about 5.54 miles per kWh.

The graph displays the simple average, specifically (9*5 +250*1)/10 = 29.5 miles per kWh.

This should be fixed! Even a fourth grader can spot the error

Driving the Grapevine or going down the hill in RPV? I've only had this happen to me once and it was a steep extended downhill on Crenshaw from RPV to Harbor City.

 

[geraldmw]

Driving the Grapevine or going down the hill in RPV? I've only had this happen to me once and it was a steep extended downhill on Crenshaw from RPV to Harbor City.

The issue is not the actual bars on the graph .. it is the calculation of the height to place the "red line."
The high miles per kWh bar is correct (might be higher, but I think the system caps it.) The problem is the red line is incorrectly calculated (a simple math mistake!)