Hacker News Reader

Comment by sdfignaionio | Hacker News Reader

There are two important things you're missing:

1. Tesla's estimates are not "highly accurate". They are alleged to use a deliberately optimistic estimate when the battery is near full, then gradually switch to a more realistic estimate as the battery drains. In particular they are alleged to incorrectly account for the weather. This means the car will always display nearly the advertised number at full charge (and on a test drive). Supposedly Tesla used to use a more accurate estimate but then got orders from on high to fudge the numbers.

2. For other car manufacturers, the range estimates are not terribly inflated. The report from Edmunds claim that most EV models meet or exceed their range estimates, and cars that miss their estimates don't miss by anywhere near as much as Teslas.

View on HN

I know we are talking about EVs here, but my ICE has an estimated range display and _it_ isn't super accurate _and_ it doesn't take potential weather (or other factors like elevation) into account either. Why would you expect it to factor in _potential_ weather when giving a range estimate?

It's not the potential weather, the one mode for the range estimator apparently isn't taking the current ambient conditions into account. So it's not factoring in the likely load from the heat or AC when it's cold or hot and they need to run to keep the vehicle comfortable.

Allegedly,

1. Other companies have more accurate estimates than Tesla.

2. Tesla uses a different and more accurate algorithm when charge is low.

3. Tesla's estimates used to be more accurate.

It doesn't make sense to argue the problem is intractable when it has supposedly been solved by many companies, including Tesla itself.

>Why would you expect it to factor in _potential_ weather when giving a range estimate?

I wouldn't expect that. The cars have thermometers.

One important thing you are missing, there are two different estimates being referred to. The "miles remaining" display on the battery charge gauge (which can be toggled to show percentage instead) and the trip planner / navigation estimate. The trip planner is very accurate, you can view a burn down charge of realtime estimate as it changes while you drive and it says what component contributed what % to deviate from the initial estimate. The "miles remaining" display is an imaginary number that is only meaningful if you are driving the same highway mix as the EPA test in the same climate conditions

This is correct. If you tell it the destination the range estimate is astonishingly accurate. And it will list all the reasons it turned out to be wrong down to what fraction a headwind impacted range.

Reply on news.ycombinator.com