Bloomberg Model 3 Tracker

[ValueAnalyst]

The tracker largely depends on reader-submitted data on VIN sightings. However, they began realizing (and reported) that people were submitting VINs at time of assignment not knowing whether the cars had actually been built yet. It seemed Tesla started pushing out VIN notifications quickly presumably since that was the determinate in reservations becoming non-refundable. That practice has since stopped. While VINs are still being assigned buyers are not being given their VINs until just days before the scheduled delivery. So the data should become much more reliable going forward. That could explain the drop in reported VINs for the coming weeks.

You really take trolling to the next level, Donn. Lately, literally nothing that has come out of your mouth/fingertips has been true. No wonder why some people have left TMC altogether to have higher quality discussion on Twitter!

 

[Chickenlittle]

Anecdotal but just at Tyson corner (tyco rd) dealorships and a lot of model3 in lot with multiple people picking up cars. Also passed 4 model 3 with temp plates on way out.

 

[ggr]

The tracker largely depends on reader-submitted data on VIN sightings. However, they began realizing (and reported) that people were submitting VINs at time of assignment not knowing whether the cars had actually been built yet. It seemed Tesla started pushing out VIN notifications quickly presumably since that was the determinate in reservations becoming non-refundable. That practice has since stopped. While VINs are still being assigned buyers are not being given their VINs until just days before the scheduled delivery. So the data should become much more reliable going forward. That could explain the drop in reported VINs for the coming weeks.

For the record, a number of things in the above statement are false.
1. The reservation becomes non-refundable when the actual configuration is finalized. The VIN isn't released until later.
2. The change by Tesla from telling the customer when the VIN was assigned to only telling them when the car was in late production was done some months ago.
3. ... hence cannot explain the drop in reported VINs. I don't know what does, but this isn't it. The most logical explanation (to me) is some other change in Tesla's internal procedures.

 

[dondy]

One of the basic rules in any modeling is proper scaling. Using single digits in models accurate to hundreds is preposterous.

I can continue, but this crap is beyond boring. As I was writing before I don't believe they use any "model". Current 6k per week come from the sales rates in USA which will reach 25k+ in august.

 

[neroden]

I currently believe that there are known supply bottlenecks which account for why the recent production rates have been higher than the projected quarterly rate would seem to indicate. The leaks I've heard indicate that the GA line has been let off early on some days, which says to me that something further back in production is bottlenecking. I'm sure they're working on it but I wouldn't expect it to be debottlenecked until Q4.

 

[jhm]

Tracker as of Aug 26 is reporting 4588 per week and 77,7740 cumulative.

So far this is 36,710 for Q3. Extrapolating to the whole month via QTD daily average is 58,607 and by current weekly rate on remain 34 days is 58,995. The two method are nearly identical because current weekly rate is close the average weekly rate QTD.

Either way, we are still on pace to 59k for Q3. The recent slow down is not severe enough to imperil 50k to 55k guidance.

 

[jhm]

One of the basic rules in any modeling is proper scaling. Using single digits in models accurate to hundreds is preposterous.

I can continue, but this crap is beyond boring. As I was writing before I don't believe they use any "model". Current 6k per week come from the sales rates in USA which will reach 25k+ in august.

What sort of modeling experience do you have? Any degrees in a quantitative discipline?

 

[dondy]

What sort of modeling experience do you have? Any degrees in a quantitative discipline?

Yes. physics of course.
but I don't see how it can be relevant. Degrees are crap everywhere. Regrettably "model massaging" and "clip thinking" is the norm nowadays.

"with faster computers we get better results" and "models prove" lol. I don't play this crap.

 

[jhm]

Yes. physics of course.
but I don't see how it can be relevant. Degrees are crap everywhere. Regrettably "model massaging" and "clip thinking" is the norm nowadays.

"with faster computers we get better results" and "models prove" lol. I don't play this crap.

I see. That explains the "scale" comment. Perhaps your physics degree is not relevant. You just don't sound like someone who has had to draw meaningful conclusions from messy real world data, but you've got plenty of attitude to make up for it. People with solid data analytic experience generally are much more inquisitive and respectful of the challenges of working with data.

 

[Tempus]

Anecdotal but just at Tyson corner (tyco rd) dealorships and a lot of model3 in lot with multiple people picking up cars. Also passed 4 model 3 with temp plates on way out.

I can confirm that as i was one of those picking up on friday

 

[Orbit]

Tesla is on track for a production record this quarter but slightly behind ambitious Model 3 goal

Electrek puts the number of M3 produced in Q3 at quite a bit lower than the Bloomberg tracker does.

 

[jhm]

Ok, just to keep tabs on the Bloomberg tracker, as of Sep 4, it is posting 83,575 cumulative and 4838 per week. With 27 days remaining in Q3, this is 42,545 qtd which may be extrapolated out to 60,494 using average rate this quarter or 61,206 using the current rate of 4838/week.

So initially these extrapolation were 59k and 68k. We seem to be closing in on the lower end around 61k.

Granted the Bloomberg sources and models may prove poorly calibrated. Fred claims to have a "reliable source" that says true numbers are lower than what Bloomberg reports. Fair enough, we will learn in early October who has the better sources. Unless, Bloomberg changes their analysis, they look to close in around 61k for the quarter end, just before Tesla releases their actual numbers.

We need to be prepared for how this plays out. Many people may take Bloomberg as more credible than Electrek. This could set them up to have expectations that are 8k or 9k above what proves actual. Would this then be counted as some sort of miss on Tesla's part? What sort of market reaction could this trigger?

For my part, I am uncomfortable going into the quarter end with such divergent estimates floated about by the media. It creates a situation where Tesla just can't win, if expectations are pushed too high by the media. To be fair, Tesla has guided 50k to 55k and may realize numbers well within that range. But if Bloomberg continues to over-report, that could create a problem for Tesla.

 

[Orbit]

I am uncomfortable going into the quarter end with such divergent estimates floated about by the media. It creates a situation where Tesla just can't win, if expectations are pushed too high by the media.

This is a key point I believe. "creates a situation where Tesla can't win".
Tesla "winning" is not necessarily in the interests of many stake holders, at least not within the time scale they are working on.
Lots of people win when there is confusion and widespread FUD.

Likewise, for many of us who support Tesla (and TSLA), we may have a tendency to go the other way. Overstating the case due to confirmation bias.


For example:
From what I can tell of the Bloomberg tracker, it's predictions are slightly lower than both your extrapolations.

I am assuming your calculated weekly average rate for Q3 so far, is around 4500 cars/week.
(83575-41040 =42525 cars in Q3)
Divided by 66 days
x 7 for weekly rate
= 4512 weekly average.

However, Bloomberg's forecasted production after this week, looks to be lower than that.
Something like 4150 per week.

So from what I can briefly estimate, Bloomberg's current forecast seems to be production of around 59,500 for Q3.

 

[jhm]

This is a key point I believe. "creates a situation where Tesla can't win".
Tesla "winning" is not necessarily in the interests of many stake holders, at least not within the time scale they are working on.
Lots of people win when there is confusion and widespread FUD.

Likewise, for many of us who support Tesla (and TSLA), we may have a tendency to go the other way. Overstating the case due to confirmation bias.


For example:
From what I can tell of the Bloomberg tracker, it's predictions are slightly lower than both your extrapolations.

I am assuming your calculated weekly average rate for Q3 so far, is around 4500 cars/week.
(83575-41040 =42525 cars in Q3)
Divided by 66 days
x 7 for weekly rate
= 4512 weekly average.

However, Bloomberg's forecasted production after this week, looks to be lower than that.
Something like 4150 per week.

So from what I can briefly estimate, Bloomberg's current forecast seems to be production of around 59,500 for Q3.

This may be, but it would be better for Bloomberg to state its forecast numerically rather than just graphically. But thanks for adding your interpretation.

 

[EV-lutioin]

That Bloomberg has a Model 3 tracker speaks volumes for the interest in Tesla. Do they have a tracker for any other car or car manufacturer?

 

[jhm]

The Bloomberg Model 3 Tracker is a case in point why Tesla should be releasing monthly sales numbers. If it is poorly calibrated, it would be better to recalibrate it on a monthly basis rather than quarterly.

This brings up an old controversy. Should Tesla publish monthly sales numbers?

 

[Orbit]

Bloomberg Model 3 tracker is now predicting 3857 per week for the next few weeks.
They seem to change and update their numbers quite a bit. You can't really rely on what they are suggesting at any particular time, it's just a guide.
Should still meet / exceed guidance for Q3 production.

 

[Webeevdrivers]

Down to 2266 this morning. Dropping a 100 or do a day now. We’ll be in the high teens within a few days and 1500 by next week I think the tracker is either completely relevant now or there are some serious line issues.

 

[Webeevdrivers]

Whoops. Forgot to spell check. Meant to say the tracker is now completely irrelelevant.

I’m kinda wondering why they even leave it up. Doesn’t look good on Bloomberg.

 

[Cloxxki]

There's an insane up tick in production being reported on Bloomberg. How much of it might be real and how much due to confusing practices by Tesla skewing the model?

Confusing: 49,000 Model 3 this quarter in 50-51 days, 960+ cars a day, 6,725+ per week.
Counter says 5,608 weekly now.
Graph shows peaks well beyond 8,000/wk.
If you look at the 13 weeks average of bars, you get higher than 5,68, eyeballing it.

If they ride the quarter out, that'll be a big fat record. And the annual product forecast may actually have been (let that settle in for a moment...) conservative.