Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
In 4 out of the last 5 quarters, my final Tesla delivery estimate was off by 2.6% or less. In Q3, I will post my final estimate on Twitter here on the last day of the quarter as usual. I'm again aiming for less than 3% error. Here is what I think of error rates:

  • 0-2.0% Excellent
  • 2.1-3.0% Very good
  • 3.1-4.0% Good
  • 4.1-5.0% Average
I'm using DMV data and VIN data for US deliveries. Getting the Model S number wrong could cause more than 2.6% overall error on its own. I'm not aware of a better source than DMV data for Model S deliveries in Q3. I'm purchasing some of the DMV data so I can have data for September before September is over,

Instead of publishing all content only on Patreon, I post the results on Twitter too with a 7-day lag, except for the final estimate on 30 Sep. That one won't have a lag.
"I think US Q3 deliveries will be lower than expected because some customers will not take delivery and will chose to wait for the tax credits."

"I think Tesla will produce more cars than deliveries because it wants to refresh low inventory in the US".

Sorry Troy, but you are making too much assumptions each quarter and that is not being data driven.

You estimated 37K deliveries for August in China and the actual was 44K. That is being utterly wrong (more than 15%). That is having no idea what is going on in China.

For the last several quarters, the ones that make estimates based on Tesla's production are way closer, much closer than you.
 
In 4 out of the last 5 quarters, my final Tesla delivery estimate was off by 2.6% or less. In Q3, I will post my final estimate on Twitter here on the last day of the quarter as usual. I'm again aiming for less than 3% error. Here is what I think of error rates:

  • 0-2.0% Excellent
  • 2.1-3.0% Very good
  • 3.1-4.0% Good
  • 4.1-5.0% Average
I'm using DMV data and VIN data for US deliveries. Getting the Model S number wrong could cause more than 2.6% overall error on its own. I'm not aware of a better source than DMV data for Model S deliveries in Q3. I'm purchasing some of the DMV data so I can have data for September before September is over,

Instead of publishing all content only on Patreon, I post the results on Twitter too with a 7-day lag, except for the final estimate on 30 Sep. That one won't have a lag.
Sure but what is your accuracy at the beginning of the quarter, after 1 month and after 2? Having a value on Sep 30 is hardly predictive. I can predict the Oct 1 stock price on Sep 30 with Very good accuracy (2.1 - 3.0%) most of the time, but so what. It's earlier predictions that actually matter.
 
Sure but what is your accuracy at the beginning of the quarter, after 1 month and after 2? Having a value on Sep 30 is hardly predictive. I can predict the Oct 1 stock price on Sep 30 with Very good accuracy (2.1 - 3.0%) most of the time, but so what. It's earlier predictions that actually matter.

Yes, I agree. It's important to try to estimate the quarter accurately in advance. That's why I measure the accuracy of my first estimate as well as the last one. I'm very transparent about this. All of these details including links to each estimate on Twitter can be found here.

The first table shows my first estimate early in the quarter. The second table shows my final estimate on the last day of the quarter. Yes, I know I messed up badly in Q1 2021.

QZX0l4M.png
 
You got to love Jay Leno and his support for change to EV's.


I got to to give him credit for stating this in such a common sense manner.

“I mean the advantage, if there is one, to an EV fire is, it doesn’t blow up,” Leno explained. “You’re in it, you smell something, there’s smoke, and then it doesn’t go up in a ball the way a gasoline car would. That’s not to say it’s not dangerous and, hopefully, they’ll fix the problem.”
 
Yes, I agree. It's important to try to estimate the quarter accurately in advance. That's why I measure the accuracy of my first estimate as well as the last one. I'm very transparent about this. All of these details including links to each estimate on Twitter can be found here.

The first table shows my first estimate early in the quarter. The second table shows my final estimate on the last day of the quarter. Yes, I know I messed up badly in Q1 2021.

QZX0l4M.png
Respectfully Troy
Why are you doing this?
What is the purpose of these estimates?
How is this helping the investment thesis ?

to the TMC members,
How are these quarterly estimates helping our investment decisions
To date , I have not made my investment decision based on Troy’s estimates
These are just estimates and more often than not have been off .

the way I see is, Troy is conservative in his estimates and recently his margin of error has been ever so slightly creeping up

Infact , Would it be wrong to assume that the shortzes are using these lower estimates to drive down the share price a few days prior to the actual delivery report and then cap it .

please do correct me if I am wrong .
 
No, I don't own any TSLA stock, I don't normally trade TSLA and have not traded it for a long time.

I'd be curious to know the reasons WHY you don't own any TSLA, especially seeing how you spend so much time analyzing the stock? The moment I started making my own TSLA model and forecasting it's future is when I really started buying TSLA as quickly as possible, the math just seemed so appealing to me from an investment point of view.
 
"I think US Q3 deliveries will be lower than expected because some customers will not take delivery and will chose to wait for the tax credits."

"I think Tesla will produce more cars than deliveries because it wants to refresh low inventory in the US".

Sorry Troy, but you are making too much assumptions each quarter and that is not being data driven.

You estimated 37K deliveries for August in China and the actual was 44K. That is being utterly wrong (more than 15%). That is having no idea what is going on in China.

For the last several quarters, the ones that make estimates based on Tesla's production are way closer, much closer than you.

I have not said the things you quoted me saying. I don't think Q3 deliveries will be affected by the EV rebate. If somebody cancels their delivery, there will be another buyer in line who will take delivery instead. So, it's not an issue at all in Q3. As for Q4, if Tesla doesn't do anything then the number of buyers that are willing to take delivery at an $8,000 disadvantage will be clearly less than the 80,000 Model 3/Y buyers Tesla would normally have in the US in Q4.

As for the China numbers in August, Tesla decided to send two more ships to Europe at the end of August even though the ships will be late for Q3 deliveries. My estimates are based on what's most likely to happen. I had calculated the trip times and I wasn't expecting more ships. But there were 2 more ships and the exports were higher than I expected. Again I was very transparent about it here. There will be a delivery spill over into Q4 because of that.

If you want to beat my estimates, you don't need to make any calculations based on production. Just estimate 2% off in either direction and you will be more accurate than me 50% of the time.
 
Troy has a patreon page and his estimates give his patrons early access. So he doesn't need to buy/sell stocks to profit from his estimates. He is providing people a service and there is an audience. There are paywalled services all over the internet on analysis of companies or complete sectors.

So either you care for his estimates or not, he has paying customers for this data and it's his job to be as accurate as that's what his patrons pay for. I appreciate his work, some may seem it trivial or unimportant. But you find analysis from paywalled wall street journal articles unimportant as well so..
 
In 4 out of the last 5 quarters, my final Tesla delivery estimate was off by 2.6% or less. In Q3, I will post my final estimate on Twitter here on the last day of the quarter as usual. I'm again aiming for less than 3% error. Here is what I think of error rates:

  • 0-2.0% Excellent
  • 2.1-3.0% Very good
  • 3.1-4.0% Good
  • 4.1-5.0% Average
I'm using DMV data and VIN data for US deliveries. Getting the Model S number wrong could cause more than 2.6% overall error on its own. I'm not aware of a better source than DMV data for Model S deliveries in Q3. I'm purchasing some of the DMV data so I can have data for September before September is over,

Instead of publishing all content only on Patreon, I post the results on Twitter too with a 7-day lag, except for the final estimate on 30 Sep. That one won't have a lag.
Two issues with that:

1) You don’t actually get to decide the grading system for yourself; that’s decided by your peers

2) Being off 2% in 2016, or 2017 or even just last year is not the same as being off 2% now or moving forward. 2%ish used to be mouse nuts on the financial sheet; not so much anymore
 
I appreciate what @Troy has done over the years - his estimates were the only ones out there for a long time and used to be quite accurate, these days it's a bit more hit-and-miss, but it's still one data set to take into account. We have others now coming with some numbers, including Rob Mauer, our very own @The Accountant, and of course the most accurate of all from @Krugerrand

So take a chill pill folks, you can disregard the numbers if you don't think their correct
 
Two issues with that:

1) You don’t actually get to decide the grading system for yourself; that’s decided by your peers

2) Being off 2% in 2016, or 2017 or even just last year is not the same as being off 2% now or moving forward. 2%ish used to be mouse nuts on the financial sheet; not so much anymore
I think being 2% off is still 2% off on the balance sheet. Don't know how being 2% off today is more than 2% off on the balance sheet. Infact with more energy revenue coming in it should mount to less than 2% off.

Also his paying customers get to decide what error rate is acceptable by voting with their wallet.
 
2) Being off 2% in 2016, or 2017 or even just last year is not the same as being off 2% now or moving forward. 2%ish used to be mouse nuts on the financial sheet; not so much anymore

Analyst consensus for Q3 2021 is currently 221K. It consists of estimates by 11 analysts that range between 204K and 246K. It's a wide range. So, it's not like Q3 deliveries are crystal clear and it's easy to estimate it with a small error.
 
I think being 2% off is still 2% off on the balance sheet. Don't know how being 2% off today is more than 2% off on the balance sheet. Infact with more energy revenue coming in it should mount to less than 2% off.
The issue is bottom line profitably. Backing out fixed cost from gross margin, the last cars produced are the ones that make the profit. For example, 2% of cars could mean the difference between break even and a profit (even though revenue changes 2%). Or, if 2% of cars were the profit makers, 2% more would double profits (roughly).
 
The issue is bottom line profitably. Backing out fixed cost from gross margin, the last cars produced are the ones that make the profit. For example, 2% of cars could mean the difference between break even and a profit (even though revenue changes 2%). Or, if 2% of cars were the profit makers, 2% more would double profits (roughly).
Believe that's Teslas problem and not Troy's. Tesla wants to have a razer thin margin on profitability that's their choice. Their operating margins are more than enough to switch it up if they wanted to.
 
Talking to a new owner of a model 3, I guess Vancouver is in full delivery mode with dozens if not 100’s of new Tesla’s leaving the delivery centres everyday.

Interesting to note on auto trader, we are down to

- 3 new Leafs available in the province,
- no Bolts as there is a stop sale order
- 11 Konas
- 143 Mustang Mach E’s.

The Mach E number surprised me. I thought they were all pretty much pre-sold.

cheers.
 
I think being 2% off is still 2% off on the balance sheet. Don't know how being 2% off today is more than 2% off on the balance sheet. Infact with more energy revenue coming in it should mount to less than 2% off.

Also his paying customers get to decide what error rate is acceptable by voting with their wallet.
Right?! 1 million in revenue is the same as 10 million in revenue all day long, every day. 🙄

And that’s what I said - his PEERS (whether paying customers or not) get to decide the grading system. But if you go back to his post, HE came up with his own grading system. 🙄
 
Right?! 1 million in revenue is the same as 10 million in revenue all day long, every day. 🙄

And that’s what I said - his PEERS (whether paying customers or not) get to decide the grading system. But if you go back to his post, HE came up with his own grading system. 🙄
Telsas total revenue also went up so the percentage doesn't change. When Tesla makes a trillion dollar revenue in the future, a 10 billion dollar revenue miss from estimates is immaterial at that time.
 
Analyst consensus for Q3 2021 is currently 221K. It consists of estimates by 11 analysts that range between 204K and 246K. It's a wide range. So, it's not like Q3 deliveries are crystal clear and it's easy to estimate it with a small error.
So? This isn’t about those 11 analysts. This is about you and your analysis. You’re here defending your analysis, they aren’t.

To be clear, I don’t care what you estimate anymore than the guy down the street. But I’m still going to point out the flaws in your approach because there are people here who take you at your word and estimates, and invest/trade according to your data. You have a level of responsibility, as do they.