It's difficult to make projections...

It's difficult to make projections...

It seems like this should be obvious, but based upon all the questioning of Tesla missing a production target of 1500 units at the beginning of an S-curve type ramp up makes me realize it isn't.

Some problems and delays are impossible to predict, but completely normal and expected in any new process. So, Elon used a well tested method, the S-curve, to predict production ramp up. Even if you can't predict specifically what will cause delays, you can anticipate their impacts on number of units produced to some extent. The amount of optimism/pessimism the estimator uses to draw the S-curve will be a guess at best. However, the bottom line is that a small number get produced at the beginning and momentum builds over time. (Again, I feel like I am stating the obvious, but the evidence says I'm not.)

So, if Tesla misses forecasts by 1200 units (yes I know this is 80% percent of the forecast and 400% of the units produced), it won't matter when they are at full production. When they are producing 20,000 units a month is it inconceivable that they might produce 22,000 in a month and wipe out any backlog from missing in September. The point is, the small numbers anticipated at the beginning make it easy for skeptics to claim the sky is falling when they are missed by a little bit.

Now, if delays continue and are drawn out for several months, there is a problem. But, you can't predict success or failure at this point based upon these tiny numbers. (See Chaos Theory in Wikipedia.)

Shock | 9 ottobre 2017

I believe they are already at least two months behind based on the tweets from musk about when customers can configure the cars.

Saying that the 1500 will be produced fast later and thus production is barely behind is incorrect. They are behind by the amount of time it takes to produce 1500 cars NOW, not 1500 cars LATER. The entire s-curve has shifted to the right where x = time.

hoffmannjames | 9 ottobre 2017

I completely agree which is why I think the delivery estimator should not be taken too seriously at this point. A lot can happen in the next year to either speed up production or slow it down.

RJMIII | 9 ottobre 2017


The fault in your thinking is that the rest of the S-curve is fixed. If they out produce the S-curve this month, they will be back on track and shift the S-curve back to the left (I know left shits are rare). The point of my post is that there is too much uncertainty at this point to make predictions of ultimate success.

And I fully admit the possibility exists that the curve could shift further to the right.

andy.connor.e | 9 ottobre 2017

Have you really ever been able to accurately predict anything when it comes to Tesla?

hoffmannjames | 9 ottobre 2017

According to the S curve on the website, production should reach its peak by the end of December. So by the end of this year, we should have a much better understanding of how production is going. Reaching that maximum, is the key, I think.

mntlvr23 | 9 ottobre 2017

the S-curve graph was just trying to convey a representation of the shape of the curve and not necessarily accurate as to when it will peak. Tesla has already stated that there will be a doubling in production rate continuing after Dec (5,000 units/week) through end of 2018 (projected 10,000/week)

hoffmannjames | 9 ottobre 2017

@mntlvr23 Granted but I think my point still stands that we will have a better grasp on the state of production around the end of this year, because when production is on the top side of the S curve, we will see what production is like when it is fully cranking.

mntlvr23 | 9 ottobre 2017


carlk | 9 ottobre 2017

Some said it the best. I take misses of Tesla over celebrated wins of other companies any day. Let's say the Model 3 is late for 3 months it's still years ahead of everyone else.

Also remember this is a brand new design not most "new" models other companies product which are 90%+ the same as previous ones. In the case of truly new design no other company had precisely predicted schedules either. Most of them don't even bother to give you a prediction.

ReD eXiLe ms us | 9 ottobre 2017

Naysayers predicted Tesla would miss production of Model 3 by as much as 99% for the year, and probably 100%. Both predictions are destined to be wrong. Being supposedly 80% behind at such an early stage forgets their previous anti-Tesla projections entirely. It also ignores that every bit of improvement through December 31, 2017 makes those detractors even more wrong. They can't say Tesla was late, they can't say they were right, they can't say they 'knew this was gonna happen', at least not honestly.

The instant Tesla delivered one Model 3 in 2017, every prognostication of 'zero' or 'none until 2019' was proven wrong.

The instant Tesla Delivered 30 Model 3, of 50 built, on July 28, 2017 the predictions of 'only a handful by December 31, 2017' were wrong.

The instant that more than 1,000 units are Delivered, guesses of 99% missed on a 100,000 unit goal will be wrong. The instant that more than 2,000 units are Delivered, impressions of a 99% miss on a 200,000 unit goal will be wrong.

Every unit sold will improve the ratio throughout this year and every year that follows. That is my prediction. So therr.

Frank99 | 9 ottobre 2017

You're such an over-the-top optimist, ReD. Glad to have you on board.

ReD eXiLe ms us | 9 ottobre 2017

Frank99: Thank you. I think it is all about context. No matter how Naysayers attempt to spin a yarn, it remains flammable in the presence of facts. It only requires a slight change in perspective to spy the errors in their methodology. Facts conquer lies. Logic banishes rhetoric. Giving any ground, yielding to any of their points, accepting their position as valid -- any of that guarantees loss on the field of battle. Even if they somehow believe their cause is just (I doubt it) they cannot be allowed to be 'right' when I know damned well I'm not wrong. So I work not only to prove my case correct, but to demonstrate why theirs is not. Red herrings, non sequitur, terminal idiocy, and rampant rhetoric be damned.

PhillyGal | 9 ottobre 2017

@DrRob - Agreed but everyone isn't good at math, hence all the freaking out.

Xerogas | 9 ottobre 2017

Maybe Tesla could publish an S-curve with error bars on it. Then not as many people would point to precise dates and numbers on it.

bmz | 9 ottobre 2017

What the S curve really shows is that the original scheduled production for the months of October and November are less than one week's worth of production in, say, May 2018.

Shock | 9 ottobre 2017

"What the S curve really shows is that the original scheduled production for the months of October and November are less than one week's worth of production in, say, May 2018."

But this means as much as me saying that one week of production vehicles in, say, May 2018 will take them several months at current rates. If Tesla's ramp up prediction remains accurate, and what they thought they would have produced by now ends up not being produced until December, they are two months behind.

On Friday in regard to customer deliveries must said "Bear in mind that the production ramp is an exponential curve. Late Oct is possible, but not certain. Literally every day makes a big difference.".

This can't possibly be taken seriously.

El Mirio | 9 ottobre 2017

All the disclaimers you are looking for Tesla stated in their SEC Q-10 filling in August. Investors have been made aware well in advance that there will be significant ongoing risk for production delay, Q-10 is filled with warnings.

ReD eXiLe ms us | 9 ottobre 2017

Shock: As El Mirio points out, Tesla covers themselves fully in EVERY 10Q Report to the SEC, and those are publicly available on the internet. Just look through the 'RISK FACTORS' section and everything that anyone could consider a potential problem, from natural disaster to competitors' products selling better is listed. If Tesla moves to countersue, the case will be dropped immediately.

About your, "If Tesla's ramp up prediction remains accurate..." comment...? You are looking at this both upside down and backward. Because you presume that for the next couple of months they will still be looking for the cause of the 'bottleneck' and that they will neither develop nor implement a solution to it. That is why claiming a target or goal is "possible, but not certain" is absolutely serious.

CYA. Words mean things and stuff. The phrase, "Past performance is not indicative of future blah-blah, blah-blah, blah..." goes both ways, positive and negative.

karlmitchell | 9 ottobre 2017

There's also the issue that the reduction in the federal tax credit is based on which quarter Tesla sells their 200,000th car. It wouldn't surprise me in the least if that just happened to be in the first week of a quarter, and the delays in the S curve allow that to align perfectly. Hitting it just right could save new Tesla owners up to hundreds of millions of dollars in total.

ReD eXiLe ms us | 9 ottobre 2017

karlmitchell: +21! Precisely. Elon Musk stated long ago that Tesla would "do the obvious thing' in regard to maximizing eligibility for the Federal EV Tax Credit in the U.S.

SUN 2 DRV | 10 ottobre 2017

Dr Rob: I fully agree with your premise, but I'd tweak it a bit...

"Although it's very easy to make projections, it's very hard to make ACCURATE projections, especially when "pushing the envelope" on things that have never been done before."

Tesla never makes promises, they always make stretch goals and that why they can accomplish so much more than anyone else. 80% success on a Tesla stretch goal (Model 3) is much better than 100% success on a GM goal (Bolt)

ReD eXiLe ms us | 10 ottobre 2017

SUN 2 DRV: Very well put. I wonder, what percentage of effort was put into BOLT compared to ATS, LaCrosse, and MALIBU? A 238 Mile range fully electric LaCrosse for $37,000 dubbed 'ELECTRA' would certainly be more impressive than the BOLT.

Rutrow | 10 ottobre 2017

"It's tough to make predictions, especially about the future." - Yogi Berra*

*or Gilbert Godfrey?!? ¯\_(ツ)_/¯

Haggy | 16 ottobre 2017

"They are behind by the amount of time it takes to produce 1500 cars NOW, not 1500 cars LATER. The entire s-curve has shifted to the right where x = time."

Not at all. The number of cars made while there's a bottleneck has nothing to do with when the bottleneck will go away. It's not as if having made 1000 more cars would mean that the bottleneck would go away a few days sooner. The reality is that if the bottleneck gets cleared up a couple of days before Tesla anticipated, it would mean they'd be ahead of where they'd be a couple of months from now, compared to where they'd be if they had made an extra 1500 cars last quarter but it took them two days longer to clear up the bottleneck.

People are still not putting this in perspective. I think this would be easier if people drew a picture. Get a piece of paper and a ruler. Draw a line 10 inches long. Label one end 0 and the other end 60,000. 1/4 inch from the beginning, make a mark and label it 1500. Make another mark about 1/16 inch from the beginning. It doesn't have to be exact since the distances are so small. Notice how both of those marks are right near the beginning and close to each other.

This line represents a capacity of 5000 cars per week over a quarter. The marks at the beginning represent how far Tesla got and how far they estimated they would get. Notice that they are almost in the same place relative to the line and that they both represent hardly moving from the start at this point. The next time somebody says that Tesla fell behind by 85%, look at that line so you can put things in perspective. If you consider only the first 1/4 inch of the line, it's easy to take things out of perspective. But if you consider how far they expected to get vs how far they got, you can see that it didn't make a significant difference.

It's like looking at a faucet with a flow restrictor that cuts off 99% of the flow and counting how many drops came out instead of asking when the restrictor is coming off. The latter question is the only relevant one.

KP in NPT | 16 ottobre 2017

If the bottlenecks are due to suppliers, then Tesla isn't totally in control of the delay.

WantMY | 16 ottobre 2017

"Rusch places Tesla’s total Model 3 deliveries at 3,005 units this year" - I would be more precise with projected ?,005 units this year, LOL(,005). There is some mixed signals, T says no suppliers issues involved in delays, but then they blamed and fired a supplier. Anyway, whatever it is - ramp it up please, please, please.