-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Fair question. We originally wanted to ask for the payment confirmation time, unfortunately some credit card companies don't store this information and we didn't think Tesla would appreciate a bunch of people calling in asking for the time stamp on their confirmation. The other big issue is that who really wants to call their credit card company? We decided to use the email confirmation time, and then made an assumption that the email's were sent out sequentially. Our algorithm can then work backwards and provide the estimate. It will be getting better with time - as we get more information some of our assumptions can be quantified and become statistically more accurate and reliable.
ummgood
Member
I tried your estimator and I think it is WAY off. If it isn't I would be majorly upset. I was number 106 in line in Austin, TX and submitted my deposit at 11:15a.m. Your calculator is saying I am number 296183. There wasn't even 150k sign ups by the night of the reveal. I am personally hoping I am about 10k in line in the US. This is based on ~100 (people in line before me in Austin) * 100 (stores in the US) = 10k. I know this isn't a precise number but at least it gives me a feel for where I am. Yes I know that the preowned cars will be in front of me but I am hoping I'll still be in the 50k range if everyone who owned a tesla bought a model 3.
For reference my reservation email came:
Fri, Apr 1, 2016 at 4:58 PM
Last edited:
We're working on refining this, will hopefully be getting a better algorithm up soon. I know its off by quite a bit in some cases, but its just an estimate. There's quite a few Tesla employees and current owners (plus people who are currently on a Model X or S waiting list), this fact really pushes everyone back in the line quite a bit. We did a fair amount of research to come up with our estimates - but with some quick Googling you can find that there's at least 115,000 current Tesla owners and employees (not including people currently on a waiting lists)
ttupper92618
Member
I have to question the very purpose of your site. In simple terms, I cannot imagine any way in which you could determine the delivery timing of these cars based on any of the data points you are collecting, because Tesla simply hasn't revealed what factors they will more heavily weight, and the actual production start and ramp time for the car is unknowable. However, this data can be used to fish.
I think it's really interesting that in your code, you refer to the email date and time as a credit card authorization date and time. I'd like to hear your explanation for how you use this data, and why you chose those specific terms for these values. Why? Because to me, the data you are collecting do not look like they could present a very useful mechanism for predicting something as nebulous as when Tesla will manufacture these cars. But these data points look like a great way to fish for data points that might allow the connection of, perhaps, something like credit card transactional logs with data which might identify real people.
If there is any sense to you, you will spend some time explaining precisely how your 'algorithm' works, and why you need these specific data points. And until you do that, I'm certainly not going to give you a shred of data.
That's just my opinion.
Being a septuagenarian, I figured the only way I can be sure of getting a model 3 during my lifetime was to make my reservation as early as possible on March 31, and so I was around #60 at the Short Hills, NJ store. I figured if production is really delayed, I can offer my son my reservation number. I found out recently that he also left a deposit.
ummgood
Member
Even if I was behind the 115k current owners add say 10k to that and I should be around 125k. I personally still would be unhappy being number 125k in line. I hope either not all owners/employees order OR tesla comes to the realization that they need to reward the people who waited in line so long. What would be nice is if they prioritized current orders based on when in line the person was AND/OR if they are a current owner. Say allocate 25% of the cars to non-owners so those of us that waited in line for hours before the store opened can get some of the first cars. Instead of waiting for queues of owners to get ALL of their cars before starting to fulfill the first non-owner.
Basically I need another car by the time my daughter is 16. That is end of 2018. If I don't have the ability to configure my car by say 4th quarter 2018 I will request my money back and go buy an ICE. I waited in line to make sure I had a fair shot of getting the Model 3 when I needed it. If I can't then I'll get something else and in 7 or 8 years see what Tesla has to offer when my existing car wears out.
Very fair points. And we appreciate skepticism and critique - if our model/algorithm is going to become better its going to have to go through a few different versions.
To address your point about the terminology in the code: we were originally going to use the credit card authorization date and time, but then (after we launched the site) we learned that some credit card companies don't keep the time stamp so we had to switch on the fly pretty quickly. Thanks for pointing this out, my friend is the website programer/coder so I'll bring it up to him. We are going to change this but only for consistency not because we are secretly fishing for something. Honestly I can't stress this enough - the purpose of this site is to build a tool that can estimate your position in line and provide an estimate of when you can expect delivery. It's a pretty unique opportunity that presents a very cool challenge to overcome.
We are justifying using the email confirmation time because we've made the assumption that the Tesla servers behaved rationally during the first few days of the release and queued up the emails so that they were released in sequential order (first in first out). For now this means there's a delay factor that we have not incorporated into our estimates, but we are working to incorporate it (we also all have jobs so time is limited, but believe me this if nothing else this project is becoming very effective at filling our spare time, haha, in the best way possible though).
You're first point about the purpose, and no disrespect, but we would like to believe your wrong. I spend most of my days working with numbers and models and algorithms that try and predict what's going to happen based on a few inputs. I really enjoy it sometimes (other times it gets quite frustrating), but the point is that there is a fair amount of data publicly available, and by combing through press releases, interviews, analyst reports, financial statements, etc. we've been able to bring the relevant parts together that allow us to make an educated data driven estimation based on your few inputs.
I hope this answers a few of your questions? Please feel free to ask more or provide recommendations. Like I said there will be a few versions of the algorithm, and each time we update we will put notes to explain what we've done. You'll also be able to see the history of your estimates so you can compare and see how our refinements affect you.
When I saw the project, and saw that you were in Toronto, I figured you did something like this.
I don't think that basing this on the confirmation e-mail date is going to work. E-mail delivery can be unpredictable sometimes, using various queues in different servers and spam filters. In my case it looks like Tesla sent the confirmations out as a bulk mail delivery the next day (4/1) by using messagelabs.com as the bulk e-mail service (look at your e-mail headers). When I log on to mytesla, it correctly says that I reserved it on 3/31. I would say that the date and time when the order was actually placed would be much more accurate, and most people should remember the actual time, more or less. Besides, that's exactly what the confirmation e-mail actually says:
Effective Date - Your Reservation becomes effective when (1) you place your Reservation and (2) we receive your Reservation Payment.
Last edited:
I am missing something...When I saw the project, and saw that you were in Toronto, I figured you did something like this.
Canada, and specifically Toronto, has a very strong machine learning contingent. The University of Toronto has some graduates and staff who are among the stars in the field. @BQuinn says he's doing predictive analysis using limited data; it seems to fit.
Agreed. I reserved in store on 3/31, but this says that my position is almost at 300k. I really don't think this is very accurate.
I'm fully expecting to get my model 3 in late 2018 or 2019. Even though I reserved late, I'm on the West Coast (just a few hours from the factory), and I'm not buying a base model (I plan on adding 5 - 10k of options). So that will bump me up by perhaps 100 - 200k in line if most people are just looking to get the 35k model and roughly half of the current preorders are not in the West Coast.
Dan Detweiler
Active Member
My guess is that most people will actually bump up their option list, if for no other reason than to try to get their cars earlier. I will probably get a lot of options anyway (bigger battery, AP, leather, paint, premium interior, cold weather package, pano roof). Hopefully for under $60,000. Maybe that will bump me up a year or so (I just reserved a week ago).
Dan
Dan
This may well the case ... or maybe not: Tesla has not indicated the full layering model to the fulfillment plan and Elon has indicated that the "West Coast First" rollout is only for early production vehicles, not their entire production capacity. Without understanding how the various factors interweave it's pretty much impossible for anyone to predict with any accuracy how the queuing will work: For example: Will an Order placed in Florida Today to an Existing Tesla Owner skip ahead of an order placed last March 29th to a line-standing non-previous owner? Will a Tesla employee who won't be hired until 2017 have his order skip ahead of the entire queue? Unless Tesla eventually documents the queuing (and I would not if I were in their shoes), it's pretty much total speculation where everyone thinks they are in the line.
Here is what we do know: the push toward both EV's and full autonomy are quite strong, many players are starting to become more serious in this market segment. Tesla has a limited opportunity to grow their market-segment beachhead before competition will make this more challenging. Personally I'd rather see the Model3 delayed an extra year than to have it released quicker only to then learn that it has service issues comparable to existing Model S or X's. It's more important that Tesla get it right than to get it out fast at this point, but doing both would be great!
Similar threads
