The batching of cars and then putting them on hold till production for European cars starts would be going on til approximately January 15 (first two weeks of the quarter dedicated to production of cars slated to Asia/Pacific region). After the 15th and up to the end of the period on Jan 30, in "full maximize Quarter deliveries batching mode" the production would 100% switch to European cars, so your approach on "draw down" of the wait time would not quite work here, but it is exactly as it is shown on your chart.
Not exactly sure where you get the significance of Jan 15th. But let's say that's the period when Tesla started to work through their European backlog which was large at that time. All orders during that period of backlog work would go to the end of the batch (near the 30th let's say). So those ordered at the 15th would wait a two weeks before going into production those near the 30th would wait a few days. Draw down works just nicely.
This approach does not represent what is actually happens when Tesla is NOT operating in "full maximize Quarter deliveries batching mode" , either by running uniform production split in first, second and third month of the quarter, as they did in Q3 of 2015, or some combination of the two as they did in Q2 of 2015
Actually European orders are always delivered in batches whatever part of a cycle Tesla is in. This is because shipping creates another hot point, there are only relatively few ships Tesla considers for transport so cars have to wait at Fremont before they are punt on a train together. Again, if you follow the European delivery threads this is more than obvious since people are tracking exactly on which ship their car is.
Then there is example with the fictitious 6 weeks. First I pointed out that it is not possible because just transit/European Assembly time to Europe is 6 weeks, and, in fact you were the one who commented that this transit/assembly time is currently 8 weeks based on the data point obtained in one of the TMC threads on European delivery.
If it's not possible then how come it actually happened? Now I am not a native speaker but even I am pretty sure fictitious means roughly 'imaginary' or 'didn't happen in reality'. Unless you are accusing the fellow (well respected in the Dutch community) forum member of lying I would look for a different adjective.
Then you came up with another data point, that indicated outlier case that had compressed shipping/assembly time. When I pointed this contradiction, you indicated that there is no contradiction because your comment on increased transit/assembly time was regarding the average time. Well, fair enough. So if you recognize that outlier case is not the same as average, how can you insist that fictitious 6 weeks wait time is ok based on the outlier case, while the way it shown on the graph makes it applicable to ALL orders that are received at that point in time?
I assumed it was obvious I meant 'average over time'. Cars that get produced at the same time generally end up on the same transport so there is little to no variation in shipping time. So it makes no sense to talk about averages in shipping time at that point in time, which leaves the only possibility (logically) which is average over time.
I think that this illustrates well that my initial assessment about 6 weeks wait time for European delivery shown on the graph is fictitious indeed, and I believe that it misrepresents the actual data.
I show you actual deliveries happen in time frames consistent with my 'draw down' hypothesis yet you keep refusing to acknowledge their existence even claiming it's impossible. I am not sure how I can convince you anyway else.
There was also the point that I made about your approach resulting in accounting for zero cars delivered in the second half of March-First half of April.
See #891 where I explicitly state such a thing is indeed not plausible and explain how your conclusion is not the only possibility.
In addition, I do not have your data series, and actually do not have a detailed description of your method either, which would be helpful to make sure that my understanding of it is correct. I've analyzed couple of other periods from your chart for Europe, and was not able to justify/explain the graph based on my understanding of your method as I was able to deduce it from the various posts on the subject.
It's pretty simple. At each date of a change of waiting time (when you posted a table basically) I filled in the week number for the current date and the week number for the expected delivery (second week of month for whole months delivery, last (full) week for end of month estimates) Subtract them and there you have it. Do that for each of the three regions at that point in time. If there was a difference in between one region or between models I took an average. As I said earlier, My data suffers from what I see as your problem in the data as well since I really should sample every two weeks. But generally there is an update every two weeks so I let that slide. But even in my graph some dips are erased.
(If possible could you insert tables directly next time, I can't quote them right now, I briefly recap them hopefully it is clear what refers to what)
First of all, you stubbornly refuse to acknowledge that it's just not possible to use sub week (or even sub two week) resolution since the input data doesn't provide it. Really, that's just not my opinion, it is truly statistically established fact. Can we agree you don't mention sub week precisions anymore? Secondly, for all your no-drawdown questions, see my answer above. That said to the remaining questions :
4/9 : 10 instead of 11 weeks Switzerland still showed June at that time, important country being two weeks earlier brings the end period from week 15 (end of june) to week 14 (3rd week)
2/6 : 16 versus 17 weeks 70s were already projected for June so averaging with Late May for the 85 gives first week of June as targeted average delivery
2/4 : Same thing but now with May for the 85 bringing the average even more to the front
1/30 : Same thing but now with Late April/June
26/11 : 15.7 to 15 because the March starts on a Tuesday so I considered second week to be week 10.
There you have it. Now maybe you can explain how you accounted for variations in delivery waiting times across models and countries in Europe at a given date? It looks like due to you not understanding how I came to my numbers you didn't take them into consideration at all which makes me question the accuracy of your charts once again (on top of the precision issue which you have yet to argue against or acknowledge) Or update your graphs and trendline to also use the history you conveniently left out (you know the part where Tesla worked itself out of really being production constrained after the Dual/AP announcement with port strikes double whammy)
