twohoos

Independent Estimate of USDM S2000 Monthly Production (AKA: Mapping VINs to build dates)

This post is in four parts – apologies in advance for the length. The seed of this project was my own casual curiosity about the S2000's production lines in Honda's Tochigi and Suzuka factories. These facilities were used only for specialty models –NSX, S2000, first-gen Insight, etc.– and total capacity was reportedly limited to about 100 cars/day. How might that limit S2000 production? Did production of a model year always start in the same calendar month? Were there any gaps, i.e. months where no US cars were made? For that matter, what were the very first and last months of US production?

I’ll address methodology, sources, and caveats in the posts below, but for now, let's get to the numbers.

USDM S2000 Monthly Production, Part 1: New Estimates and Overall Observations
Below is my current (see end of Part 2 for date of last update) best estimate of monthly production numbers for the U.S. market (USDM) S2000. Each model year is plotted with a unique color, and the total monthly production is also indicated with the black asterisks. (The total for a month differs from the model year estimates whenever there was overlap -- usually due to model-year changeovers that occurred mid-month.) Solid lines indicate production at the Tochigi plant, while dashed lines indicate Suzuka production. The changeover from Tochigi to Suzuka occurred at the end of April 2004, at which point Honda re-started the MY04 VIN serials, using an "S" prefix instead of "T" – therefore I've broken the MY04 data into two traces (MY04T and MY04S).



The corresponding numerical estimates are in the table below. The calendar months in the left column correspond to the calendar years indicated in the top row (e.g., MY00 production spanned from June '99 through August '00; MY01 production spanned August '00 through August '01; and so on).

**See Note below regarding "estimated" vs. "actual" totals.

Table entries that are shaded yellow indicate uncertainty as to whether ANY production of that model year occurred in that month (because I have not actually verified any existing VIN sticker for that model year and calendar month). Specifically, it is possible that:
- MY01 production may have extended into Sept '01, or MY02 production may have begun in August '01;
- There may have been zero U.S. S2000s produced in May '04, following the changeover from Tochigi to Suzuka [edit: I've now confirmed that Suzuka production DID begin in May '04];
- [Edit: I've now confirmed that the first 4 production AP2 cars (MY04) were made in July '03, while MY03 production was still continuing. After MY03 production ended in September, MY04 production resumed. Thus there were no MY04 cars made in Aug.'03. See Part 4.]
- MY04 production may have extended into August '04, or MY05 production may have begun in July '04;
- MY00 production may have begun in May '99 (this seems unlikely);
- UPDATE: Thanks to @migz_s2k we now know that MY09 production ended in Jan '09 (or *possibly* Feb '09)! See Part 3 for discussion of why I believe that this followed a month-long hiatus (in Dec '08) during which no U.S. S2000s were built.

Some immediate observations of overall trends and unusual events.
- The duration of production varied significantly from model year to model year. That is, we'd expect a 'regular' model year to have a 12-month production run, which in general would span 13 calendar months (e.g. mid-August to mid-August). However, even ignoring MY00's early start and MY09's early end, we see production vary from 11 months (MY06) to 14 months (MY05 & MY07).
- Correspondingly, we see the start or end of a model year's production occurred as early as July, and as late as September (again ignoring the MY00 and MY09 outliers).
- It appears that when there was a model refresh ('02, '04, '06, '08), production didn't start until September (late August for '08); but when there were no year-on-year changes, production began in August or even July. (MY03 is an exception here -- possibly due to the introduction of the all-new tan interior.) Perhaps --and this is speculation-- Honda has found that U.S. buyers are willing to wait a bit for a "new" (refreshed) model, but when there are no substantive changes, incrementing the model year early acts as an enticement to buyers.
- Monthly production peaked in February and March of 2001, with ~2,500 made in those two months alone – that's 3.7% of U.S. S2000s, all made in a 59-day span! (And again, that's just U.S. cars – remember, JDM and Euro sales were very hot then too, so the factory must have been buzzing!)
- There was a nearly month-long hiatus in U.S. production during the changeover from Tochigi to Suzuka in May '04. It’s perhaps surprising, then, that Tochigi’s run ended with a burst of productivity: April '04 saw the highest output in 18 months, posting a mark that was (to within the error of my estimates) equaled only once--in 9/'04--by any single month of Suzuka production.
- U.S. production typically peaked between February and May for a given model year, although it varied. This makes sense given that the car’s a sporty convertible: by accelerating production in late winter and early spring, Honda would be sure to have plenty of cars in dealerships when U.S. buyers were eager to enjoy the summer weather.
- The MY01, MY03, and MY05 production runs tailed off significantly in their final few months, perhaps indicating that resources had been diverted to development of the "refreshed" MY02, MY04, and MY06 models, respectively.
- Despite the previous observation, no production dropoff is evident for MY07 in advance of the CR/Type S introduction. However by that time, overall demand was down significantly and the production line would not have been highly stressed.
- Overall, production of U.S. cars spanned 116 calendar months (June '99 thru Jan '09). JDM production, by comparison, spanned 126 months: March '99 through August '09. That's nearly an entire extra year of JDM production!


Continued in Part 2 below...

---------
**Note: Counting from 0001 to the highest valid VIN of each model year, there are 67,060 total USDM S2000 VINs. However, total U.S. production is known (thanks to @patinum) to be 66,860: it turns out there are exactly 200 "invalid" VINs across all model years. These VINs correspond to chassis which were never delivered or offered for sale; nearly half of them appear to be early-production prototypes/development mules: 42 of them are early '00 VINs (the first-ever U.S. cars), 25 more are early '04 VINs (the first-ever AP2s), and another 25 are early '06 VINs (the first drive-by-wire cars). The rest were presumably isolated production units that failed quality checks or were subject to destructive testing. My estimates here only account for 100 of the 200 known-invalid VINs; therefore my production numbers sum to 66,960 (100 more than the total "known valid" production).

twohoos

USDM S2000 Monthly Production, Part 2: Methodology (Or, Mind The Gaps)

For this project I wanted to use primary sources (i.e. not Wikipedia, personal blogs, etc.). The ideal source, of course, would be Honda; but Honda does not typically publish monthly production statistics on its public sites – and in the rare cases one finds, the data is not broken out by model. Similarly, U.S. government sources tend to focus on either sales (which I already know) or U.S. production (n/a for the S2000). Alternatively, auto-industry trade publications (Ward’s, Edmunds, et al) routinely publish both sales and production statistics, and these sources can be both reliable and detailed, since they have longstanding relationships with the major manufacturers. Indeed, over a decade ago, I had real success finding and tracking monthly S2000 sales (the annual sales statistics you'll find on the Wikipedia S2000 page are mine ). Unfortunately, at that time I wasn't interested in production numbers, and didn’t save most of the ones that I ran across. That was a missed opportunity: today, such industry statistics are A) generally only made available to paid subscribers (typically at the manufacturers’ request, to prevent wholesale leakage of potentially sensitive data), and B) probably no longer available anyway for the S2000, due to the car’s age. Bottom line: it seems that “official” monthly S2000 production numbers, if they were ever available, have long since disappeared. A different approach was needed – and I found it, stuck on the S2000 itself.

The federal ID sticker on the driver's door jamb of all US-market S2000s contains two pieces of information: the car's VIN, and its build date (month/year only). For example, here's the sticker from my own S2000:


It had occurred to me long ago that if you 'collected' enough of these stickers, you could estimate the total production of U.S. cars for every single month: just compare the highest and lowest VINs you'd found for each month! Now, since the stickers don’t specify a calendar day, it’s impossible to know when during the month a given car was made; you can only infer if it was early or late in the month by finding another car: one with a very close VIN, but which was made in the previous or next month.

For example, here are the stickers from two MY00 cars (slightly image-processed to enhance the dates):

Note that the two VINs differ by just 10, but #2372 was made in August ’99 and #2382 was made in September ’99. So 2372 must be very close to the highest (last) 8/'99 VIN, and 2382 must be very nearly the lowest 9/'99 VIN. (In fact, they can be no more than 9 VINs away from those points, respectively.)

So, at last, here’s the game: For each and every pair of consecutive months during the production period, find the best possible “close pair” of stickers: two stickers with numerically close VINs, but produced in successive months. Note that each calendar month will contribute to two close pairs, since it “partners” with both the preceding and following month.

Now consider the theoretically ideal result of such a search: we'd not only find the “close pairs” for every single month, but every close pair would be as close as possible: a pair of consecutive VINs. This would mean that every single sticker corresponds to either the first or the last S2000 made in some particular month. In that case, those 230 or so stickers would be sufficient to determine the production date of every other U.S. S2000, since every possible S2000 VIN could be slotted between the lowest and highest VINs of only one unique month.** In turn, of course, this means that the total number of S2000s produced in each month would also be fully determined.

Of course the immediate question is how one could ever hope to find such a set of stickers! That’s quite a subject in its own right, and I’ll address it in Part 3. For now, let’s continue by observing that our theoretical ideal case leads us directly to a practical goal: minimize the numerical gap between the VINs of each “close pair”. The size of this gap represents the number of “unverified” VINs near the end/start of the successive months. Numerically, this gap is computed as
ClosePairGap = ClosePairHigherVIN – ClosePairLowerVIN – 1.

For example, in the close pair pictured above, the gap is
2382 – 2372 – 1 = 9
which corresponds to the nine MY00 VINs numbered 2373 thru 2381, which cannot be definitively assigned to either 8/99 or 9/99.

Analogously, the number of “verified” VINs for a given month is simply the total number of VINs from the lowest known/verified VIN for that month to the highest known/verified VIN. Numerically, this is calculated as
MonthVerifiedProd = MonthHighVIN – MonthLowVIN + 1.

Now for the crucial observation. In our example above, suppose that in the future I happen to find the sticker for one of those 9 unverified VINs – say, for example, we discover that VIN 2377 has an 8/99 build date. Then VIN 2377 simply replaces VIN 2372 as the new “highest verified VIN of 8/99”. Together with VIN 2382, it forms a new “close pair” with an improved gap: just four unverified VINs remain between the new pair. As for the old VIN 2372 sticker, we need only keep it if we have no other (lower) VINs from 8/99. Assuming we do already have one, then 2372 can safely be discarded. It’s through this iterative process of “narrowing the gaps” (i.e. continually searching for stickers from VINs which fall within an existing gap) that we generate better and better production estimates.***

Having obtained a set of close pairs, the production estimate for each month is computed as the total number of verified VINs for the given month, plus half the gap to the previous month, plus half the gap to the following month:
MonthProdEstimate = MonthVerifiedProd + 0.5*(PrecedingGap + FollowingGap)
(Note: For my estimates, rounding was employed to avoid non-integer values.)
In other words, each month is “credited” with half of the unverified VINs that might possibly have been produced in that month. Note that this estimate is self-consistent and unbiased: the monthly values sum precisely to the actual number of possible VINs.

Given the data gaps inherent in this methodology, it’s useful to quantify the uncertainty levels associated with a given set of close pairs. The following chart provides the current best estimates (solid line) of total USDM monthly S2000 production, along with strict upper and lower bounds (dashed lines). That is, for each month, the actual number of vehicles produced must fall between the upper and lower curves (inclusive).


The bounds are formulated as follows. The lower bound is simply the verified production for the month:
LowBound = VerifiedProd
This bound is pessimistic, since it “discounts” all unverified VINs, omitting them from any month’s totals.

The upper bound is the verified production for the month, plus the entirety of both gaps adjacent to it:
HiBound = VerifiedProd + PrecedingGap + FollowingGap
This bound is optimistic, since it “credits” each month with all of the unverified VINs that might possibly have been produced in that month. Thus, it double-counts the unverified VINs, applying them to both months in which they might have been produced.

So how did I do overall? After a couple of months’ [edit: now, years' ] effort, I’m actually pleased with the relatively tight bounds that are evident in the curves. Here are some statistics on the “close pair” gaps that I’ve ended up with (as of 8 Sep 2023) :
Largest gap: 95 (Oct-Nov ’02)
Average gap: 23.2
Smallest gap: Nil. Eight examples -- April-May ’02; six model-year changeovers; and the Tochigi-Suzuka changeover. (Also Jul-Aug-Sep '03, for MY04 only, which overlapped the end of MY03.)
Total verified VINs: 64,300 (96.0%)
Total unverified VINs: 2,660 (4.0%)

With only a few major exceptions (mostly the months with very low estimated production), the absolute uncertainty –-i.e. the sum of the preceding and following gaps, or the difference between upper and lower bounds-– is on the order of a third [edit: now, an eighth ] of the estimated production value. However, because these are strict bounds (i.e. 100% confidence ranges), in those cases the true production value is statistically likely to be within 10-15% [edit: now, 4-8% ] of the estimated value.


Continued in Part 3…
-------------
**This assumes that no VINs were ever built “out of order”, i.e. a higher VIN produced in an earlier month than a lower VIN. EDIT: This assumption proved false! However, I believe the exceptions are rare enough that they don't make any material difference in the overall numbers. UPDATE: See Part 4, which discusses these cases.
***A technicality: It was also important to determine in which months the changeovers between model years occurred. In most cases I succeeded, having found VINs from consecutive model years that were made in the same month. In the two cases where that never happened (’01-’02 and ’04-’05), I assumed that the changeover occurred right at the month boundary.

twohoos

USDM S2000 Monthly Production, Part 3: Sources (AKA Drive More Carefully!)

Regardless of the mathematical validity and assumptions underlying the methodology described above, the biggest problem with the approach is, indeed, the obvious one: finding all those stickers! Ten years of S2000 production means ~120 months, which means ~120 “close pairs”. Assuming you found stickers at random, you’d surely need to collect thousands before finding that many close pairs! And until recently, I’d have agreed: such a brute-force approach seemed essentially impossible; or, if not, at least doomed to yield so much uncertainty as to be effectively meaningless.

Then came the breakthrough: This summer I stumbled across the Poctra website, which archives the listings of millions of salvage vehicle auctions nationwide. There were over 1,100 (! ) S2000s listed, and the vast majority included –yes!– photos of the federal sticker! Suddenly –actually, quite literally within a few hours– I had a very solid data set to start with, including at least one candidate close pair for almost every month of production.

The next step was to simply comb the web for more stickers, hoping that each one would narrow a gap. For this, BaT listings (where “please show all 10 VINs!” is an inevitable request) were particularly helpful; so were general for-sale listings, many of which included a photo of the driver seat/door jamb at high enough resolution to read the sticker. Still, there were too many big gaps (“close pairs” that simply weren’t). That's when I remembered the S2000 Registry. From the more than 2,000 owners listed, I identified and PM’d a select few s2ki members (ahem...well ok, a couple hundred) whose VINs fell in my remaining gaps; happily, dozens of them were willing to help. (You know who you are: thanks again! )

And yet...it *still* wasn't enough! So, gentle reader, I’m here to confess that yes, at long last, I did it: I started emailing used-car dealers, and even private sellers, when I found S2000s for sale whose VINs would help close the gaps. Now, this stirred up some mild ethical reservations (since I was asking for info/help but had no intention of buying)…but I was unfailingly polite, and in the case of dealerships, I left dozens of positive google reviews for those that went out of their way to help. Conscience soothed.

Some (possibly) interesting sticker examples...
The lowest MY00 VIN I found, #0058, which like too many others, is a wrecked/salvaged car: This is (was) the 16th U.S. production S2000 ever made. Because there are so few earlier VINs, it seems unlikely that U.S. production began in May, but as I said above, we can't *quite* rule out May '99 yet. (And perhaps we never will, unless/until we find VIN #0027!)

And the 23rd U.S. S2000, also a wreck:


An example of the ideal "close pair" -- consecutive VINs in different months:


Here's one of just 49 S2000s ever made--worldwide--with a 5-digit VIN/chassis serial number:

(We can say this with certainty because we know that no other model year in any market, and no chassis series in Japan, ever saw 10,000 production units.)

There are just four valid "0001" U.S. S2000 VINs (from MY01, MY05, MY07, and MY09). I've now run across two of them (sadly couldn't read the date for MY05):

In a similar vein, here's the first-ever production MY06 sticker (VINs 0001-0024 are invalid). That makes it the first-ever U.S. car with drive-by-wire, VSA, TPMS, DRLs, etc... The seller confirmed for me that the date is 9/'05:

[Edit: Sadly, MY06 VIN #0025 later turned up on a salvage auction website.]

Here's one of the last S2000s ever produced at the Tochigi plant (highest '04 "T" VIN is 6062, a Silverstone/Black car).

And one of the first Suzuka VINs (lowest "S" VIN is 0002, a GPW/Tan car), which also happens to prove that Suzuka production began in May '04, not June:

(Sadly, these two '04 examples are both wrecks; moreover, though I didn't get the sticker, MY04 VIN #6061 is also a wreck.)

[UPDATE - March 2021] And now a lengthy aside, in which I'll discuss the end of S2000 production -- both for the U.S. and the world.

To start, here's the sticker from MY09 VIN #348, which is the last-ever USDM base-model S2000 (obviously a GPW car):

(Thanks to the kind owner, who's a member here on s2ki and provided the photo.)
Note that the build date is Nov. '08. After this car, just nine more U.S. S2000s were built, all of them CRs. (Three were GPW; the other six were Berlina Black, including the very last US S2000 ever built, VIN #357.)

Now here's the sticker from one of those last nine cars: MY09 VIN #353 (just 5 cars later):

[Thanks to s2ki user @migzs_s2k.]
As you can see, the build date is Jan. '09, two months later than VIN #348!!

So how could it be that (nominally) two months separate the build dates of MY09 VINs #348 & #353, which were just 5 cars apart? (Remember: at its peak, Honda was producing some 50 U.S. S2000s a day!)

Here's my guess about what happened: After the global financial crash in the fall of '08, U.S. demand essentially vanished overnight. Meanwhile, even before the "Lehman shock", inventory had been piling up at U.S. dealers, and Honda had already slashed U.S. production to a trickle (this is obvious from the data). So, I speculate that Honda must have suspended production of U.S. cars that November, following that final U.S. base model, VIN #348 pictured above. In the weeks after, they would have been busily updating their global market forecasts (for every car in the lineup, most likely) so they could decide how best to respond. Now at that point, a total of 690 CRs had been built. With the S2000’s outlook unquestionably bleak, the final decision to kill the roadster must have been made in late December '08 or early January '09. However (again speculating) the top bosses probably granted approval for a final few CRs: these would serve as a swansong for the model’s ten-year run, and (just as important) would help clear out their inventory of CR-specific parts. (To the latter point: note that Honda also started selling the Type S worldwide for 2009, even though it had been Japan-only when it was introduced – another way to use up those distinctive CR/Type S parts!)

With the decision made, production of those final nine CRs would have occurred in January ‘09, and would have taken just a few days at most. Then it was time to go public: late that month (Jan. 29) Honda issued a press release announcing the demise of the car, with production scheduled to end forever in June. In the U.S., the announcement was greeted with mostly yawns, as there were still hundreds of '08s and '09s sitting around dealers, priced at steep discounts from MSRP. However, the Japan and Europe/UK markets sat up and took notice, and orders came pouring in! With the higher margins available from EDM and JDM versions (where the car was priced, relative to median income, about 20% higher than in the U.S.), Honda found itself with its highest JDM and EDM production quantities since 2001! In fact, JDM sales peaked at 273 units in July 2009, higher than any single month since the 2001 calendar year. The response was so strong that Honda took two unusual steps: first, that March, it announced a limited run of 100 European “Ultimate Edition” cars, plus 100 UK “GT 100 Edition” cars, each in Grand Prix White and equipped with hardtop, unique shift knob, door sill serial, and other bespoke touches. And second, Honda pushed back the end of production from June to August. The last S2000 ever built (one of those special editions, in Grand Prix White) rolled off the line on August 7, 2009 – ten years and five months after production started.

Again, it’s worth noting: U.S. demand for the car tanked in Fall ’08 and never recovered, even after the car’s “death announcement” in January ‘09. From a pure profit perspective, U.S. production probably “should” have stopped entirely after the November ’08 build of MY09 VIN #0348. But there were two more things to consider. First, existing commitments to parts suppliers and unused inventories of CR-specific components almost certainly mean that the Suzuka plant had enough parts on hand for several more cars. Second, the S2000 was the last remaining “halo car” in Honda’s global lineup, with the NSX having been axed years before. So those nine U.S. CRs made in January 2009 (VINs #0349 thru #0357) were almost certainly considered “Final Edition” U.S. cars. True, they were indistinguishable from other CRs, but with the decision to end production at 699 CRs total (“one less” than a round number – a custom started by Ferrari with the F50) Honda unquestionably saw those last few cars as a special tribute to the American market, which had, through sheer volume, sustained the business case for worldwide S2000 production for a full decade.


Continued in Part 4…

twohoos

USDM S2000 Monthly Production, Part 4: In which I undermine my own results

New info and discussion, 11-26-2019:

Well, I guess it had to happen eventually. The fundamental assumption behind my methodology --that no VINs were produced "out of order", i.e. an earlier VIN built in a later month-- has proven false.
Here are a couple of examples where a lower VIN is dated after a higher VIN:



So what's the implication for my results? Well, in the strictest sense, I suppose it throws everything in the bin! If I can't count on the "close pairs" being accurate, then my entire methodology is fundamentally flawed.

Still, I think the old engineering maxim applies: "Better" is the enemy of "Good Enough". Certainly, I can no longer claim that the bounds I calculated above are absolute. But I'm convinced that they must be very close, and that the basic estimates are valid.

How to explain these? Given Honda's well-known quality control, I think these examples point to the vagaries of manufacturing lines, and routine quality control measures. First it's probably unreasonable to expect that cars always proceed down the assembly line in strict VIN order -- even if it's very close (say, plus or minus 10-20 for routine production), there will inevitably be cases where a lower VIN happens to roll out just after the month boundary. Bigger discrepancies (100 or more?) could easily occur if a lower VIN was "held back" briefly when a minor defect was discovered, or as a random/routine quality-check example. Meanwhile, that month's production continued on with higher VINs, and by the time the "hold" on the earlier car was lifted, the next calendar month had begun. Such holds must be brief (in the examples above, the VINs are so close that the two cars would have been completed within a production day or so), and they appear to be extremely rare (I've looked at a *lot* of VIN stickers!).

So overall, I can't imagine these instances making much, if any, impact on the "true" production numbers. Nevertheless, I felt it important to "come clean" about any counter-examples to my assumptions.

...

[Update, Nov'19] Another new data point came in from a recent BaT auction featuring MY04 VIN #0026, the fourth-ever production AP2! Up until now (late-Nov '19) the data in the first post of this thread concluded that MY04 production had begun in Sept. 2003. Well, not so fast:


July!? WTF?! MY03 production was still going strong: over 400 '03s were made in July, over 300 in August, and even a few dozen in September!

Now, MY04 VINs 001-022 are invalid, i.e. presumably AP2 development and prototype examples. What's surprising is how some of these very early AP2 chassis were certified as production examples. [Update, Apr'21]: Things suddenly became clear when I ran across this:

Yes, VIN #028 was apparently two cars and two *months* later! But it gets better: MY04 VIN #027 is invalid! So in fact VINs 026 and 028 are *consecutive*! At last, the pieces all fit: Once the plan for the AP2 was approved, Honda would naturally have begun making early prototypes while MY03 production was still going. Eventually, they would have decided they'd worked out the kinks and had demonstrated a handful of production-quality chassis (VINs 023-026). So they pause production during August while they get final approval and build up the necessary inventory of AP2-specific parts, so that "full-rate" production can continue in September without a hitch. That first September AP2 chassis, VIN #027, is set aside as the final quality-check example, and then production resumes as normal. This scenario fits nicely with the fact that the first Suzuka chassis, VIN #S0001 (built 8 months later), is likewise invalid -- presumably also reserved as an initial quality-check example.

As of 3-Jan-2020, the tables and plots reflect the overlap of late-MY03/early-MY04 production. Separately, I've confirmed that Suzuka production did in fact begin in May '04, having found an example (MY04 VIN #S0009, 5/'04).

-----------------
Original post 10-05-2019:
USDM S2000 Monthly Production, Part 4: Community Contributions (AKA: Carnac)

Nobody:
twohoos: I've just spent three months scouring the web for thousands of S2000 VIN stickers to create an estimate of monthly USDM S2000 production for every model year!
Everybody: ...
twohoos: And, um, now if you give me your U.S. S2000's model year and last 4 VIN digits, I can almost certainly tell you its build date
Everybody and their cousins: .........
twohoos: Cool, right?


In other words: Yes, I know my latest S2000 trivia project might peg your 'why bother'-meter. :lol: But I went down the rabbit-hole anyway, and if you’ve made it this far, congrats: so have you!

Now, the full Part 4 that I’ve planned involves more tables, maybe more charts, heck maybe even more equations! But let’s postpone that, and play a little game instead: You tell me your S2000’s model year and last four VIN digits*, and I’ll “guess” its production date. Here’s the kicker: if I’m wrong, I’ll Venmo/PayPal you one shiny U.S. dollar! (You have to prove I’m wrong with a photo of course.) And if I’m right – nothing! It’s completely risk-free for you!

So, fire away, folks! This is a limited-time offer. [Edit: I'll keep this going indefinitely. In fact, I'll [b]double the offer if your car is an MY03 with VIN between 1675 and 1765!]

*For MY04 models with VIN<1450, please also state whether it's an "S" or a "T" VIN.

GuthNW

Jeez, after all of the effort you deserve some participation...

2005 / 6557

I'm looking forward to more tables and charts. (I had enough equations in school to last me a lifetime but I'm sure they would be welcomed by others, lol.)

twohoos

5/05

Chuck S

Thanks for posting this. I'm going to read this later as I need more coffee! ... but my data is:

MY: 2006
Last 4 VIN: 5421
(Data plate in door jam shows build date: 7/'06)

-- Chuck

stock06

LBP 2006
S-2873

twohoos

2/06

cosmomiller

2009

0086

Nice article.