9% of VINs are valid

[wildjeff]

Here is something probably no one but me would find interesting.

I wrote a little app to generate valid VIN numbers for thirdgens and with any given check digit there are 9091 valid VINs.

for example for VIN 1G2AW87H0CL5xxxxx where the 0 is the check digit, there are 100000 different possible VINs but only 9091 or about 9% will calculate out correctly to a valid VIN.

of course changing any other part of the VIN, such as changing the W (T/A) to a X (S/E) will create a whole new set.

[TAdan]

Nice work Jeff. You are not the only one, thanks for posting it.

Once a Vin is generated how do you know it is valid? How do you cross check your generated Vin against cars that were actually built?

This is a pretty cool app. Any other cool stuff in the works?

[wildjeff]

Glad you like it.

The ninth digit is the check digit and it is calculated based on all the other digits in the VIN. When Pontiac went to build a car, say a T/A, they give the car a sequence number and already know everything else.

1G2AW87H(unknown check digit)CL500001

Now they have everything except the check digit, then they then calculate (the formula is on the internet, I don’t know it off the top of my head) the check digit and WHAM they have a valid VIN which goes on the car and to the government.

So let’s assume that 1G2AW87H7CL500001 is a built car. Even though 1G2AW8778CL500001 is also a valid VIN it will never exist because there is only one 500001 per plant. (1G2AW8772CN500001 could exist because it is a different plant and note the check digit is different too).

As far a cross checking if the car was actually built. I emailed both Autocheck.com and carfax.com and they both said they don’t do custom jobs. You can enter them one by one but that would be time consuming. It would also only show the cars registered within the last 15 years or so since they started using computer records. IF a car was bought in 82 and wrecked in 83 then scraped in 84 it probably won’t show up in their records. So I’m still working on that one.

I have a bunch of projects going on.
*I’ve updated my VinTool program but haven’t uploaded it yet to my site
*working on redesigning my site with a database backend, this will make updates easier
*working on putting all my data in database.
*still working on D W B N S
*trying to collect Firebird magazine or anything 82-84 firebird

To much stuff

[TAdan]

Good Stuff Jeff!

I didnt realize that there was a Formula for the check digits. Hmmmm... Could this type of thing be applied to 4th gens? I keep a registry of 25th T/A's... and the ability to formulate legitimate vin numbers could be useful in filling in some of the blanks in my list... I'll have to look around online and see if I can find what formula the factory used to generate the check digit.

[wildjeff]

Dan, I can mod my app to work with 4th gens pretty easy. I'll give it a go tonight and upload it to my site tonight and post a link. you can try it and see if it does what you want.

[TAdan]

That would be great jeff thanks. Dont knock yourself out if it is too difficult.

A normal 1994 LT1 Pontiac Vin would be: 2G2FV22P-R22----- (the first - is the check, the last - are the build sequence).

On my list I have a few cars where I am missing the check digit.

[wildjeff]

I've have the new version of VinTools ready. I have added decoding for 93 to 02 firebirds and Camaros. I also added a vin finder feature that will allow you to find all the valid vins in a given range.

You can down load it at

www.82firebird.com/VinToolsv3.msi

My old version was written in VB6 and the new version is written in VB.net 2008. What does that mean? It means you'll need to install .NET Framework 3.5 Service Pack 1, if you don't have it already installed. The installer will check if you have during setup. It also requires Windows Installer 3.1 to be installed. the app I wrote is .5 meg but the .NET Framework 3.5 is 58 meg. FYI

If you try it, let me know how it goes. There are some small things I have to do but the VinTools is fully functional .

[jeremynyr]

I think your calculation, in a roundabout way, merely shows the obvious.... that there are 10 possible check digits, and only one will be valid for a particular VIN (10% validity rate compared to your finding of 9%). Instead of having your check digit be variable and all other digits fixed, you just reversed it and kept the check digit valid while all other digits were variable.

Please don't take my post as being negative... I know I'm new on this site and don't want to make a bad impression. I could be wrong in my thinking. I just saw what I described above as a distinct possibility from how I read your post.

[wildjeff]

No problem, That about sums it up. I built my app to help me verify VINs. For example when I find a car online, like ebay and the VIN is not listed I can use my app to narrow down the possibilities and use details about the car to find the correct VIN.