Starcraft Keys

[Trejo]

I have a program under development that is about 65% done, it is in C++ http://www.grimhosting.com/grim`mass - Check it out. I'm going to put the list of 1.39GB of keys I have (after using the check the installer uses and running 0000000000000-9999999999999 through it) and let it just run, it auto-refreshes proxies and i'm going to have a scanner running 24/7 outputting the the bot's proxie file. I'll post the results here.

That's really not a good plan.  It'll probably take you a few years or so to complete...

[GoaL]

It's not just a list from 0000000000000-9999999999999 it's all them run through the check the installer uses then discarding the ones that don't meet the check. So how do you figure that so joe?

[Arta]

Many of them will have the wrong product value. This has already been discussed. If you want to do this, you should create valid decoded keys and then encode them for testing.

[Lenny]

I haven't gone through all the pages, but the only feasable way I see this problem can be approached is statistically. 

No one is going to find the something that generates keys at a 100% success rate.


But I'm not sure why this post was ignored.  IMO, this is very useful. 

Is value 1 always equal to value 2?  At a glance it appears to.  Or is this a bug in your output?

No it's a bug. I have generated a csv file full of these cdkeys and their public/private/product values in decimal. One thing I noticed is the private values range from [0, 999], and the public values range from [1111, 7405972]. So brute force of the private value given the cdkey hash is extremely easy if the range holds true for all cdkeys.

Edit: I have generated some scatter plots of the data.

Public vs Private for Product Id = 1


There is a big ass line going down the middle. Interesting.

Public vs Private for Product Id = 2


This is even more interesting. Might just be bad data though.

Let's assume if dxoigmn had the entire population to graph that it would follow the pattern in the second graph giving it solid lines.  Now if the public and private values fell on those lines, wouldn't it have a much higher chance of being valid than it falling somewhere else?

So this might be worthwhile to check on a decoded key (I haven't worked on anything Bnet in a very long time, so I won't be testing this):
-If it's one of the 10 public values seen in the second graph
-Because for each public value, there are 200 private values (integers only), check if the associated private value falls in the 300-400 range or the 600-700 range

Now one of you has generated a list of decoded CDkeys that can verify or prove this inference wrong, since the population in that list is bound to include invalid bnet keys.  If you were to graph it and the same pattern arose, then what I have said is useless....

One problem I already see if all valid keys were to follow that pattern, there would only be at most 200,000 valid keys in existence on Bnet.  But maybe we don't have the entire population.  It's very possible that Bnet uses multiple CDkey algorithms. 

Or (if 9 million copies of SC were sold according to wikipedia), Bnet has been banning 45 people for every one CDkey , which people wouldn't really notice.

[UserLoser]

What program was used to create those graphs?  I have 35151 CdKeys that should be graphed...could be helpful .  A few thousand are sequential via public value...And I even have keys with public values 2-9..that's interesting, eh?

[PaiD]

UL: I thought you where against this?

[GoaL]

So wait, if I was to input my 1.39GB of keys into that.... wouldn't the graph that gives me/everyone be like the exact values needed?

[dxoigmn]

What program was used to create those graphs?  I have 35151 CdKeys that should be graphed...could be helpful .  A few thousand are sequential via public value...And I even have keys with public values 2-9..that's interesting, eh?

excel

[GoaL]

Figures, has to be one I don't have. 

[Lenny]

Could someone generate a small list (500 or so) of cdkeys with their public and private values (with a product ID of 2) based off the installer check only?

This way I can determine if the trends in dxoigmn graph are useful or not...

[GoaL]

How bout all of them? 0000000000000-9999999999999 with installer check? :-D ?

[UserLoser]

Could someone generate a small list (500 or so) of cdkeys with their public and private values (with a product ID of 2) based off the installer check only?

This way I can determine if the trends in dxoigmn graph are useful or not...

How about 900?  Check PM.

[dxoigmn]

Could someone generate a small list (500 or so) of cdkeys with their public and private values (with a product ID of 2) based off the installer check only?

This way I can determine if the trends in dxoigmn graph are useful or not...

Not going to do you any good.

[Lenny]

Here is the figure generated from the data UL sent me (thanks UL).

As you can see, the graph shows that the private values range from 0-1000 when invalid keys are included.  From I have now, it's fair to conclude that if the private value does not fall in the 300-400 range or the 600-700 range, it's not valid to Bnet.


Also, to correct my earlier post, I have found that there is a unique public value assigned to every cdkey (or at least there is for the csv file dxoigmn posted) (thanks dxoigmn).  But this unique public value only spans certain ranges, the lines in dxoigmn's figure are actually blocks which represent those number ranges.  Because the resolution of the x axis is so high, they only appear to be lines.  My figure shows one of those 20 blocks.

[Joe[x86]]

GoaL, I hate to break it to you, but I'm not stupid. If you would be so kind as to look at SCKMG, you'd realize that I've already done exactly what you're telling me I should be doing.