Skybuck's Racing Random Number Generator

Jim, please, not even you are that stupid.

Did you actually look at the code.

Logically it's impossible for such a sequence to occur.

It's some kind of hardware problem.

And this is ofcourse just experimentation.

I will try again and then you shall bite maybe the dust ! =D

Bye,
Skybuck
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Skybuck Flying schrieb:
> It's some kind of hardware problem.

Yeah! Today's hardware is incompatible with Skybuck's code.
ROTFLABMFTGTVEGOOMF.

--
Reporter, n.:
A writer who guesses his way to the truth and dispels it with a
tempest of words.
-- Ambrose Bierce, "The Devil's Dictionary"

....und wech
Danny <dannys9 (at) gmx (dot) de>
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Skybuck Flying wrote:
> Jim, please, not even you are that stupid.
>
> Did you actually look at the code.
>
> Logically it's impossible for such a sequence to occur.
>
> It's some kind of hardware problem.
>
> And this is ofcourse just experimentation.
>
> I will try again and then you shall bite maybe the dust ! =D
>
> Bye,
> Skybuck
>
>
wrong - - it is your understanding of the random aspects of
threads and code and timing - - -

are you blaming all PC hardware for not being compatible with your code?
and only with your code?

while it works for everything else - - - is that what you just said

maybe stop the smoking or what ever you do - - -

Jim P.
 >> Stay informed about: Skybuck's Racing Random Number Generator 

"Skybuck Flying" <spam RemoveThis @hotmail.com> wrote in message
news:e5bb5$476da400$541983fa$20646@cache5.tilbu1.nb.home.nl...
> Here ya go guys and gals.
>

Here ya go, Skybuck: here is a way to generate a random number
that really works. (Arithmetic follows the normal rules of
integer arithmetic: division truncates toward zero).

Start_Time = ( System.TimeInMillseconds() / 1000 ) * 1000;
Random_Number = SHA_256( Start_Time );

While ( ( System.TimeInMilliseconds() - Start_Time ) < 1000 )
{
Random_Number = SHA_256( Random_Number );
}

Analysis: in order to predict the result, the attacker must
know the exact time in milliseconds when the process started,
and he must know the exact number of iterations that will occur
in the while loop before the exit condition is satisfied. This
will, in general, depend on the speed of the particular CPU and
the implementation of SHA_256 as well as on the Start Time. In
a multiprogramming environment, system load and cycle-stealing
by drivers will also have a randomizing effect on the number of
iterations that can be completed.

You can depend on this algorithm for *one* random number.
However, this one random number can be used to seed a complex
chain of pseudo-random numbers by continuing the "Random_Number
= SHA_256( Random_Number )" theme.

For the benefit of the uninitiated, SHA_256 is a hash function
approved by NIST for cryptographic applications used by the US
Government. So far as is known, it's about the closest thing
there is to a one-way (non-invertible) function.

http://csrc.nist.gov/publications/PubsDrafts.html#FIPS-180--3
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Seems pretty easy to predict if the algorithm is known.

Suppose the time is returned in a longword which is 32 bits.

There are 4 billion possibilities for the starting time.

Simply try all values from StartTime zero to 4 billion until a match is
found.
Or simply start from the current time.

From that moment onwards execute the while to find the next number until
again a match is found.

One might not be able to predict a single number accurately but one might be
able to predict a likely range wherein the number will be

The hash itself is useless at preventing prediction since it will simply be
called by the attacker in a trail and error fashion.

It only makes the number look a bit more random

Also why first divide by 1000 milliseconds and then multiply by 1000
milliseconds again ?

Is that to illiminate lower fast cycling bits ? Why ?

Seems like the timer will vary with one second intervals.

Could be very predictable.

Needless to say I don't like it very much but thanks anyway =D

Always fun to read other people's tries

Bye,
Skybuck.

"John E. Hadstate" <jh113355.DeleteThis@hotmail.com> wrote in message
news:13n64s99e18ci31@news.supernews.com...
>
> "Skybuck Flying" <spam.DeleteThis@hotmail.com> wrote in message
> news:e5bb5$476da400$541983fa$20646@cache5.tilbu1.nb.home.nl...
>> Here ya go guys and gals.
>>
>
> Here ya go, Skybuck: here is a way to generate a random number that really
> works. (Arithmetic follows the normal rules of integer arithmetic:
> division truncates toward zero).
>
> Start_Time = ( System.TimeInMillseconds() / 1000 ) * 1000;
> Random_Number = SHA_256( Start_Time );
>
> While ( ( System.TimeInMilliseconds() - Start_Time ) < 1000 )
> {
> Random_Number = SHA_256( Random_Number );
> }
>
> Analysis: in order to predict the result, the attacker must know the exact
> time in milliseconds when the process started, and he must know the exact
> number of iterations that will occur in the while loop before the exit
> condition is satisfied. This will, in general, depend on the speed of the
> particular CPU and the implementation of SHA_256 as well as on the Start
> Time. In a multiprogramming environment, system load and cycle-stealing
> by drivers will also have a randomizing effect on the number of iterations
> that can be completed.
>
> You can depend on this algorithm for *one* random number. However, this
> one random number can be used to seed a complex chain of pseudo-random
> numbers by continuing the "Random_Number = SHA_256( Random_Number )"
> theme.
>
> For the benefit of the uninitiated, SHA_256 is a hash function approved by
> NIST for cryptographic applications used by the US Government. So far as
> is known, it's about the closest thing there is to a one-way
> (non-invertible) function.
>
> http://csrc.nist.gov/publications/PubsDrafts.html#FIPS-180--3
>
>
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Skybuck Flying wrote:
> Seems pretty easy to predict if the algorithm is known.
>
> Suppose the time is returned in a longword which is 32 bits.
>
> There are 4 billion possibilities for the starting time.


>
> Simply try all values from StartTime zero to 4 billion until a match is
> found.
> Or simply start from the current time.
>
> From that moment onwards execute the while to find the next number until
> again a match is found.
> \

match to what - - - random numbers do not have a match - - - that would
remove the random number if run till a certain number shows up - - -

> One might not be able to predict a single number accurately but one might be
> able to predict a likely range wherein the number will be

predict - - you want a random number generator - - not a number
predictor - - - correct - - -

if random that is not possible - - it is random -- no ability to predict
- - only if the system does not generate random numbers can one predict
- - - - so why this talk about random number generator and predicting
the next number - - - - strange but then you are
>
> The hash itself is useless at preventing prediction since it will simply be
> called by the attacker in a trail and error fashion.
>
attacker in a random number generator - - - Strange
but government level encryption takes billion of tries or years
on supper computers to crack so that is not a valid concern


> It only makes the number look a bit more random
more random than your concept of course it does.
>
> Also why first divide by 1000 milliseconds and then multiply by 1000
> milliseconds again ?
>
> Is that to illiminate lower fast cycling bits ? Why ?
>
> Seems like the timer will vary with one second intervals.
>
> Could be very predictable.
>
could it - - you have to know the exact second that the code was run
and only then

> Needless to say I don't like it very much but thanks anyway =D
>
you do not like any idea but your own - - - as bad as it is - -


> Always fun to read other people's tries

and but them down without knowledge of what you are saying
of course you do.

and note that your BS and stuff on this subject is going to all of these
newsgroups so you are really making a fool of yourself all over the
place.

Jim P.
>
> Bye,
> Skybuck.
>
> "John E. Hadstate" <jh113355 RemoveThis @hotmail.com> wrote in message
> news:13n64s99e18ci31@news.supernews.com...
>> "Skybuck Flying" <spam RemoveThis @hotmail.com> wrote in message
>> news:e5bb5$476da400$541983fa$20646@cache5.tilbu1.nb.home.nl...
>>> Here ya go guys and gals.
>>>
>> Here ya go, Skybuck: here is a way to generate a random number that really
>> works. (Arithmetic follows the normal rules of integer arithmetic:
>> division truncates toward zero).
>>
>> Start_Time = ( System.TimeInMillseconds() / 1000 ) * 1000;
>> Random_Number = SHA_256( Start_Time );
>>
>> While ( ( System.TimeInMilliseconds() - Start_Time ) < 1000 )
>> {
>> Random_Number = SHA_256( Random_Number );
>> }
>>
>> Analysis: in order to predict the result, the attacker must know the exact
>> time in milliseconds when the process started, and he must know the exact
>> number of iterations that will occur in the while loop before the exit
>> condition is satisfied. This will, in general, depend on the speed of the
>> particular CPU and the implementation of SHA_256 as well as on the Start
>> Time. In a multiprogramming environment, system load and cycle-stealing
>> by drivers will also have a randomizing effect on the number of iterations
>> that can be completed.
>>
>> You can depend on this algorithm for *one* random number. However, this
>> one random number can be used to seed a complex chain of pseudo-random
>> numbers by continuing the "Random_Number = SHA_256( Random_Number )"
>> theme.
>>
>> For the benefit of the uninitiated, SHA_256 is a hash function approved by
>> NIST for cryptographic applications used by the US Government. So far as
>> is known, it's about the closest thing there is to a one-way
>> (non-invertible) function.
>>
>> http://csrc.nist.gov/publications/PubsDrafts.html#FIPS-180--3
>>
>>
>
>
 >> Stay informed about: Skybuck's Racing Random Number Generator 

"Skybuck Flying" <spam.DeleteThis@hotmail.com> wrote in message
news:d0ff$47733ab3$541983fa$14805@cache4.tilbu1.nb.home.nl...
> Seems pretty easy to predict if the
> algorithm is known. Suppose the time is returned in a
> longword which is 32 bits.
> There are 4 billion possibilities for the starting time.

How many possibilities are there if the time is returned as a
64-bit longword?

> Simply try all values from StartTime zero to 4 billion until
> a match is found. Or simply start from the current time.

Maggie's drawers! You entirely missed the target.

> From that moment onwards execute the while to find the next
> number until again a match is found.

Once again, your reputation as an illiterate preceeds you. I
said it was good for *one* random number.

> One might not be able to predict a single number accurately
> but one might be able to predict a likely range wherein the
> number will be

You have two dimensions of complexity: you have to guess the
number of iterations through the while() loop, and you have to
guess the starting time. Be wrong by a single bit, and you're
hosed.

> The hash itself is useless at preventing prediction since it
> will simply be called by the attacker in a trail and error
> fashion.
> It only makes the number look a bit more random

If you hash a random number a random number of times, will the
result be any more predictable? Have you ever tried to work
backwards through a crytographic hash? Ever succeeded?

> Also why first divide by 1000 milliseconds and then multiply
> by 1000 milliseconds again ?
> Is that to illiminate lower fast cycling bits ? Why ?

Ah-ha! Now Skybuck is getting smarter!

> Seems like the timer will vary with one second intervals.
> Could be very predictable.

It seems like it, but you have to think this one through.
Here's a hint: suppose the first line is executed when the
system time is 12345678 milliseconds. Start_Time will be
assigned a value of 12345000. The while() loop will continue
until the returned time in milliseconds is at least 12346000.
But how many times did the while() loop execute? It depends on
the actual start time, and the system load, and the speed of
your particular CPU, and any variations in clock tick rate
caused by thermal effects, etcetera, etcetera, and so forth.

> Needless to say I don't like it very much but thanks anyway
> =D Always fun to read other people's tries

Think of it as a gentle laxative for Skybuck's brain
 >> Stay informed about: Skybuck's Racing Random Number Generator 

"Jim P" <Jim_P.RemoveThis@mad.scientist.com> wrote in message
news:tOqdnapty7V02e7anZ2dnUVZ_tWtnZ2d@comcast.com...

>
> match to what - - - random numbers do not have a match - - -
> that would remove the random number if run till a certain
> number shows up - - -
>

Some background information: "Skybuck Flying" is an
afficionado of his own ad-hoc cipher algorithms. He has
trolled sci.crypt and sci.crypt.random-numbers for several
years under a couple of nyms. His posts always follow the same
tired pattern: drop a mental turd in the middle of the dinner
table and then spend a week or so smearing it around while the
rest of the company tries to clean it up. Then vanish for
several months. Lather, rinse, and repeat.

>
> predict - - you want a random number generator - - not a
> number predictor - - - correct - - -
> if random that is not possible - - it is random -- no ability
> to predict - - only if the system does not generate random
> numbers can one predict - - - - so why this talk about random
> number generator and predicting the next number - - - -
> strange but then you are

In cryptography, we talk about randomness (measured as "bits of
entropy") and unpredictability, which we really don't know how
to measure. Perforce, a cipher must be both deterministic and
unpredictable (otherwise, the intended recipient could not
decrypt the message or else any unintended recipient could).
However, for various technical reasons, we would like to have
truly random numbers for such things as cipher keys (we then
wrap these using techniques like asymmetric encryption and send
them along so the intended recipient (and only he) can recover
the message. The random cipher keys need to be random so that
if an attacker should unravel the asymmetric key wrapper, he
won't be able to compromise more than one message (he won't be
able to predict the next random key value).

> attacker in a random number generator - - - Strange but
> government level encryption takes billion of tries or years
> on supper computers to crack so that is not a valid concern

Actually, with computers that are 1,000,000 times more powerful
than current technology allows, it is longer than the estimated
lifetime of the universe. But, that's only if the technology
is applied correctly, and if Skybuck is the implementer, it
probably won't be
 >> Stay informed about: Skybuck's Racing Random Number Generator 

For that to work you would:

1. Need to discover a pattern.

2. Need to gamble on the pattern.

In this case the pattern and gamble would be to bet on the last outcome of
the lotery ?!

Do you know any persons that bet on the outcome of the last lotery ?

Also such betting during a poker game might be deadly if you assume you ll
get another 4 aces or whatever

But ok point taken

Bye and thanks,
Skybuck

<sayembara DeleteThis @gmail.com> wrote in message
news:83093c20-f3b0-484c-9eb9-02bf337b2271@s12g2000prg.googlegroups.com...
> On Dec 23, 7:57 am, "Skybuck Flying" <s... DeleteThis @hotmail.com> wrote:
>
>> Number 11 A: -1290957144
>> Number 12 A: 830628039
>> Number 13 A: 830628039
>> Number 14 A: 830628039
>> Number 15 A: 830628039
>> Number 16 A: 830628039
>> Number 17 A: 830628039
>> Number 18 A: 830628039
>> Number 19 A: 830628039
>> Number 20 A: 830628039
>> Number 21 A: 830628039
>> Number 22 A: 830628039
>> Number 23 A: 843795460
>> Number 24 A: -1164565777
>
>> Number 71 A: -1104423537
>> Number 72 A: -2127767234
>> Number 73 A: -2127767234
>> Number 74 A: -2127767234
>> Number 75 A: -2127767234
>> Number 76 A: -2127767234
>> Number 77 A: -2127767234
>> Number 78 A: -961332962
>
> The predicter program's main code: Number[A]=Number[A-1];
>
> If this is a lottery (assume $1 ticket wins $1,000,000 jackpot), after
> losing $11, I will win $14,000,000 at series 12 A. Then after losing
> another $47, I will win $6,000,000 at series #71 A. The lotto's hitman
> surely assigned to wreck some programmer's life.
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Skybuck Flying wrote:
> For that to work you would:
>
> 1. Need to discover a pattern.
>
> 2. Need to gamble on the pattern.
>
> In this case the pattern and gamble would be to bet on the last outcome of
> the lotery ?!
>
> Do you know any persons that bet on the outcome of the last lotery ?
>
> Also such betting during a poker game might be deadly if you assume you ll
> get another 4 aces or whatever
>
> But ok point taken
>
> Bye and thanks,
> Skybuck


what does a random number generator have to do with the lottery?
what does a random number generator have to do with poker and 4 aces?

Why would a random number generator have a pattern? if it is a random
number generator it should not have a pattern - -

and why not use the last outcome to seed the random number generator
that runs many cycles - - - really makes for a random number - - as good
as can be - - - -

again your total lack of understanding of a random number generator is
showing - - - big time

Jim P.



>
> <sayembara.TakeThisOut@gmail.com> wrote in message
> news:83093c20-f3b0-484c-9eb9-02bf337b2271@s12g2000prg.googlegroups.com...
>> On Dec 23, 7:57 am, "Skybuck Flying" <s....TakeThisOut@hotmail.com> wrote:
>>
>>> Number 11 A: -1290957144
>>> Number 12 A: 830628039
>>> Number 13 A: 830628039
>>> Number 14 A: 830628039
>>> Number 15 A: 830628039
>>> Number 16 A: 830628039
>>> Number 17 A: 830628039
>>> Number 18 A: 830628039
>>> Number 19 A: 830628039
>>> Number 20 A: 830628039
>>> Number 21 A: 830628039
>>> Number 22 A: 830628039
>>> Number 23 A: 843795460
>>> Number 24 A: -1164565777
>>> Number 71 A: -1104423537
>>> Number 72 A: -2127767234
>>> Number 73 A: -2127767234
>>> Number 74 A: -2127767234
>>> Number 75 A: -2127767234
>>> Number 76 A: -2127767234
>>> Number 77 A: -2127767234
>>> Number 78 A: -961332962
>> The predicter program's main code: Number[A]=Number[A-1];
>>
>> If this is a lottery (assume $1 ticket wins $1,000,000 jackpot), after
>> losing $11, I will win $14,000,000 at series 12 A. Then after losing
>> another $47, I will win $6,000,000 at series #71 A. The lotto's hitman
>> surely assigned to wreck some programmer's life.
>
>
 >> Stay informed about: Skybuck's Racing Random Number Generator 

On Thu, 27 Dec 2007 05:09:15 -0500, "John E. Hadstate"
<jh113355.DeleteThis@hotmail.com> wrote:

>Ah-ha! Now Skybuck is getting smarter!

A rare compliment indeed.

>Think of it as a gentle laxative for Skybuck's brain

You are a bad man..that's way to subtle for him!

--
Z
 >> Stay informed about: Skybuck's Racing Random Number Generator 

On Wed, 26 Dec 2007 22:23:35 -0800, Jim P <Jim_P.DeleteThis@mad.scientist.com>
wrote:

>and note that your BS and stuff on this subject is going to all of these
> newsgroups so you are really making a fool of yourself all over the
>place.

He's testing you - how red can you make that wall just by using your
head?

--
Z
 >> Stay informed about: Skybuck's Racing Random Number Generator 

Skybuck Flying wrote:
>
>For that to work you would:
>
>1. Need to discover a pattern.
>
>2. Need to gamble on the pattern.
>
>In this case the pattern and gamble would be to bet on the last outcome of
>the lotery ?!
>
>Do you know any persons that bet on the outcome of the last lotery ?

If the lottery uses Skybuck's Racing Random Number Generator and thus
has the same winning numbers 7 days in a row, I rather expect a LOT
of people betting on them repeating again on day 8!

Time to ignore the crossposting troll.

*plonk*
 >> Stay informed about: Skybuck's Racing Random Number Generator