Singapore Expats Forum

Cheating?

Compilation of interesting and useful articles for Singapore Expats readers.
Sporkin
Chatter
Chatter
Posts: 180
Joined: Fri, 20 Jun 2014

Re: Cheating?

Postby Sporkin » Thu, 27 Apr 2017 9:33 am

Does a non random sampling of a random phenomenon e.g radioactive decay still give you a random distribution? I.e you have a random number series and you obtain slices or uniformed points from it how does that affect the randomness?

Sent from my E6653 using Tapatalk

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Thu, 27 Apr 2017 10:13 am

Here's one for you. Write some code that mimics random coin flipping. My university first year statistics prof always assigned this task at the first of the year: Flip a coin two hundred times and record the results.

He could always tell when people just made up the results without doing the flips. Why? In a truly random environment it's just as likely to get 10 or 12 heads in a row as it is heads/tails/heads/tails.

But, nobody believes that... so they have maximum sequences of 3 or 4 heads.

User avatar
x9200
Moderator
Moderator
Posts: 9288
Joined: Mon, 07 Sep 2009
Location: Singapore

Re: Cheating?

Postby x9200 » Thu, 27 Apr 2017 12:49 pm

Sporkin wrote:Does a non random sampling of a random phenomenon e.g radioactive decay still give you a random distribution? I.e you have a random number series and you obtain slices or uniformed points from it how does that affect the randomness?

Sent from my E6653 using Tapatalk

Providing infinite population of numbers and ignoring deterministic part of the quantum physics - yes.

User avatar
PNGMK
Director
Director
Posts: 4998
Joined: Thu, 21 Mar 2013
Location: Sinkapore

Re: Cheating?

Postby PNGMK » Thu, 27 Apr 2017 6:17 pm

x9200 wrote:PNGMK, what I mean, before the time of easily available PCs, random numbers from white noise were actually available in the printed form in the tables. I am surprised back in 80s you could that easily sample the noise with a receiver and convert it with some sort of DAC with probably needed 16 bit or better resolution and all this real time in a compact device. I am not saying not possible, but I am just surprised. Sounds to me like a challenge for that era, even in 1988. I would think, what was mentioned earlier, or even getting the pseudo numbers from the sound or light ambient noise would be much more optimal solution in any way.



What we actually did was generate an 8 bit random number from a shift register whose timing was a white noise input. The timing was random therefore the out the output was random. To get a bigger number we just added more bits together. This was on a PDP-11 originally and then I ported it to the PC XT platform which from memory was native 8 bit?
I have gay, black, Asian friends and then JR8.

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Thu, 27 Apr 2017 11:10 pm

PNGMK wrote:...and then I ported it to the PC XT platform which from memory was native 8 bit?


Going totally OT for the moment... your comment about native 8 bit (which is correct, BTW), reminded me of the perils of thinking that floating point representation covers the real number universe. It doesn't.

Take a 16 bit floating point variable... 5 bits exponent, 1 sign, 10 bits mantissa... this works out to about 4 decimal digits of precision... at higher... or lower... powers, the gaps between the numbers could be significant... so much so that trying to use 16 bit reals for chaos theory calculations or other applications that assume a continuous real domain, leads to all sorts of errors.

User avatar
PNGMK
Director
Director
Posts: 4998
Joined: Thu, 21 Mar 2013
Location: Sinkapore

Re: Cheating?

Postby PNGMK » Fri, 28 Apr 2017 9:30 am

If you enjoy thinking about bits SE... try to imagine why the Soviets needed a 6 bit ADC that would run at 100MHZ and withstand 500G of accel/deceleration.

https://zeptobars.com/en/read/1107PV3A- ... ADC-soviet

(Zeptobars is a fascinating blog of IC xrays and analysis).
I have gay, black, Asian friends and then JR8.

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Fri, 28 Apr 2017 10:39 pm

PNGMK wrote:If you enjoy thinking about bits SE... try to imagine why the Soviets needed a 6 bit ADC that would run at 100MHZ and withstand 500G of accel/deceleration.

https://zeptobars.com/en/read/1107PV3A- ... ADC-soviet

(Zeptobars is a fascinating blog of IC xrays and analysis).


This is pretty weird. Only 6 bit so only 64 levels of resolution. A 100MHz ADC means a 50MHZ signal can be digitized, something in the range of an ultrasound device, or maybe... I dunno.

But 500G's... whoa! Almost sounds like this is part of some kind of crash test bed setup... a digital streaming of whatever device is being run into a wall at a high rate of speed.

Sporkin
Chatter
Chatter
Posts: 180
Joined: Fri, 20 Jun 2014

Re: Cheating?

Postby Sporkin » Fri, 28 Apr 2017 11:21 pm

Super secret radio homing artillery guidance?

Sent from my E6653 using Tapatalk

User avatar
x9200
Moderator
Moderator
Posts: 9288
Joined: Mon, 07 Sep 2009
Location: Singapore

Re: Cheating?

Postby x9200 » Sat, 29 Apr 2017 11:59 am

500g is actually not that much. In a car crash accident the deceleration may range between (ca) 20 and 200g.
500g is just enough to prevent damage when dropping such chip on a concrete floor from a rather low height (less than half a meter?)

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Sat, 29 Apr 2017 10:12 pm

x9200 wrote:500g is actually not that much. In a car crash accident the deceleration may range between (ca) 20 and 200g.
500g is just enough to prevent damage when dropping such chip on a concrete floor from a rather low height (less than half a meter?)


Your brain will suffer significant damage if subject to a deceleration of 100g over about 5 milliseconds.

9g's is about the maximum a fighter pilot will sustain. The record, for survived deceleration on a test sled is about 46g's.

500 is mighty big for humans.

Acceleration/deceleration is computed as a = (v - v0) / t and if v0 (end condition for deceleration) then simply v / t. So, the shorter the length of time it takes to go from some velocity 'v', the higher the deceleration.

This implies that the shorter the distance over which the deceleration occurs, the higher the deceleration value will be. This makes sense. You whack your head into a concrete floor, which is relatively incompressible, and the resultant g forces will be very high. Encase your head with a motorcycle helmet with 1.5 inches of foam and the distance over which the deceleration occurs is much larger resulting a net reduction in deceleration force.

If motorcyclists were willing to wear larger helmets, injuries would go down even more.

User avatar
x9200
Moderator
Moderator
Posts: 9288
Joined: Mon, 07 Sep 2009
Location: Singapore

Re: Cheating?

Postby x9200 » Sat, 29 Apr 2017 10:55 pm

Yes, fine, correct, but a Russian ADC chip is hardly human so 500g for a chip or as a matter of fact, most modern electronic devices is nothing impressive.

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Sat, 29 Apr 2017 11:28 pm

x9200 wrote:Yes, fine, correct, but a Russian ADC chip is hardly human so 500g for a chip or as a matter of fact, most modern electronic devices is nothing impressive.


True - the average SSD these days are good for 1500 G's over .5 milliseconds.

User avatar
PNGMK
Director
Director
Posts: 4998
Joined: Thu, 21 Mar 2013
Location: Sinkapore

Re: Cheating?

Postby PNGMK » Sun, 30 Apr 2017 4:44 pm

this was for constant 500 G linear acceleration though... not just a single impulse.
I have gay, black, Asian friends and then JR8.

User avatar
x9200
Moderator
Moderator
Posts: 9288
Joined: Mon, 07 Sep 2009
Location: Singapore

Re: Cheating?

Postby x9200 » Sun, 30 Apr 2017 5:50 pm

PNGMK wrote:this was for constant 500 G linear acceleration though... not just a single impulse.

I think it is something different:
1) not angular
2) linear over a fixed period - when a car accelerates, it's roughly linear, but deceleration (e.g. during crash) is almost never linear (could be exponential or worse).

User avatar
Strong Eagle
Moderator
Moderator
Posts: 10570
Joined: Sat, 10 Jul 2004
Location: Off The Red Dot
Contact:

Re: Cheating?

Postby Strong Eagle » Mon, 01 May 2017 2:55 am

x9200 wrote:
PNGMK wrote:this was for constant 500 G linear acceleration though... not just a single impulse.

I think it is something different:
1) not angular
2) linear over a fixed period - when a car accelerates, it's roughly linear, but deceleration (e.g. during crash) is almost never linear (could be exponential or worse).


The point that PNGMK is trying to make is that a car crash, linear deceleration or not, is of a very short duration... .5 to just a few milliseconds.

But, a constant 500g acceleration is something altogether different. A 500g acceleration from rest would have you at traveling about 25,000 meters per second after only 5 seconds of acceleration at that rate.


Return to “Articles”

Who is online

Users browsing this forum: No registered users and 0 guests