Lot Hacking - Is it possible?

[Ambir]

Seriously reading too much into this, the server generates the randoms, and it goes pretty much like this:

iRandomValue = floor(rand() % 1000); // generates a number between 0-999

there's no moon phase/day of the week shit, it's just a C function that comes with the standard libraries

I think the point is that there is no completley random number generation possible by computer (rand() doesn't just pluck a number out of nowhere), and if that is a given, is it possible to determine exactly how the server is creating the number, and then influence the number you get.

The fact that someone achieving this would almost certainly take advantage of it and that would let the mithra out of the bag pretty quickly...

[aurik]

It is possible, but without access to the code that generates the number, you won't figure out the pattern by observation until you've observed a whole ton of consecutive results, and then done a ton of subsequent analysis. Come back to us in 10^x years when you've cracked the code.

IOW, it's random, get over it.

[Raji]

Random number generation.

Far too important to be left to chance.

[Zosi]

I think the point is that there is no completley random number generation possible by computer (rand() doesn't just pluck a number out of nowhere),

Correct.

and if that is a given, is it possible to determine exactly how the server is creating the number, and then influence the number you get.

Toooooootal bullshit.

Assuming they're using system rand(), you can't "influence" anything except via picking when you /random. You'd have to precisely predict two factors: the initial seed, and the number of iterations done at the time your /random is processed. Given both of those, you still have to account for network latency, and when we're talking numbers that can change on a nano/microsecond time scale, a variation of a few milliseconds in your ping means you have zero control over the precise time your /random request hits the server.

The point isn't to have "completely random" numbers, that's impossible with just the mathematical methods that CPUs are limited by. It's to have numbers that are sufficiently hard to predict that they serve the same purpose as random numbers. These numbers are called pseudo-random, and functions that generate them are called pseudo-random number generators. It's a well-researched field, you can read all about it on Wikipedia, let me get you started.

http://en.wikipedia.org/wiki/Pseudor...mber_generator

http://en.wikipedia.org/wiki/Linear_...tial_generator

Now, if they implemented their own PRNG, all bets are off. They could just be doing something astonishingly stupid like calling time() and chopping off all but the last few digits/bits. If they are, well... I have no words, that'd just be stupid beyond measure.

[Delayed]

A long time ago a member in one of our sky Linkshells (back when sky was a big deal) was able to lot on the same item 3, 4, 5 tiems in a row
it would show up in the chat log, and also in the treasure pool as him lotting, and his numbers changing when it would go up.
I don't think he was able to make his lot go higher, but I do know that he was somehow able to lot the same item multiple times without leaving the PT, or changing zones.
He later stole osode ( omg wow ) changed servers and i assume quit life.

[tymora]

Sometimes odd client side glitches can happen if an update is screwed up. There was once I didn't have to move my rod at all to reduce HP on ANY fish while fishing. There were other glitches which forced me to re-install which fixed the free fishing glitch. My bet is that the lot check was done on the client and this was somehow skipped that allowed that person to lot multiple times on the same item.

-----------
That said, why are people thinking they have to do something to the random number generator in order to purposely skew results according to superstitions like moon phase and stuff?

Assuming a fixed available number range (leaning towards 0-1023 since it is used for other ranges like BRD song potency), they just simply could choose to leave rand() as is and modify the range for which a particular result will happen.

Taking a T0 synth for example...~10% break (96/1024), ~3% HQ (32/1024). 0-95 on the scaled random will give a break, 991-1024 will give a HQ. Perhaps Luna will bless your synths by making the HQ range 950-1024 instead? Or perhaps scratching your butt while kissing the floor will make it 900-1024? A considerably simpler implementation assuming they did something like this.

As a side-note, I believe in a base chance of 2^-5 T0(3.125%), 2^-4 T1(6.25%), 2^-2 T2(25%), 2^-1 T3(50%). This should fit in rather decently with observed results by crafters.

[Freysi]

Everybody here is also assuming that each player has their own random function generated by the server. I'll bet there is just one global random function that is called by every single random and lot over all the servers. That way, the seed could even be static and there would be no possible way for someone to figure out the seed, let alone know when to lot/random for a specific number. Just imagine how many people are lotting/randoming across all servers at any given moment.

[Pendulum]

I don't anything about random number generators or manipulating them but it reminded me of this ;x

http://img177.imageshack.us/img177/8659/dilbertbf0.jpg

[Sylvrdragon]

The funny thing about this statement is that nothing is ever truly random on a computer, and as technology progresses, the likelihood of us working toward encountering truly random anything becomes better and better, not vice versa.

Technically, nothing in the entire universe is/was/ever can be truly random. Every wave, and particle's position and velocity etc have an exact cause, which in turn has it's own cause regressing all the way back to the Big Bang. We just don't have the tools to extrapolate all of the information (or even a microscopic fraction). Probability theory itself is just a stand-in; there is actually no such thing.

I just realized what a fucking nerd I am... I'm going to sleep...

[isladar]

Its not possible; POIDH, also excell spreadsheets with 100,000 (at least) results showing an ability to manipulate the lot system, then we reopen. Otherwise we just get into "well one time this one guy I knew? In sky? Well he lotted a bunch of stuff one time, and the numbers? they were really close/high/low/I'm fishing for patterns".