Regarding the Jumbo Cactpot

Hi! I just had another night of cashing in my weekly jumbo cactpot tickets. I actually got one 4th prize this week!

That got me thinking.

Let's assume I only wish to match the last number and the cactpot winning number is created by a randomly generate number program.

Are my chances of matching the last number truly three in ten or are certain numbers more likely to be generated by a RNG program?

Funny enough, I'm in the same boat. First time I got 4th in a while.

For a presumed 3 in 10 chance, I sure do fail to get anything better than consolation most weeks...

Quote:

---

Originally Posted by Hasrat

Funny enough, I'm in the same boat. First time I got 4th in a while.

For a presumed 3 in 10 chance, I sure do fail to get anything better than consolation most weeks...

---

Same! I would assume I would get a 4th prize way more often than I do but it's so rare than it's a surprise when I win. It's definitely rarer than three hits every two and a half months. Something feels off.

Quote:

---

Originally Posted by Kacho_Nacho

Hi! I just had another night of cashing in my weekly jumbo cactpot tickets. I actually got one 4th prize this week!

That got me thinking.

Let's assume I only wish to match the last number and the cactpot winning number is created by a randomly generate number program.

Are my chances of matching the last number truly three in ten or are certain numbers more likely to be generated by a RNG program?

---

It's known that no computer can truly draw a random number, however if you wanted to predict what an RNG program was going to draw, you would need access to its code and the higher math skills to draw the correlations.

I wouldn't worry too much about it. I've been playing the same three numbers for four years and I've never done better than 3rd.

Yes, the RNG is explicitly rigged to draw a different last number than you specifically have chosen the vast majority of the time. This all despite the fact that the jumbo cactpot number is the same for the whole server.

All jokes aside, if you pick a different last number for all three tickets, your chances to win the 4th prize is 30%. Given you can only play once a week, a 30% rate isn't that great as even on average, you could only be expected to win the 4th once a month, which is pretty much in line with my experience over many years.

While computer number generation cannot ever be 'true random', pretty much every software program these days that requires RNG uses formulas and generation techniques to make it as close to true random as it can possibly achieve. Unless you were to go through datasets of trillions of generations, it'd look like true random to the average person.

Keep in mind, it's only a 30% chance (assuming you play 3 different end numbers) that occurs once a week, meaning you're going to have negativity bias purely because the odds of winning are low vs how infrequently you can actually participate. Go back to any old pokemon game and spam 1-hit K.O moves and tell me how many times you manage to actually get them to land over hundreds of uses - that's exactly the odds you're working with here. Each time you fail to land it is another week in FF14 of the Jumbo cactpot.

A 30% chance to win one and lose two versus a 70% chance to lose all three is in line with my personal experience. I roughly win one ticket every three weeks to a month.

The total number of won tickets should be 10%.

Quote:

---

Originally Posted by JamieRose

It's known that no computer can truly draw a random number, however if you wanted to predict what an RNG program was going to draw, you would need access to its code and the higher math skills to draw the correlations

---

Thats because random isnt actualy random, every single particle can, and will affect the result. The trick is making this too difficult to predict, and use a system to translate this to a number that cannot be influenced.

For example. To generate a random 8 bit digit, you want a system in which you can ensure that all 8 bits are perfectly randomized. And there are tricks for this (inefficient, but because of that reliable). But it does require you to know weaknesses in the system:
- For example, each of the 8 bits, must effectively be influenced by all other bits before and after. (this waterfall effect is also demanded in encryption for the very same reason)
- For any spike in activity (throwing off RNG normaly), the spike must also be balanced out.

A simple way to do this (and i probably miss taking some factors here):
Take the number of options, this is your base. Now you take any weakness numbers (for example if powers of 2 are weak, you take 2), this normaly means all primes below that number you selected. Now multiply all these together (for example if you want a random number between 1 and 4, you take 2x3x4=24 bits), and then take the modulo of the number of options from them. Using tricks you dont need a GB or more of data even when calculting numbers within millions (while it generates, you can already apply modulo at certain steps).

This method is already that reliable, that even if a burst of 1's gets generated by outside activity, the primes cause this to cascade as a waterfall upon the number. And because the speed of generating is faster than the backward calculation (even in quantum physics! if you are using a QP safe method of obtaining the bits), the final number is safe from external influences. Especialy since you must set a treshold of bits in which you can ensure no external manipulation is done (at least 40% of the bits have to be a 1, and 40% must be 0). And these are kept relatively open because RNG has bursts naturaly happen. Its just to avoid a burst attack from changing too much values. Because of all the prime factors, even if half of them are negated in value, cascading still happens reliably.

Its slow for generating a lot of numbers, but lotteries do depend on things like this, and probably have even more tests to detect potential external influence.

Most RNG systems demand to be faster, which causes the weaknesses to appear. But perfectly random info can be generated, as long as you can be sure that the external source is not influenced.

Thank you for reminding me this exists damn

With the CPU RNG out of the way, I just want to point out that buying three tickets does not increase your chance to 3/10. Each ticket is a 1/10 chance to get 4th prize.

I run the same numbers on all tickets. 9947 (no special reason). Every week, since my start. I have played 90-91 weeks.

I call it Bingo/Keno style. All or nothing.

I have won 856k MGP for its worth. My best finish is 3rd, once. I have no idea its thats good or bad tbh.

This style will generate quick big wins or you will be stuck for years.

Quote:

---

Originally Posted by Gemina

With the CPU RNG out of the way, I just want to point out that buying three tickets does not increase your chance to 3/10. Each ticket is a 1/10 chance to get 4th prize.

---

If you buy 3 tickets and a different last number for all 3, yes, it does increase it to 30%, but only one of them can win. You have 10 possible last numbers and you have picked 3 of them (3 out of 10). The jumbo cactpot number doesn't change between tickets, it's the same for the whole server.

Quote:

---

Originally Posted by Gemina

With the CPU RNG out of the way, I just want to point out that buying three tickets does not increase your chance to 3/10. Each ticket is a 1/10 chance to get 4th prize.

---

It is 3/10 per weekly draw (assuming the last digits of each ticket are different), because "random number matches ticket A, B, and C" are not independent, as you're only comparing to one random number.

Imagine I'm rolling a die, for instance, and you have three guesses as to what number I'm going to roll. Naturally, since 3 of those guesses will be correct for some roll of the die, you have a probability 3/6 that one of them is right.

TBF, even if a different number was drawn for each ticket (so that we are speaking of independent events where you multiply instead of add probabilities), it would be hard to notice the difference over the small sample size caused by the weekly lock: the chance of getting three consolations then becomes (0.9 ^ 3) = 0.729, or a little under 73%, vs. 70% of a triple whiff with the single drawn number. Similarly, there is a (5/6 ^ 3) chance you'd fail to guess any of three consecutive die rolls with one guess per roll (0.578, or a little more than 4/7).

It does explain why fourth prizes seem common and the higher ones are so rare though, as assuming you're betting this way, you're down to one potential ticket for the higher prizes.

(It also explains in part why random pick up groups are such a doggone bone of contention in MMOs too: assuming we replace "winning roll" with the chance that someone doesn't have the necessary know how, and considering the multiplicative property of independent events, it doesn't take too many randos in difficult content before the chance that at least one person is making the endeavor moot from the first gets to be frustratingly high, right?)

Quote:

---

Originally Posted by Sove92

If you buy 3 tickets and a different last number for all 3, yes, it does increase it to 30%, but only one of them can win. You have 10 possible last numbers and you have picked 3 of them (3 out of 10). The jumbo cactpot number doesn't change between tickets, it's the same for the whole server.

---

I see what you're saying, but what you're talking about is a 30% of winning 4th prize once. You still have two losing tickets. Technically each ticket is still 1/10. Because the payout for 4th is so low I almost always triple down for jumbo cactpot, but I do get why a player would want to increase their odds for a single prize.

Quote:

---

Originally Posted by Sove92

Yes, the RNG is explicitly rigged to draw a different last number than you specifically have chosen the vast majority of the time. This all despite the fact that the jumbo cactpot number is the same for the whole server.

---

HAHAHAHA!

We all know that's only the case with our retainers and the job gear they bring us.

Quote:

---

Originally Posted by Sove92

All jokes aside, if you pick a different last number for all three tickets, your chances to win the 4th prize is 30%. Given you can only play once a week, a 30% rate isn't that great as even on average, you could only be expected to win the 4th once a month, which is pretty much in line with my experience over many years.

---

*nods*

I guess 30% seems bigger than it really is.

I recently got the luckiest I ever have with Jumbo Cactpot ever have since it was released. I got 3rd place 2 weeks in a row. It probably won't ever happen again, but I'm feeling pretty satisfied about it, lol.

Quote:

---

Originally Posted by JamieRose

I wouldn't worry too much about it. I've been playing the same three numbers for four years and I've never done better than 3rd.

---

If you played every week without fail, that's 52 weeks * 4 years * 3 tickets = 624 chances. Crude math would say you have a 0.1% chance of winning 2nd prize on any given ticket. By ticket number 500, odds are you should have seen a 2nd prize winner. At ticket number 624, you are getting outside a standard deviation for the odds.

I'm more annoyed that I've been playing the Cactpot since it's release 8 or 9 years ago and never once won the Jumbo ever. I figured after 9 years I would have won this silly digital lottery at least once but nope.