Anyone want to teach a class on UO math?

[sablestorm]

So I here imbuing a suit for training and adding this second property, the gump tells me I have about a 40% chance of success. Why is, then, that I get four or five failures before I actually succeed, not to mention the last piece on this suit where I've failed 10 times already and need to go buy more gems? In regular plain old math, my best success rate is 20%.

Now I know I probably just hit one of UO's legendary bad streaks, but when you hit one of those streaks, you just want to go complain somewhere!

 

[DHMagicMan_1]

So I here imbuing a suit for training and adding this second property, the gump tells me I have about a 40% chance of success. Why is, then, that I get four or five failures before I actually succeed, not to mention the last piece on this suit where I've failed 10 times already and need to go buy more gems? In regular plain old math, my best success rate is 20%.

Now I know I probably just hit one of UO's legendary bad streaks, but when you hit one of those streaks, you just want to go complain somewhere!

More then likely, no one fed the rats running the RNG so they are throwing out as many failures as they can... temper tantrum time.

 

[Pinco]

40% doesnt mean that you have 4 chance of success every 10 tries.
You can fail 100 times if you are not lucky, there are no warranties

 

[EnigmaMaitreya]

A common thing people do, they think that a failure changes the chance (even in real life). It doesn't.

Every time you try, you have exactly the same chance each time, unless the system were weighted to accommodate your failure, which in this case it is not.

In short, Probability does not equal Chance.

 

[Maplestone]

Even if RNG was working perfectly, if you assume that a streak of sufficiently bad luck with motivate someone to rant on stratics, then statistically, given a large population of stratics readers, we expect to see rants about runs of bad luck that, looked at in isolation, seem impossibly unlikely.

The chance of failing on your next 10 straight attempts when there is a 40% chance of success is: 0.6%. So we expect this to happen once every 165 readers. There are 209 people currently reading U.Hall, so I expect at least one person here to have the same run of bad luck next time they log in

 

[Smokin]

Ever buy scratch tickets that say 1 in 4 chance of winning, can buy 10 and still lose on all of them.

 

[Uvtha]

I hear tell, that if you flip a quarter irl, 100 times, its will always be 50 heads and 50 tails!

 

[EnigmaMaitreya]

I hear tell, that if you flip a quarter irl, 100 times, its will always be 50 heads and 50 tails!

Go ahead and do it, it is not guaranteed to be the missing side .... because even if you were 50 of one and 49 of the other, the last flip still has the same 50/50 chance as the previous 99 flips had.

You are NOT guaranteed which one it will be.

And should we wager on it, we each have a 50/50 chance of winning.

 

[Doubleplay]

Well, I quickly forget the "impossible" streaks of good luck, but will never forget when I failed three times in a row with a 98.5% success chance a couple weeks ago. Naturally, it was when using my precious relic frags.

 

[EnigmaMaitreya]

Well, I quickly forget the "impossible" streaks of good luck, but will never forget when I failed three times in a row with a 98.5% success chance a couple weeks ago. Naturally, it was when using my precious relic frags.

Ever play a card game called Ace Ducey (aka in between), get Ace 2 bet pot, draw a 2, next time get a King 2, draw a King, next time get an Ace 2, draw an Ace. Now that is bad luck, but oddly enough the next time was a Queen 4 and I bet the pot and won.

 

[Basara]

Actually, folks, probability has nothing to do with it.

The devs have already stated that the RNG is prone to streaks of similiar values, and that they were looking to correct the issue.

Since the streaks can occur at any point in the distribution, the overall probability numbers stay the same, but it means that small samples are much more likely to be skewed from what standard probability calculations would indicate.

A way to look at it would be an unfair six-sided die (As opposed to a fair one), where there was an additional modifier that made it 20% likely that any number rolled would be repeated in the next roll, instead of being a fair die roll. As this could occur to ALL 6 possible results, the overall results would appear to be a perfectly normal probability curve - but the actual data, if examined closely. would show that the chance of two consecutive rolls being identical to be much greater than the expected 1 in 6 chance, in a manner that would indicate some sort of external bias.

To use another gaming analogy, since EM brought up Acey Ducey, imagine if you were shooting craps, and there was something that altered all die rolls in the same manner as my example above. The Probabilities would balance out in the end, but given that one wins by rolling sevens, or rolling the same value that you previously rolled before hitting a seven, you'd end up with house seeing lots of big losses, if people played intelligently, and didn't let the pot ride too long.

 

[Maplestone]

Actually, folks, probability has nothing to do with it.

By my point is that even if UO had a perfect randomizer, we'd still see the exact same kind of posts

 

[Stupid Miner]

So I here imbuing a suit for training and adding this second property, the gump tells me I have about a 40% chance of success. Why is, then, that I get four or five failures before I actually succeed, not to mention the last piece on this suit where I've failed 10 times already and need to go buy more gems? In regular plain old math, my best success rate is 20%.

Now I know I probably just hit one of UO's legendary bad streaks, but when you hit one of those streaks, you just want to go complain somewhere!

No, it's not a bug, they actually put a "sablestorm" exception in the RNG code. The devs just dislike you that much.

... but seriously, if you think it's bugging on you do something else. recall somewhere, log-in/log-out, craft another item, or whatever.

 

[Poo]

UO Math 101


1 + 1 = 14.2 (3.14*223(+ or - the numbers barding difficulty rounded to the nearest 1/10,000 - 1))


end of lesson.

 

[Ni-]

UO Math 101


1 + 1 = 14.2 (3.14*223(+ or - the numbers barding difficulty rounded to the nearest 1/10,000 - 1))


end of lesson.

mmmm, pie...

 

[yorky]

UO Math 101


1 + 1 = 14.2 (3.14*223(+ or - the numbers barding difficulty rounded to the nearest 1/10,000 - 1))


end of lesson.

You forgot the 0.6 variance of latency.

 

[Frarc]

Sosaria's math is not like Earth math. You should know now after 12 years.

 

[Jermosh]

RNGs can do wierd things, there was a study on monitored RNGs accross the world that all started to flip to the same side during 911. I think it was Columbia that did it.

 

[Basara]

Also consider that in some of the UO code, 2/3 is actually "2/3" in the code, "0.66" in others (and I'm sure there are other repeaters similarly truncated), causing oddities in the math

The above example, is mining as it affects blacksmith recycling. Its ingot return rate is 0.66 times the Mining skill (as a percentile), so that GM mining only returns 66.0% instead of the intended 2/3, and one has to have 101.1 Mining, via the Studded Leather or Ringmail Mining Gloves, to get the full, intended ingot return.

 

[BartofCats]

With everything in this thread, 40% chance of success is also 60% chance of failure. PER CHANCE. Like with that coin being heads or tails, you have 40% chance of calling it, and 60% chance of it being the other. Believe me, I have been where you are, 40% chance of imbuing something and 10 tries later it took.

 

[Zyon Rockler]

I think the problem with an RNG is that it tries to predict what the outcome should be without knowing all the factors or possibilites.

Like, when you flip a coin, there's a chance it could be heads or tails but it could also land on its' side. The problem is that as a human we wait for it to fall and then decide that it is either heads or tails and lose the possibility of its' landing on its' side.

If you hold something in your hand and you let go of it, its' most likely going to fall to the ground but of course there's a possibility of it going up, as it could be a helium balloon or you could be in space.

The intereting word of random is being used but when dealing with a computer system its' basic analog or language is numbers.

This means that it's not really a random machine but in actuality it's predicting the highest possibility. If you program 1+1 it should be 2, so when you ask it to make that a random number, your actually telling it what random number you want it to produce. This is how it becomes more of a prediction.

For example: To predict what i'm going to say next you would need to know all of the possibilities, once you do then you know because you have all of the outcomes in a data base. Understanding the variables, for example: what someone's favorite color is or photograph that they like. You would be able to pinpoint an exact reaction.

If we are to assume there will be an outcome at all and if we can convert everything into a number then it's no longer randomized, it's catagorized.

 

[Maplestone]

*looks at Zyon as if he had just grown a second head and flippers*