Enchantments are Evenly Weighted. Why?

"

MadRabbitPoE wrote:

There you go, guys. Math. It's got numbers and it's a bitch

(And holy shit, man. How have I lived this long and still not known about this website? That's life changing, man. Thanks for that. I can now stop relying so heavily on analogies and metaphors to obfuscate my laziness with actually doing the math.)

"

Wruntjunior wrote:

"

grepman wrote:

you arent meant to brute force an enchanment for your skill easily. the enchantments are a quirky perc that may facilitate things like unique with enchantments trade/white base trade/auxilary skill trade. its neither mandatory nor meant to be easily achievable

I dont understand that quip about odds. odds are only meaningful in context. especially in card table analogy having, the EV is all that matters. if you have 1 to 218 odds to win 1000 dollars and you invest only 1 dollar, if you fold, you are doing a very dumb thing since you EV is almost +800 dollars

Actually, your EV isn't 800 dollars (you can't just subtract it like that, probability is not a guarantee of success after x number of tries) - you have an estimated 99.54% chance of failure. At 200 bets, you only have a 60.13% chance of a single win. To have even a 90% chance of success (literally 90.0087%), you'd have play 501 times. At this point, it's technically worth it if you fall into the 90th percentile (by definition, you're very likely to do so), and you'd double your money - roughly $500 value.

Taking this back to the same odds with mods, if you wanted to be practically guaranteed to get your mod at least once (>95%, or >99%, I'll give both), you'd be looking at 652 runs (95.0099%) or 1002 runs (99.0017%). Even at 1002 runs, there's a 1% chance you won't have seen the mod you're after at least once, and that percentage never hits 100% because the formula always forms an asymptote - it just gets increasingly unlikely with every run.

Source: WolframAlpha for quick math, and Geometric Distrubition for the formula (cumulative distribution function, 1 - (1 - p)^k, where p is probability and k is number of trials)

"

Wruntjunior wrote:

"

MadRabbitPoE wrote:

There you go, guys. Math. It's got numbers and it's a bitch

(And holy shit, man. How have I lived this long and still not known about this website? That's life changing, man. Thanks for that. I can now stop relying so heavily on analogies and metaphors to obfuscate my laziness with actually doing the math.)

I literally do the math as an excuse to use WolframAlpha at times...as a programmer, it's magical how well it works. That thing runs on farkin' magic.

And yeah, as I started running the numbers just there, I feel worse and worse about actually trying to farm the helmet mod myself. Easier to get a decent helmet (like rat's nest or abyssus) and get any of several good enchants, then sell it.

1% is a very high value for not seeing something after (if 10 minute runs) ALMOST A WEEK (167 hours) of running the same thing.

"

MadRabbitPoE wrote:

"

grepman wrote:

you dont have to cling to numbers I provided.

30 minutes of your time is your investment, but how much is a perfect helmet with perfect enchantment for your build worth to you ? or to a buyer ?

a rat's nest with a perfect attack-based enchantment for your build will be worth a lot, several exalts. how much is an exalt worth to you time-wise ? an hour ? two hours ?

you running the lab 50 times while enchanting a rat's nest for a really good enchant will most likely pay off. eventually your investment will be like 10 mins because you'll get better understanding of the lab and run through it. if you like economy simulator, you could even buy up rats nests and brute force them, reselling them with good enchants, making a big profit over long run. that possibility is also factored in the odds.

I am not clinging to any numbers. I'm really well aware of the circumstances around running for enchantments and times speed runners have acquired. I'm still stating I think a pool of very specific skill enchantments amounting to a total of 218 with the bulk of percentage beings ones I can't effectively use without a completely different character is too much for evenly weighted odds.

My player experience would be better if the odds were better in a way was relevant to my unique character. I would be willing to take better odds if they were balanced around less powerful enchantments or an inability to sell them on the market.

If it helps to understand where this perspective comes from, I don't buy off the market. I am purely self found.

"

grepman wrote:

you're right,
I simplified it for a purpose of understanding because you would have to bring an arbitrary # of trials in. we dont want to be discussing poisson distributions here do we ? :p

the only thing that matters is that your EV is positive for that example. you want positive EV over long run.

"

MadRabbitPoE wrote:

"

Wruntjunior wrote:

"

MadRabbitPoE wrote:

There you go, guys. Math. It's got numbers and it's a bitch

(And holy shit, man. How have I lived this long and still not known about this website? That's life changing, man. Thanks for that. I can now stop relying so heavily on analogies and metaphors to obfuscate my laziness with actually doing the math.)

I literally do the math as an excuse to use WolframAlpha at times...as a programmer, it's magical how well it works. That thing runs on farkin' magic.

And yeah, as I started running the numbers just there, I feel worse and worse about actually trying to farm the helmet mod myself. Easier to get a decent helmet (like rat's nest or abyssus) and get any of several good enchants, then sell it.

1% is a very high value for not seeing something after (if 10 minute runs) ALMOST A WEEK (167 hours) of running the same thing.

I am a programmer and developer myself and you first colleague I have ever encountered who has even mentioned this website. It's blowing my mind right now. I've always done stuff like trig by hand with old textbooks.

And that feeling is why I am posing this suggestion. The actual numbers come out to be so daunting that you lose all motivation to even try.

"

grepman wrote:

in other words, you would like to invest as least time as possible to find a useful enchant for your specific character (not even characterS). isnt that the opposite of what GGG wants ?

"

grepman wrote:

"

Wruntjunior wrote:

"

grepman wrote:

you arent meant to brute force an enchanment for your skill easily. the enchantments are a quirky perc that may facilitate things like unique with enchantments trade/white base trade/auxilary skill trade. its neither mandatory nor meant to be easily achievable

I dont understand that quip about odds. odds are only meaningful in context. especially in card table analogy having, the EV is all that matters. if you have 1 to 218 odds to win 1000 dollars and you invest only 1 dollar, if you fold, you are doing a very dumb thing since you EV is almost +800 dollars

Actually, your EV isn't 800 dollars (you can't just subtract it like that, probability is not a guarantee of success after x number of tries) - you have an estimated 99.54% chance of failure. At 200 bets, you only have a 60.13% chance of a single win. To have even a 90% chance of success (literally 90.0087%), you'd have play 501 times. At this point, it's technically worth it if you fall into the 90th percentile (by definition, you're very likely to do so), and you'd double your money - roughly $500 value.

Taking this back to the same odds with mods, if you wanted to be practically guaranteed to get your mod at least once (>95%, or >99%, I'll give both), you'd be looking at 652 runs (95.0099%) or 1002 runs (99.0017%). Even at 1002 runs, there's a 1% chance you won't have seen the mod you're after at least once, and that percentage never hits 100% because the formula always forms an asymptote - it just gets increasingly unlikely with every run.

Source: WolframAlpha for quick math, and Geometric Distrubition for the formula (cumulative distribution function, 1 - (1 - p)^k, where p is probability and k is number of trials)

you're right,
I simplified it for a purpose of understanding because you would have to bring an arbitrary # of trials in. we dont want to be discussing poisson distributions here do we ? :p

the only thing that matters is that your EV is positive for that example. you want positive EV over long run.

"

grepman wrote:

you should try MATLAB. anything you want to do math-wise, it can. not free of course, but hey.

and, the game is about grind and based around trading. 167 hours is nothing in the game which is played by many for thousands of hours

"

MadRabbitPoE wrote:

"

grepman wrote:

you're right,
I simplified it for a purpose of understanding because you would have to bring an arbitrary # of trials in. we dont want to be discussing poisson distributions here do we ? :p

the only thing that matters is that your EV is positive for that example. you want positive EV over long run.

In your example, yes, that's what matter, but it's not all that matters in the context of this discussion. The example was using dollar amounts which have objective value. The input for this calculation is time invested per run and the output is an enhanced item in a video game. The value of both of those are subjective to individuals and for me personally, the investment is too high for the reward.