Is in-set conversion worth it?

[seebs]

huh.

 

now i'm wondering, because i'm pretty sure i've seen winter's gift way more often than a lot of other sets, and i just assumed that was intentional, but it might just be coincidence.

[Yomo Kimyata]

14 minutes ago, seebs said:

huh.

 

now i'm wondering, because i'm pretty sure i've seen winter's gift way more often than a lot of other sets, and i just assumed that was intentional, but it might just be coincidence.

I don't know how the RNG does the rarity thing.  One logical way would be to look at every rare set that this *could* be converted to, choose one, then assign which enhancement in the set it is.  Another logical way would be to look at every rare enhancement that this could be converted to and choose one.

 

If it's the former, then you might have a (for example, making this up entirely) 1 in 45 chance of getting a winter's gift, but if it's the latter you would have a lesser chance (since winter's gift only has three enhancements in the set).

 

There may also be observer bias.  I tend to notice when I get things that are worthless to me, like winter's gift or snipes, but don't pay as much attention if it's something middling, like a melee damage.

[seebs]

So it turns out, the "successfully converted" messages go in the chat log, so I have ~4600 conversion messages in my chat history. I can't tell which ones were attuned and which were of specific levels, but I can distinguish "by-rarity", "by-set", and "by-type", I believe -- so far as I can tell, by-rarity always picks a thing of a different type, by-type always picks a thing of a different set but the same type, and only in-set can ever pick another thing of the same set.

 

Thus, if you know the type of every enhancement, you can distinguish those categories reliably.

[Yomo Kimyata]

5 hours ago, seebs said:

So it turns out, the "successfully converted" messages go in the chat log, so I have ~4600 conversion messages in my chat history. I can't tell which ones were attuned and which were of specific levels, but I can distinguish "by-rarity", "by-set", and "by-type", I believe -- so far as I can tell, by-rarity always picks a thing of a different type, by-type always picks a thing of a different set but the same type, and only in-set can ever pick another thing of the same set.

 

Thus, if you know the type of every enhancement, you can distinguish those categories reliably.

I cannot guarantee this 100%, but I feel very confident that I have converted by rarity and gotten the same category but in a different set, e.g., converting a Smashing Haymaker by common, getting a Focused Smite.

[seebs]

That is possible, in which case I don't actually have a way to tell them apart. Hmm.

[AboveTheChemist]

On 1/17/2020 at 6:43 AM, Yomo Kimyata said:

I cannot guarantee this 100%, but I feel very confident that I have converted by rarity and gotten the same category but in a different set.

I can guarantee this 100%, from my logs from just the other night (with color code tags removed for readability). I was converting by rarity, with the intent of landing in a type that would yield a guaranteed uncommon-to-rare conversion by type.

 

2020-01-13 06:28:53 Maelstrom's Fury: Damage/Endurance successfully converted to Thunderstrike: Damage/Recharge.

 

Just this week I added code to my log parser to automatically compute the number of converters used, and due to cases like the one above where it is otherwise ambiguous whether the conversion was by rarity or by type, I had to generate a warning if the parser couldn't tell which conversion type had taken place.

 

Two-month edit: I was converting this weekend, had this conversion happen, and thought of this thread. I was converting by rarity and it stayed in-set.

Edited March 25, 2020 by AboveTheChemist
add another example

[seebs]

Hmm. That does make it hard to tell for sure. I suppose there's unambiguous cases (rarity changes, or type changes), but if you reroll from one set to another in the same category *and* rarity, it's ambiguous.

 

[MunkiLord]

I've started tracking my converter use when converting in set and how much I'm making for various sets. Once I have a decent sample size, I'll share. 

 

Also, how many sold enhancements and/or converters used do y'all think it will take before the data is reliable? Couple thousand enhancements, 100k converters, etc?

[Yomo Kimyata]

On 2/3/2020 at 12:36 PM, MunkiLord said:

I've started tracking my converter use when converting in set and how much I'm making for various sets. Once I have a decent sample size, I'll share. 

 

Also, how many sold enhancements and/or converters used do y'all think it will take before the data is reliable? Couple thousand enhancements, 100k converters, etc?

I’d say a few hundred enhancements.

[MunkiLord]

So far since I started tracking I have converted 429, using 6576 converters. For an average of 15.329 to reach a specific enhancement. And that seems about right. I buy my converters at just over 82k each, so that adds an average cost of 1.266 million. I have a ton more to sell, so I'll have more data, including buying, posting, and selling prices for specific enhancements in a few weeks to a month. 

[Yomo Kimyata]

Sounds dead on (assuming a 6-piece set).  A fair RNG would have a mean of 5 conversions (15 converters) and a median of 3 conversions (9).  Out of curiosity, what was the longest non-conversion streak you had?

[MunkiLord]

That I don't know. I convert in batches of 10-80 and track converters used. I know I've hit at 15 attempts, but it's also not uncommon to get multiple in a row, don't recall hitting more than 4 consecutive times.