I didn't see a topic around for it yet, so I figured I would start one.
Rules are simple: Rate the monk above you on a scale from 1 to 10 while offering constructive criticism. Please be polite and, of course, make sure you add a link to your own monk and a brief explanation if necessary after you're done.
I [used to] run with 275% MF before stacks. I could farm about 45-50 rares per hour with a sample size of about a dozen hours. This was act 1 mind you, same route.
I'm currently running with a tad less so that I can clear act 2 with my farm gear...
If you use the AH, you're going to be predisposed to gear that is really good. Truth is, there is a lot of "good" gear that drops. It's unlikely that it will be good for you, but probably good for someone else. What do you do? You put it up on the AH for some odd amount.
The idea here is that you're basically taking the combined farm time of a lot of different people to get you gear. For you to find a good piece of gear, not only does it have to drop, but it has to drop for you (not your potential farmers). If you decided that you weren't going to use the AH at all, you would have much worse gear and you would be finding gear at about the same rate.
I started a similar exercise (isn't having something really boring to avoid at work great!) trying to reverse engineer the "item roll" process and noted similar trends.
I was trying to work out how far down each affix type an item would roll - the floor on most seem to be level 51/52 affixes (with the exception of pickup radius).
Then I tried to work out if it was 1 roll or 2 roll system - all possible affixes thrown in a big pool and rolled or roll "group", then roll within the group. Its clearly a lot more complex than that as you note there is clearly a level of weighting towards the highest affix in any given group.
At this point I was pretty much beyond my analytical capability so I'd be interested in suggestions on how to move forward (with the goal being to be able to work out the proportion of items with a given set of stats or better - a "rarity" score as it were).
I'm not really a statistician, but I think the graphs do support the premise that there are a few different stackable affixes a chest (or any piece of gear) can roll.
It lists several different combinations between Vitality and Dexterity rolls.
If you roll the best combination of affixes you can between Vitality and Dexterity, you come up with a maximum possible 300 Vit and 200 Dex: http://d3db.com/item...5994,3561578828
I believe those statistics were datamined from the game, but I could be wrong. Assuming we believe them, there is probably a relatively even distribution of possible affix combinations between each of them.
I'm willing to bet that the game randomly decides what affixes the item has based on a simple "pick X affix combos and roll the stats for each applicable combo" I'm also willing to bet the game won't allow you to roll more than 2 different combinations of any given stat (no more than 2 vit rolls, 2 dex rolls, etc.) and probably has a set probability of rolling each combination.
Best possible roll for a Vit chest would be a combination of 2 different Vitality rolls including the "pure" vitality roll (170-200) as well as a vitality/primary stat roll.
This is extremely important as it implies that a magic item can't have more than 2 different affix rolls, but can result in 3 different stats. If you can find a magic chest with more than 200 vitality AND more than 100 of any given primary stat, it would invalidate that claim as doing so would actually require more than 2 different affix rolls.
***EDIT***
And just looking around on the AH, I was unable to find any Magic items that had those stats. I suppose the correct word for these "rolls" is a property which can be a combination of different affixes.
We can confirm this hypothesis with the affix quantity distribution curve with relation to primary stats on the chests that Tybrone crafted as it simply says that it's much more likely to only roll a single set of vitality affixes than it is to roll double vitality affixes. Given the minimum and maximum possible ranges for those affixes, we can identify a "deadzone" so-to-speak of a range of vitality (or any stat really) that simply cannot occur on a given piece with given ilvl.
What would truly be interesting is identifying what affixes were rolled on a piece of gear. I think we could do so if we ran through each piece of gear and basically just confirmed or denied that a given combination of affixes is possible for that piece of gear. It is only then that we could determine what the probability of rolling a particular property is and then coming up with a distribution curve for a particular affix.
From there, it's just a matter of figuring out the probability of finding that gear and, assuming it's a good combination of properties, valuing it accordingly.
****EDIT #2****
Figuring out what properties a piece of armor had would largely be trial and error but mostly process of elimination. It doesn't look like that website I linked is entirely accurate... but regardless, it's a lot to work with.
I think I'm most interested in the affix combinations and limitations. I'm loving this data.
The most interesting things I've noted are the probabilities with stat rolls. For example: I never realized that the resists were a positively skewed bell curve (it's very very likely to roll high resists as opposed to rolling low resists) and I didn't realize that there were two different affix strengths that the primary stats fall into.
Maybe this info will help you in identifying the rarest of your chests:
Keep up the good work and give me more data to work with!
I would love access to the Helm spreadsheet you've been keeping. Depending on the rolls, I might be interested in an upgrade to my own..
I hate to be pedantic, but if a probability distribution has its bulk toward higher values, then it has negative skewness, *not* positive skewness. It is also said to be negatively skewed or left skewed.
Doh... I suppose it's been a while since I studied actual statistics. So yes. Negatively skewed.
Just sent an access request for that spreadsheet. I'll start looking into trends and such. I'm not necessarily interested in what combinations of affixes happen, just the affix probabilities themselves.
I know everyone and their mother has a DPS calculator, but what I didn't like was how a lot of them weren't in spreadsheet form! I love spreadsheets and find them easier to use than many other tools so I just had to make my own item comparison that I thought was simple and easy to use.
If you like it, let me know!
Be sure to copy this spreadsheet (File->Make a copy) for your own use so you can fill it in with your character's stats.
I think I'm most interested in the affix combinations and limitations. I'm loving this data.
The most interesting things I've noted are the probabilities with stat rolls. For example: I never realized that the resists were a positively skewed bell curve (it's very very likely to roll high resists as opposed to rolling low resists) and I didn't realize that there were two different affix strengths that the primary stats fall into.
Maybe this info will help you in identifying the rarest of your chests:
Awww man... All other classes are jealous of that DH calculator.
I've been using a really nice Monk spreadsheet that calculates damage mitigation as well. The formula for DPS is actually pretty simple which is why you're seeing a lot of tools out there that utilize it.
1
Rules are simple: Rate the monk above you on a scale from 1 to 10 while offering constructive criticism. Please be polite and, of course, make sure you add a link to your own monk and a brief explanation if necessary after you're done.
I'll go first!
http://us.battle.net/d3/en/profile/Azraeil-1945/hero/3438373
0
I'm currently running with a tad less so that I can clear act 2 with my farm gear...
0
If you use the AH, you're going to be predisposed to gear that is really good. Truth is, there is a lot of "good" gear that drops. It's unlikely that it will be good for you, but probably good for someone else. What do you do? You put it up on the AH for some odd amount.
The idea here is that you're basically taking the combined farm time of a lot of different people to get you gear. For you to find a good piece of gear, not only does it have to drop, but it has to drop for you (not your potential farmers). If you decided that you weren't going to use the AH at all, you would have much worse gear and you would be finding gear at about the same rate.
0
I'm not really a statistician, but I think the graphs do support the premise that there are a few different stackable affixes a chest (or any piece of gear) can roll.
Take this page for example: http://d3db.com/item/i/archon-armor
It lists several different combinations between Vitality and Dexterity rolls.
If you roll the best combination of affixes you can between Vitality and Dexterity, you come up with a maximum possible 300 Vit and 200 Dex: http://d3db.com/item...5994,3561578828
I believe those statistics were datamined from the game, but I could be wrong. Assuming we believe them, there is probably a relatively even distribution of possible affix combinations between each of them.
I'm willing to bet that the game randomly decides what affixes the item has based on a simple "pick X affix combos and roll the stats for each applicable combo" I'm also willing to bet the game won't allow you to roll more than 2 different combinations of any given stat (no more than 2 vit rolls, 2 dex rolls, etc.) and probably has a set probability of rolling each combination.
Best possible roll for a Vit chest would be a combination of 2 different Vitality rolls including the "pure" vitality roll (170-200) as well as a vitality/primary stat roll.
This is extremely important as it implies that a magic item can't have more than 2 different affix rolls, but can result in 3 different stats. If you can find a magic chest with more than 200 vitality AND more than 100 of any given primary stat, it would invalidate that claim as doing so would actually require more than 2 different affix rolls.
***EDIT***
And just looking around on the AH, I was unable to find any Magic items that had those stats. I suppose the correct word for these "rolls" is a property which can be a combination of different affixes.
We can confirm this hypothesis with the affix quantity distribution curve with relation to primary stats on the chests that Tybrone crafted as it simply says that it's much more likely to only roll a single set of vitality affixes than it is to roll double vitality affixes. Given the minimum and maximum possible ranges for those affixes, we can identify a "deadzone" so-to-speak of a range of vitality (or any stat really) that simply cannot occur on a given piece with given ilvl.
What would truly be interesting is identifying what affixes were rolled on a piece of gear. I think we could do so if we ran through each piece of gear and basically just confirmed or denied that a given combination of affixes is possible for that piece of gear. It is only then that we could determine what the probability of rolling a particular property is and then coming up with a distribution curve for a particular affix.
From there, it's just a matter of figuring out the probability of finding that gear and, assuming it's a good combination of properties, valuing it accordingly.
****EDIT #2****
Figuring out what properties a piece of armor had would largely be trial and error but mostly process of elimination. It doesn't look like that website I linked is entirely accurate... but regardless, it's a lot to work with.
0
Doh... I suppose it's been a while since I studied actual statistics. So yes. Negatively skewed.
0
0
If you like it, let me know!
Be sure to copy this spreadsheet (File->Make a copy) for your own use so you can fill it in with your character's stats.
https://docs.google....SjRac29udlFiRlE
If you have any questions on it's use, I'll be happy to help.
0
The most interesting things I've noted are the probabilities with stat rolls. For example: I never realized that the resists were a positively skewed bell curve (it's very very likely to roll high resists as opposed to rolling low resists) and I didn't realize that there were two different affix strengths that the primary stats fall into.
Maybe this info will help you in identifying the rarest of your chests:
https://docs.google....bEU0ZWgyX2NJSmc
Keep up the good work and give me more data to work with!
I would love access to the Helm spreadsheet you've been keeping. Depending on the rolls, I might be interested in an upgrade to my own..
0
0
I've been using a really nice Monk spreadsheet that calculates damage mitigation as well. The formula for DPS is actually pretty simple which is why you're seeing a lot of tools out there that utilize it.