[outdated] Caustic Arrow Solo Map MFer (20/300+)
|
Chaos Spam Simulator
Spoiler
Step 1 is now done. I've translated the data-mined affix weightings into a series of xml files - probably a little over a 3rd done now with what I'm going to end up writing. Next step is the monte carlo part to randomly generate items and see how many attempts out of say 1 million you get items that meet your threshold. At this stage I'll just make some wild speculations about how likely it is to roll 4 versus 5 versus 6 attribute items. Later this week/next week I'll save up 100c or so and see if my guesses seem to make sense (100 is a pretty small sample but it should serve as a check to see if my guesses aren't totally wrong). Over in the Gameplay Help and Discussion forum, Raics replied to one of my queries about affixes. Apparently the data-mined values posted by poe.db aren't even the values used by the game (which makes total sense - you'd data mine client side stuff and unless the client files are duplicative of stuff on the server side, there's a decent chance that the poe.db stuff isn't trustworthy. Anyhow, poe.db is the best information we have, so it's either use that in the simulator or don't write it in the first place :-). Last edited by hankinsohl#1231 on Jan 27, 2017, 8:29:48 AM
|
|
|
I cant seem to figure out why i die alot my rese is maxed out. Life 4.2k. Chaos damage 20k damage.
stuck on lvl 90 for the last couple of month No change. Lost tons of xp. always evading attacks but still am getting killed. Any one would be able to help me figure out what to do.
Spoiler
  Last edited by max1982#6617 on Jan 27, 2017, 2:44:18 PM
|
|
|
You should have more than 4.2k life at level 90, so something's wonky on either your tree or gear. You should be closer to 4.8k.
Unlock your characters and I'll take a look. (Click your name at the top, Privacy Settings, uncheck "Hide Characters"). Jul 27, 2011 - Sept 30, 2018. Last edited by Serleth#4392 on Jan 27, 2017, 3:03:57 PM
|
|
|
I have unhiden the characters
|
|
|
There's no room in this build for cwdt - immortal call setup.
Part of the reason you're getting murdered is because you don't have Grace, so you're getting hit a LOT more often than you would normally. Drop the setup in your gloves completely other than Flame Golem, add Grace + Clarity. Clarity is so you can drop the mana flask in favour of another Seething Life flask which will dramatically increase your ability to spam out of the situations where you do get into trouble. Adding Clarity will also allow you to drop Perandus Signet for another ring with life (if, and until, you want to add more MF. I'd also drop Distillate. You're not really stacking a lot of MF at the moment, so it won't make a huge difference; replace that for a quicksilver or Stibnite flask). I straight up don't take the bottom jewel socket with Evasion nodes preceding it. It's essentially wasted nodes unless you're running a QoTF build or something. Remove that for more life in the Scion tree. I also don't take the Ranger area proj/proj/jewel socket until Level 93. I'd remove that for the last three life nodes in Scion if you're still not comfortable with the extra 15% from the eva/eva/jewel change, then you can take those nodes again once you level into them. Look into upgrading your helmet. 56 life is fairly low for that slot. I also find Phase Acro incredibly useful (at least for this build, because we don't leech). What I would personally do is remove the eva/eva/jewel for Scion Life, remove proj/proj/jewel for the extra dodge after Acro and spec out of a damage node in Shadow to finish off Phase Acro. Jul 27, 2011 - Sept 30, 2018. Last edited by Serleth#4392 on Jan 27, 2017, 3:48:17 PM
|
|
|
How is this build as a league starter? Looking for something for next league, but I'm wondering if the +3 bow would be an issue for a brand new character in a new league.
|
|
" Not really. I usually end up mostly geared (5L +3 bow, double andy, goldwyrm, etc) within 70 hours playtime, sometimes earlier. I think 6Ls even were like.... 3ex after the first week this league? Very achievable. Jul 27, 2011 - Sept 30, 2018.
|
|
|
Cool, good to know.
|
|
|
Distribution of Prefixes and Suffixes Following Chaos Spam
Spoiler
GGG apparently has not disclosed the exact details about how chaos orbs work other than it randomly rerolls affixes when applied to a rare item. So it's up to the community to work out the details on their own. I've not found any community work on this issue, so, I'm tackling it myself :-) Here's what I've found so far: ================= I spent 100 chaos tonight, spamming my 6-link, iLevel 84 Vaal Regalia in an attempt to get a feel for how often 4, 5 and 6 affixes roll and what the distribution of affixes are in the case of 4 and 5. When using an int chest piece for this task some ambiguity arises when assessing results: Stun and Block Recovery can roll on its own as a suffix or as a combination trait, yielding both increased ES% as well as S/B recovery, as a prefix. Ambiguity arises in some cases because the game does not display the rolled affixes per se, but instead adds up values and displays totals. So if you've rolled a chest that shows an ES% value as well as a B/S recovery value, several possibilities arise: 1) The displayed value represents a single affix: * B/S recov + ES% suffix 2) The displayed value represents 2 affixes which might occur in the following ways: * ES% prefix --- B/S recov + ES% suffix * B/S recov + ES% prefix --- B/S recov suffix * ES% prefix --- B/S recov suffix 3) The displayed value represents 3 affixes: * ES% prefix --- B/S recov + ES% prefix --- B/S recov suffix Usually you can use Haku's crafting table along with the minimum and maximum ranges of these affixes to precisely determine which of the above cases you're looking at. However, I ran into 4 cases which I was unable to disambiguate. I set these 4 cases aside. ================= Considering just the 96 unambiguous outcomes the results were as follows: 6 affixes PPPSS (no open affix): 6 5 affixes PPPSS (no open prefix): 14 5 affixes PPSSS (no open suffix): 12 4 affixes PPSS (open prefix and open suffix): 26 4 affixes PPPS (no open prefix): 21 4 affixes PSSS (no open suffix): 17 Based on this small sample size, my working hypothesis about how the number of affixes is determined is that a random number between one and 16 is rolled. Then: 1-11 = 4 affixes 12-15 = 5 affixes 16 = 6 affixes With 96 cases, the above hypothesis would yield: 11/16 * 96 = 66 expected 4 affix cases versus 64 observed 4/16 * 96 = 24 expected 5 affix cases versus 26 observed 1/16 * 96 = 6 expected 6 affix cases versus 6 observed The hypothesis fits the data well. Given the low number of trials of course the hypothesis could be wrong. =========== As for the distribution of affixes in the case of 4-affix results, the observed distribution loosely fits the coin-flip hypothesis I made in an earlier post. ============ So, I'll go with the above in writing my chaos spam simulator. Empirically I'll gather more data too to help boost confidence in the hypotheses or to replace them with models that better fit observed data. ============= Side note - 200 chaos spent now on my chest (didn't track the first 100) and no chest upgrade yet, LOL. Last edited by hankinsohl#1231 on Jan 28, 2017, 9:12:33 AM
|
|
" Many thanks for you help Radagastt. Your the best. I have swaped some of the gear for higher hp. now i have 4.7 hp with 20k damaged. 1 last thing which maps you would recommend for lvling fast as i have lost 58% of xp again lol |
|
