If i had more than a fragment of coding knowledge, id make a bot that kept a party member behind the target, the thf behind that party member, and only use SA and TA, which would generate much better proc rate data.
Unfortunately I dont and I can only work with whats available and within a tolerable risk threshold.
Treasure Hunter Proc Rate Testing -TH 8 Versus 14
|
Treasure hunter proc rate testing -TH 8 versus 14
You're being extremely aggressive about this, but fact of the matter is your data isn't any more extensive or targeted than the data in the OP of this thread from 2021. It's less rigorous, and has a smaller sample. This isn't being said to insult you, simply to point out that the same reasons that data set was inadequate to establish a conclusion still apply to yours.
To measure the outcome in a way that will stand up to scrutiny requires measuring the variable that matters; chance of TH proc at a given TH level. Since TH proc only happens on first attack of a round, you need 3 pieces of data on each attack round: -Current level of TH on mob -Current level of TH equipment -Whether a proc occured on that attack round A sample that does not cover these data points, and is not measured in number of attack rounds, is at best a correlation to the original post's data.
[+]
If I used windower, I'd be more than happy to make something for you to measure with, but it's not worth learning their API for. I explained the layout of the packet in question on page 2, and chiaia said people in windower discord might help.
I'll try to get something set up myself, but I have a lot going on for the holidays so not the best time atm. Im still trying to get EP at a moderately decent rate, not spend 5 minutes poking an Apex mob down ¯\_(ツ)_/¯
Isnt the whole point of a larger sample size to balance out all those noisy variables? I mean, were talking something that is already something like 2% chance to begin with normal attack rounds, and it only gets worse when player TH is lower than mob TH level. 45% of fights yielded zero TH procs, and when I was checking the fight times for the TH 11 procs, I noticed a back to back TH9 then 10. Its already very low distribution to begin with, I dont think those variables youre talking about are affecting the data in any significant matter. Carbuncle.Nynja said: » Isnt the whole point of a larger sample size to balance out all those noisy variables? Quote: I mean, were talking something that is already something like 2% chance to begin with normal attack rounds, and it only gets worse when player TH is lower than mob TH level. Carbuncle.Nynja said: » when I was checking the fight times for the TH 11 procs, I noticed a back to back TH9 then 10. Shiva.Thorny said: » If I used windower, I'd be more than happy to make something for you to measure with, but it's not worth learning their API for. I explained the layout of the packet in question on page 2, and chiaia said people in windower discord might help. I'll try to get something set up myself, but I have a lot going on for the holidays so not the best time atm. Does that mean you have something that will capture and log the relevant data for Ashita? I'm not currently subbed, but free login is about to start and I don't mind generating some logs for people to math on. Means I could easily write something for Ashita, but not Windower. If I did, I'd just run the tests myself though. Like I said, it's not the best time of year for free time for me, but I'll try to get something set up this week.
1 damage weapons on uragnites would allow you to make absurdly large samples I think.
my curiosity is peaked, and I happen to be doing TH building activities at the moment.
Seems like it would be fairly straight forward programatically to do this.
I'm gonna concede that the methodology I used is flawed and would need a massive sample size flooding to *** the data being accurate, but it has nothing to do with counting attack rounds or what not. The problem the system itself: If the capped proc rate is when your TH is greater than or equal to the TH on the monster, then only procs of TH10+ are relevant. The TH9 proc doesnt matter because TH8 and TH13 sets should yield similar results. Trying to test this while EPing wont work as mobs dying in 30 seconds just isnt enough time. As much as I have "no desire to poke mobs for 3+ minutes per mob", the sample size needed based on 35 second kills where 45% of the time they didnt even hit TH9 would be massive. I still think that the proc rate isnt capped to TH8 based on the TH10 proc numbers, but after coming to the realization of most of the data is useless, I'm no longer willing to say it with all my chest.
Having said that, I do think a decent 300+ kill sample size of poking things down in a somewhat timely manner (it cant live long enough where the TH8 set routinely hits TH13+ for example) in both a TH8 and TH13 set would be more than sufficient to note if the proc rate delta is affected by the TH8 cap. If the proc rate cap is 2% for unbuffed attack rounds when your TH is greater than or equal to the mob TH, and its halved for every level under the mob TH (the amount it is cut is unknown, it could just as easily be an exponential number), then the TH12/13/14's should start to really stand out. Made up numbers incoming: If it takes 50 attack rounds per TH proc on average (based on 2%), the TH8 set should take 750 attack rounds on average just to hit TH12 (50+100+200+400, every level gets halved). The TH13 set, in 750 attack rounds, should be proccing TH14 every mob. Cerberus.Shadowmeld said: » 1 damage weapons on uragnites would allow you to make absurdly large samples I think. Carbuncle.Nynja said: » I still think that the proc rate isnt capped to TH8, but after coming to the realization of most of the data is useless irrelevant, I'm no longer willing to say it with all my chest. Pretty much where I'm at. I'm thinking tomorrow I'll set up something with a few chars (RUN tank disengaged, GEO haste cap, THF builds TH until 14 with just autos then RUN kills it). Pretty sure I can pull TH21, will have to double check.
[+]
Carbuncle.Nynja said: » Cerberus.Shadowmeld said: » 1 damage weapons on uragnites would allow you to make absurdly large samples I think. Once it gets to 14 you kill or otherwise swap to another mob. It's a mob that should consistently live to get all the way to TH14 though, so a few dozen (or hundred) would give to a good sample TH upgrade rate. I see someone bumped my old thread. If someone who knows how to design windower addons cares to take a crack at developing something to track this data, then I'd be all for using it. While the information I collected was a starting point, the sample size was small. It takes entirely too much energy to try to collect a reasonable sample size from one person. Someone mentioned in the thread that ""he hard part should be in creating the data collection method, not the data collection itself". I agree with this. Look at what was suggested for metrics. If someone creates the addon, I'll be happy to use it to help gather data.
Carbuncle.Nynja said: » We already know that the proc rate goes down when your equipped TH is below the mob TH level. Last I checked (here) this wasn't "known" care to elaborate? https://www.bg-wiki.com/ffxi/Treasure_Hunter
Quote: Upgrading TH levels from normal attack rounds is inversely proportional to the difference between your current TH level in trait/gear and the monster's current TH level[2]. Example 1: If you plant TH6 on the monster by pulling with your TH gear on and then continue to use your TH gear while you kill the monster, there is a 0-level gap between your current TH level and the monster's current TH level. This makes it "easy" to proc (approximately a 6% chance per melee round). Example 2: If you plant TH6 on the monster by pulling with your TH gear on and then swap to TH3 to kill the monster, there is a 3-level gap between your current TH level and the monster's current TH level. This makes it comparatively more difficult to proc (approximately a 2% chance per melee round). links to data is there
[+]
https://github.com/ThornyFFXI/THTest
Threw this together, not pressing anyone to change to ashita or test but figure this gives auditability for any future results in the off chance anything is wrong, and gives other people opportunity to do it if they're so inclined. Still plan to set up at least one good test with it today. Ninja edited it shortly after putting it up to add columns, currently has: Code "Timestamp","Mob TH Level","Player TH Level","Proc Index","Proc Crit","First Hit Crit","First Hit Land","First Hit Damage" If anything else seems important, let me know.
[+]
Maybe a test on a locus mob that heals in bibiki bay that won't die would be a good test target (they heal). See how many melee rounds it takes to get th 14 each time per TH level?
doing apex crabs with a single wielded centovente and th15 atm, plenty of hp to get 14 every mob, just takes a while to build up meaningful data
Honest question: Why though?
Are you gonna hyper-optimize your sets to account for the level of TH the mob has on them? That seems rather excessive, even to the people who hyper-optimize for every scenario. If you're on THF and that concerned about DPS, you'd most likely be TH-Tagging and reverting to your DPS set, considering mobs wouldnt even be alive long enough to see a TH9 proc. Do you not believe that the proc rate has a cap reached when your TH is >= mob TH? You dont need to smack something to TH14 every time to *** this though. If you have TH11 in gear and it takes ~17 rounds to hit TH9, another ~17 rounds to hit TH10, another ~17 rounds to hit TH11, then its established that there is a cap on proc rate. Additionally, you would still need a large sample size to *** that information because sometimes you never hit that 6% chance, and sometimes you'll hit that 6% chance back to back. In the time you smack around a mob 200 times to get TH14 data, you can smack around 3x as many mobs to get better data to determine if the proc rate cap is affected by the TH8 cap. Now having said that, its also possible they lowered the proc rate cap for the higher TH levels, that would be a very SE thing to do. isolating for the data point in question is what makes it high quality data, and i don't need to be botting master levels every free minute because i'm not obsessed with progress
if the proc rate is based on delta from your TH level to the level on the mob, the rate differences will be much more apparent at TH12+ than they would be at TH8.. especially if there's a cap on rate benefit at breakeven Sorry I shoulda quoted Ahlen. My reply took too long.
I do not use ashita so this is probably a newbie error but when I try to load thtest I get an error
Code Error while trying to load addon thtest. Error: \addons\thtest\thtest.lua:1: attempt to index global 'addon' (a nil value) Also, in the readme it says to edit line 27 but I assume that line 34 Code local currentPlayerTH = 15; you're on ashita3, this is for ashita4
fixed readme, that is correct i'm hoping someone can port this to windower, or make something comparable.. the actual logic isn't that complicated, i just have no idea how you filter out double sent packets in windower because i'm pretty sure you don't get sequence ids at all Got Ashita v4 working but the addon is throwing a new error
Code failed to execute task within addon 'thtest'. Error: \addon\thtest\\thtest.lua:9: module 'gdifonts.inclue' not found: it then proceeds to look for in a bunch of locations that it obviously doesnt exist at. I assume I need to get this from your github ashit v4 addons? it is in the repo i linked, you can download as zip to get everything
i'd rather not turn this thread into basic tech support, if you need more help(or anyone else does) please join the ashita discord I understand, I will note I had to put the gdifonts folder inside the thtest folder to get thtest to load.
Yea, if you downloaded as zip it should all have been in the same folder, you copy it all to the addon folder and load like any other addon on ashita/windower.
I think the "why" is simply because people can?
In the absence of SE actually coming right out and explaining every aspect of the TH mechanic, people will always experiment and try to figure it out. If I had to go out on a limb though, I'd say do what makes you happy. The difference in drop rates between TH 8 and 15 is pretty insignificant for common drops, with its value increasing the more rare the drop is, but even then, we're not talking about improving a ultra rare drop to 50%, we're talking a bump from .8% to 1.5% I do know there's been past discussions in my ls about this mechanic though, and so people will always have questions until someone either determines a formula, or the devs actually explain what happen\s. From my understanding, the level difference between your equipped TH and the mob's current TH level is what effects the proc rate. I believe the number most used is 6% improving the TH rate to 9 per attack on a mob when at level 8 TH. I've been told that each level of TH you have above 9 increases that % by 1, and I 've also been told that anything above the current TH level +1 actually reduces the proc rate, so that's very inconclusive. Each increase in TH resets the value at level +1 being 6% so naturally with using the same gear, you'd get diminishing returns with each level increase you got. what I can confirm though, and that while wearing 0 TH gear, the proc can still happen. I've had it happen on multiple occasions on my mules while leveling and running oddy chest hunting. my 99 mule has gone as high as TH 5 with no gear, and my other lower mules have gotten to 3 while leveling in the 50's. Though I 'm not sure that last bit was ever in doubt, but just confirming for consistency sake |
All FFXIV and FFXI content and images © 2002-2024 SQUARE ENIX CO., LTD.
FINAL FANTASY is a registered trademark of Square Enix Co., Ltd.