Using Python3 To Estimate BIS Gear Sets For Ninja.
Using Python3 to estimate BIS gear sets for Ninja.
which GUI module are you using? you should be able to scale it based on window resolution
Ramuh.Austar said: »
which GUI module are you using? you should be able to scale it based on window resolution
PySimpleGUI
I started that process back in November, but never got any complaints about UI scaling so dropped the priority in favor of adding more jobs.
Each GUI code already stores screen resolution:
Code
w, h = sg.Window.get_screen_size()
I just need to find a good value to use for scaling purposes. Unless you know of a better way to do it.
I wonder if it might even be possible with the "Initial Gearset" if you move the mouse over an item, you not only see the name but also the stats of the item.
This would have the advantage that missing stats or incorrect ones could be recognized more quickly.
This would have the advantage that missing stats or incorrect ones could be recognized more quickly.
Asura.Cluste said: »
I wonder if it might even be possible with the "Initial Gearset" if you move the mouse over an item, you not only see the name but also the stats of the item.
This would have the advantage that missing stats or incorrect ones could be recognized more quickly.
This would have the advantage that missing stats or incorrect ones could be recognized more quickly.
This is very easy to do. I actually had this exact same idea before I left for holiday, but I didn't like how it looked on my first attempt. The biggest problem is the stats for gear on the bottom row will occasionally overflow off-screen with low resolution displays (more UI scaling issues).
I think in the end it'll come down to me hard-coding which stats should share a line in an attempt to reduce how tall the tooltip box is. I'll push the update in January when I return and then modify it over time, just so something is there.
I've quickly re-created what I have on my main computer as an example of what it looked like. See below.
[+]
A Question about the set on Page 5 "Tachi: Fudo: 2000-2300 TP"
I was also playing around with SAM Sets, and found that ~2k TP "Doji" comes out ahead of "Masa" for Fudo, is that correct? Or is there a calc mistake somewhere.
Knobkierrie also wins for me at attack cap over Crep.
I was also playing around with SAM Sets, and found that ~2k TP "Doji" comes out ahead of "Masa" for Fudo, is that correct? Or is there a calc mistake somewhere.
Knobkierrie also wins for me at attack cap over Crep.
Asura.Cluste said: »
I was also playing around with SAM Sets, and found that ~2k TP "Doji" comes out ahead of "Masa" for Fudo, is that correct? Or is there a calc mistake somewhere.
I recall there being an issue with Fudo and Doji that I fixed a while ago. I don't remember what the issue was exactly. I think that I had an extra space in the "Tachi: Fudo" name, so the code was trying to compare "Tachi: Fudo" with "Tachi: Fudo ", which caused the +10% damage on Masamune to not trigger. If this is the case, then it would show up in your copy of check_weaponskill_bonus.py or weaponskill_scaling.py. I do not* see this issue myself.
Asura.Cluste said: »
Knobkierrie also wins for me at attack cap over Crep.
Knobkierrie should win for single-hit WSs like Fudo. That one was just me adding gear too quickly without paying enough attention. Edit: Of course this depends on total WSD, but Knobkierrie wins by an insignificant amount in the set I show there.
Izanami said: »
That's a lot of BST discussion for a Ninja/Python project thread. Interestingly, Ninja appears to be part of the discussion in the BST thread right now too. I recommend staying on topic. I don't think many users are going to search for BST testing results in the Ninja forums.
The code supports all jobs now if you wanted to download it and play with SAM and MNK. SAM sets are especially boring, though.
I'm slowly working on another FFXIAH node post that will contain sets for all jobs. So far I've created mid/high buff sets for DRK, DRG, MNK, SAM, BRD, WHM, and BST. This is a slow process since making FFXIAH item sets is not quick. If anyone knows of a way to semi-automate this, then please let me know.
See below for the copy/paste SAM and MNK sets taken from my node draft. The formatting is still a work in progress, but I think I'll settle for something similar to this anyway.
Edit: Looks like the far-right part of the sets gets cut off, though. At least you can mouse-over the partially visible item and see what it is.
Edit2: I've simply moved the cut-off sets down a row so they're visible in this spoiler format.
Bismarck.Nekhekh said: »
I wonder what the python method would look like for mnk and sam? So far this has helped me with my nin but still having trouble getting Gere ring from Odin.
The code supports all jobs now if you wanted to download it and play with SAM and MNK. SAM sets are especially boring, though.
I'm slowly working on another FFXIAH node post that will contain sets for all jobs. So far I've created mid/high buff sets for DRK, DRG, MNK, SAM, BRD, WHM, and BST. This is a slow process since making FFXIAH item sets is not quick. If anyone knows of a way to semi-automate this, then please let me know.
See below for the copy/paste SAM and MNK sets taken from my node draft. The formatting is still a work in progress, but I think I'll settle for something similar to this anyway.
Edit: Looks like the far-right part of the sets gets cut off, though. At least you can mouse-over the partially visible item and see what it is.
Edit2: I've simply moved the cut-off sets down a row so they're visible in this spoiler format.
Awesome work brother Kastra... Incredibly well-made and Resourceful App. ^_^
Sorry just to make sure :) When u talk about Impulse Drive: 2000-2300 TP, or something like Upheaval etc.
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
Izanami said: »
I think something went wrong with head here.
Both Mpaca R25+ and AdhemarB should be better (no attack on Blistering is also a problem imo, but mor eon this below)
I would also argue Mpaca feet R30 to be better too, because damage is probably almost identical, but 78 attack would help this set to work against slightly stronger targets with same buffs too.
Izanami said: »
High-buff uses the same buffs and enemy, but I reduced the enemy defense to 1 to force attack cap.
Imo this is not a good practice. It will show you completely unrealistic sets using items that will almost never cap attack in any real scenario that matters in current end-game. Like for example Malignance pieces, that has PDL, but no attack and low STR. It would probably be ok in escha era when you was able to drop everything to attack capped with bog/ea or bolster Frailty with Dia II, but it's no longer the case.
Im not sure what would be optimal approach here. I usually just assume capped attack sets should also have high attack, so multipliers like Chaos/Fury have a chance to actually push you to capped attack with PDL.
Asura.Cluste said: »
Sorry just to make sure :) When u talk about Impulse Drive: 2000-2300 TP, or something like Upheaval etc.
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
When I say "2000-2300 TP", I mean I use min=2000, max=2300 as inputs. These are the values that the player waits for before using the weapon skill. TP Bonus gets added to these values before the damage is calculated. I've been considering updating the sets with lower inputs so TP bonus may push them to the optimal range, but I haven't decided if that's the best approach yet.
For now, I simply check the TP scaling for each weapon skill and decide on a reasonable range to use.
SimonSes said: »
Izanami said: »
I think something went wrong with head here.
Both Mpaca R25+ and AdhemarB should be better (no attack on Blistering is also a problem imo, but mor eon this below)
I would also argue Mpaca feet R30 to be better too, because damage is probably almost identical, but 78 attack would help this set to work against slightly stronger targets with same buffs too.
As far as I can tell, there is nothing "incorrect" with that set or the algorithm that found it. That set is for when Impetus is up and with capped attack between 1000 and 1300 TP. If I remove Impetus, then I see the following set, which does favor Mpaca and Nyame since the set is not receiving the valuable crit rate and crit damage from Impetus and Empy+3 body.
ItemSet 389270
You are correct that Adhemar(B) and Mpaca are only 1~2% worse than Blistering Sallet in that TP range for the Impetus set. Additionally, Blistering Sallet starts to lose at slightly higher TP values when the extra TP over-caps crit rate with Impetus. In this case, I agree that Mpaca would be a better choice than Blistering Sallet, despite being worse in a vacuum.
SimonSes said: »
Izanami said: »
High-buff uses the same buffs and enemy, but I reduced the enemy defense to 1 to force attack cap.
Imo this is not a good practice. It will show you completely unrealistic sets using items that will almost never cap attack in any real scenario that matters in current end-game. Like for example Malignance pieces, that has PDL, but no attack and low STR. It would probably be ok in escha era when you was able to drop everything to attack capped with bog/ea or bolster Frailty with Dia II, but it's no longer the case.
Im not sure what would be optimal approach here. I usually just assume capped attack sets should also have high attack, so multipliers like Chaos/Fury have a chance to actually push you to capped attack with PDL.
You make a good point that I did not fully consider previously. I'll return to the original method used for all of the Ninja sets on the first page and simply add GEO buffs for the "high" buff state.
However, finding the set that provides just enough attack to trigger PDL is not feasible when checking sets across all different weapons and jobs (I'm up to about 220 sets created so far). The best I can do is provide realistic "high" buffs and reasonable enemy stats for the sets I publish and allow the interested users to download the code to check specifics themselves.
I have been following this forums and really love the work everyone is putting into theory crafting. This will be an absolute benefit to so many players. /cheers! :)
I wonder if it is possible to place a certain mDT, pDT, Meva or even Subtle Blow toggle to calculate best set options while remining defensive in fights like Odyssey T25 NMs.
Once again, just such amazing work!
I wonder if it is possible to place a certain mDT, pDT, Meva or even Subtle Blow toggle to calculate best set options while remining defensive in fights like Odyssey T25 NMs.
Once again, just such amazing work!
Izanami said: »
Asura.Cluste said: »
Sorry just to make sure :) When u talk about Impulse Drive: 2000-2300 TP, or something like Upheaval etc.
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
In the Basic Inputs, you put in Min. TP. 1550, Max. TP 1850 for SAM Impulse Drive.
Min. TP. 1150, Max. TP 1450 for lets say Upheaval (Chango, Moonshade, Empy Legs +3)
The Simulator calc the effective TP then @2000~2300?
When I say "2000-2300 TP", I mean I use min=2000, max=2300 as inputs. These are the values that the player waits for before using the weapon skill. TP Bonus gets added to these values before the damage is calculated. I've been considering updating the sets with lower inputs so TP bonus may push them to the optimal range, but I haven't decided if that's the best approach yet.
For now, I simply check the TP scaling for each weapon skill and decide on a reasonable range to use.
Do Players wait for 2000-2300 with TP-Bonus gear? For lets say Upheaval, i don't hold my TP to 2000~, but i release at 1150~ TP, so with all the TP-Bonus Gear it reaches 2000~. I thought that was the usual approach, or why do you think this is not the best approach?
If i misunderstood you, sorry English is not my main language :)
[+]
Izanami said: »
As far as I can tell, there is nothing "incorrect" with that set or the algorithm that found it. That set is for when Impetus is up and with capped attack between 1000 and 1300 TP. If I remove Impetus, then I see the following set, which does favor Mpaca and Nyame since the set is not receiving the valuable crit rate and crit damage from Impetus and Empy+3 body.
It must be something wrong then.
Maybe you dont have fTP replication on Victory Smite?
Against Mpaca its 8STR and 3%crit rate vs 3%TA 2%DA
3%TA and 2%DA is around +4.29% avg damage for Vsmite with the rest of this set.
8STR is like 0.92% and 3%crit rate is +1.2~1.5% avg damage, depends what values you assume from dDEX and Impetus.
3%TA and 2%DA should be a clear winner here.
Unless it's a sim thing that MA proc is more random and it creates lower median than more consistent 8%STR and 3%crit rate?
I'd guess it's probably the impetus stack count as the only difference. the methodology is pretty much the same as the spreadsheet in every other case.
most likely different variables you're using for any other differences
most likely different variables you're using for any other differences
Asura.Otomis said: »
I wonder if it is possible to place a certain mDT, pDT, Meva or even Subtle Blow toggle to calculate best set options while remining defensive in fights like Odyssey T25 NMs.
Asura.Cluste said: »
Do Players wait for 2000-2300 with TP-Bonus gear? For lets say Upheaval, i don't hold my TP to 2000~, but i release at 1150~ TP, so with all the TP-Bonus Gear it reaches 2000~. I thought that was the usual approach, or why do you think this is not the best approach?
No, you have it correct. Players should use the weapon skills such that their effective TP is within the optimal range.
Because the code uses raw TP, I'm not sure if it's better to intentionally give it low TP, which essentially forces TP Bonus to show up, or to intentionally give it optimal TP, which may cause TP Bonus to not show up at all. I typically just pick 1500-1800 TP as a middle-ground.
I'd like to add effective TP to the code to deal with this, but it'll likely need to wait for when TP sets are combined with WS sets for total DPS.
SimonSes said: »
Izanami said: »
As far as I can tell, there is nothing "incorrect" with that set or the algorithm that found it. That set is for when Impetus is up and with capped attack between 1000 and 1300 TP. If I remove Impetus, then I see the following set, which does favor Mpaca and Nyame since the set is not receiving the valuable crit rate and crit damage from Impetus and Empy+3 body.
It must be something wrong then.
Maybe you dont have fTP replication on Victory Smite?
Against Mpaca its 8STR and 3%crit rate vs 3%TA 2%DA
3%TA and 2%DA is around +4.29% avg damage for Vsmite with the rest of this set.
8STR is like 0.92% and 3%crit rate is +1.2~1.5% avg damage, depends what values you assume from dDEX and Impetus.
3%TA and 2%DA should be a clear winner here.
Unless it's a sim thing that MA proc is more random and it creates lower median than more consistent 8%STR and 3%crit rate?
The automated set finder does not use the simulations. It calculates average damage, similar to the spreadsheets. Maybe you're slightly overestimating damage from multi-attacks, which share FTP, but not the additional 12% WSD?
Also note that Blistering Sallet is 6% more crit rate (10% Blistering vs 4% Mpaca), not 3%. This is a small correction, but Impetus with Empyrean +3 body is giving those extra crits +45% damage dealt when Impetus is 90% built, which is what my code assumes. The TP Bonus on Mpaca's boosts crit rate by a bit as well, which makes up for some/all of the missing 3% anyway.
These are small differences, but Blistering Sallet only wins by a small amount anyway. I wouldn't worry about it too much and just use Mpaca's Cap R25+ anyway.
[+]
Izanami said: »
Also note that Blistering Sallet is 6% more crit rate (10% Blistering vs 4% Mpaca), not 3%.
I assumed 3% crit from 200TP bonus, for a total of 7% on Mpaca.
Izanami said: »
The automated set finder does not use the simulations. It calculates average damage, similar to the spreadsheets. Maybe you're slightly overestimating damage from multi-attacks, which share FTP, but not the additional 12% WSD?
I don't think I overestimate it, because it's just math. I added details like wsd on first hit and 99% hit rate to get even better number and I got 1.0421 for 3%TA and 2%DA more (3%QA,21~24%TA,14~16%DA)
Including WSD on first hit:
6.7460 hits for Mpaca
6.47334072 hits for Blistering
*Keep in mind it's not fTP, it's hits, since it does't matter as fTP would be just a multiplier for each hit, so wouldn't change % difference.
Izanami said: »
No, you have it correct. Players should use the weapon skills such that their effective TP is within the optimal range.
Because the code uses raw TP, I'm not sure if it's better to intentionally give it low TP, which essentially forces TP Bonus to show up, or to intentionally give it optimal TP, which may cause TP Bonus to not show up at all. I typically just pick 1500-1800 TP as a middle-ground.
I'd like to add effective TP to the code to deal with this, but it'll likely need to wait for when TP sets are combined with WS sets for total DPS.
Because the code uses raw TP, I'm not sure if it's better to intentionally give it low TP, which essentially forces TP Bonus to show up, or to intentionally give it optimal TP, which may cause TP Bonus to not show up at all. I typically just pick 1500-1800 TP as a middle-ground.
I'd like to add effective TP to the code to deal with this, but it'll likely need to wait for when TP sets are combined with WS sets for total DPS.
Leave it as is, the TP you would have on the TP bar, before any gear.
[+]
SimonSes said: »
I don't think I overestimate it, because it's just math. I added details like wsd on first hit and 99% hit rate to get even better number and I got 1.0421 for 3%TA and 2%DA more (3%QA,21~24%TA,14~16%DA)
Including WSD on first hit:
6.7460 hits for Mpaca
6.47334072 hits for Blistering
*Keep in mind it's not fTP, it's hits, since it does't matter as fTP would be just a multiplier for each hit, so wouldn't change % difference.
Including WSD on first hit:
6.7460 hits for Mpaca
6.47334072 hits for Blistering
*Keep in mind it's not fTP, it's hits, since it does't matter as fTP would be just a multiplier for each hit, so wouldn't change % difference.
I don't think I understand. Could you explain a bit more how you get your number of hits just for consistency?
The ambuscade cape has +10% DA. In total, I'm getting 3% QA, 21% TA, 26% DA for the Blistering Sallet set. Even ignoring that difference in double attack, I'm seeing only ~5.3 hits on average per weapon skill. This is very different from your ~6.5 and suggests that I may not be treating hand-to-hand weapon skills correctly.
Do hand-to-hand weapon skills get off-hand attacks? Victory Smite is a four-hit weapon skill, with two chances to multi-attack. My code is seeing the following number of attacks per Victory Smite in my set (3% QA, 21% TA, 26% DA):
main_hand, off_hand, total
1: 5.32308744, 0.0, 5.32308744 (1,000,000 simulations based on MA rates.)
2: 5.33934324, 0.0, 5.33934324 (A simple equation for averages.)
Izanami said: »
SimonSes said: »
I don't think I overestimate it, because it's just math. I added details like wsd on first hit and 99% hit rate to get even better number and I got 1.0421 for 3%TA and 2%DA more (3%QA,21~24%TA,14~16%DA)
Including WSD on first hit:
6.7460 hits for Mpaca
6.47334072 hits for Blistering
*Keep in mind it's not fTP, it's hits, since it does't matter as fTP would be just a multiplier for each hit, so wouldn't change % difference.
Including WSD on first hit:
6.7460 hits for Mpaca
6.47334072 hits for Blistering
*Keep in mind it's not fTP, it's hits, since it does't matter as fTP would be just a multiplier for each hit, so wouldn't change % difference.
I don't think I understand. Could you explain a bit more how you get your number of hits just for consistency?
The ambuscade cape has +10% DA. In total, I'm getting 3% QA, 21% TA, 26% DA for the Blistering Sallet set. Even ignoring that difference in double attack, I'm seeing only ~5.3 hits on average per weapon skill. This is very different from your ~6.5 and suggests that I may not be treating hand-to-hand weapon skills correctly.
Do hand-to-hand weapon skills get off-hand attacks? Victory Smite is a four-hit weapon skill, with two chances to multi-attack. My code is seeing the following number of attacks per Victory Smite in my set (3% QA, 21% TA, 26% DA):
main_hand, off_hand, total
1: 5.32308744, 0.0, 5.32308744 (1,000,000 simulations based on MA rates.)
2: 5.33934324, 0.0, 5.33934324 (A simple equation for averages.)
Yeah you are right, it was my mistake. I took additional hits from 2 attacks and still doubled it by mistake (It wasnt in sheet, just some napkin math).
That being said, you wrote Segomo: Str + crit under that set in your previous post, so difference in DA is only coming from you changing cape to DA :) Still not sure how you are getting 26%DA instead of 24%DA though? Schere(6)+Sherida(5)+blistering(3)+cape(10) should be 24.
Also the reason you are getting DA cape with Blistering is because you are probably caping crit rate on your target. Bats that you take as target only have 275AGI right? Im still getting Adhemar wins by like 2% with crit cape and Mpaca marginally wins with both capes, but crit cape only gives Mpaca 7% crit, because of a cap.
If I push agi on target to like 375 (dDEX drops to 0), then all of head pieces works best with crit cape and Mpaca wins by like 0.6%
The only time I have Blistering winning is when it's 11DEX advantage has effect on dDEX between +10 and +50
SimonSes said: »
That being said, you wrote Segomo: Str + crit under that set in your previous post, so difference in DA is only coming from you changing cape to DA :) Still not sure how you are getting 26%DA instead of 24%DA though? Schere(6)+Sherida(5)+blistering(3)+cape(10) should be 24.
Also the reason you are getting DA cape with Blistering is because you are probably caping crit rate on your target. Bats that you take as target only have 275AGI right? Im still getting Adhemar wins by like 2% with crit cape and Mpaca marginally wins with both capes, but crit cape only gives Mpaca 7% crit, because of a cap.
If I push agi on target to like 375 (dDEX drops to 0), then all of head pieces works best with crit cape and Mpaca wins by like 0.6%
The only time I have Blistering winning is when it's 11DEX advantage has effect on dDEX between +10 and +50
Also the reason you are getting DA cape with Blistering is because you are probably caping crit rate on your target. Bats that you take as target only have 275AGI right? Im still getting Adhemar wins by like 2% with crit cape and Mpaca marginally wins with both capes, but crit cape only gives Mpaca 7% crit, because of a cap.
If I push agi on target to like 375 (dDEX drops to 0), then all of head pieces works best with crit cape and Mpaca wins by like 0.6%
The only time I have Blistering winning is when it's 11DEX advantage has effect on dDEX between +10 and +50
You are correct. I am seeing STR+Crit cape. I ended up having a few GUIs open and must have moused over the wrong window's cape slot to see augments. Subjob Warrior provided the +12% DA, which pushed the total to 26%. Do players usually subjob Dragoon with Monk?
I used 356 AGI for the enemy.
The damage with Blistering Sallet is 53402, while Mpaca's Cap R30 shows 52920 in this specific situation. Still not enough to worry about, especially considering that even a little bit of extra TP (more than the 1150 average used to find the set) will cause Mpaca's to win.
Edit:
If I enter 1300 as the TP value, then Mpaca's Cap R30 wins 53501 vs 53402. Since the Blistering Sallet set did not increase damage with more TP, it suggests that Blistering Sallet only wins because it was needed to cap crit rate, and it does so faster than Mpaca's Cap R30. Increasing TP further shows that Mpaca's Cap R30 caps Crit Rate at around 1500 TP with damage "cap" of 54143. (Adhemar Bonnet +1B "caps" at 53638 damage around 1600 TP it seems).
You are correct to say that Mpaca's Cap should be considered the best option. However, Blistering Sallet wins on paper for this very specific situation with a crit weapon skill on the border of the crit rate cap.
Edit2:
This is just a rare situation in general. If I had used 100% Impetus potency instead of 90%, then we'd get +5% more crit rate and Blistering Sallet +1 would fall off even faster. Using much lower Impetus potency may result in a larger window for Blistering Sallet +1 to beat Mpaca.
Izanami said: »
I used 356 AGI for the enemy.
Izanami said: »
Subjob Warrior provided the +12% DA, which pushed the total to 26%. Do players usually subjob Dragoon with Monk?
Ok, that makes sense now.
1. You are between 0-50 dDEX, so 11DEX from Blistering provides additional crit rate
2. I was using DRG subjob. I wouldnt say DRG subjob is straight up better, but I just don't like inconsistency of 3min Berserk and 2min without it and also def drop that goes with it.
Izanami said: »
Edit2:
This is just a rare situation in general. If I had used 100% Impetus potency instead of 90%, then we'd get +5% more crit rate and Blistering Sallet +1 would fall off even faster. Using much lower Impetus potency may result in a larger window for Blistering Sallet +1 to beat Mpaca.
This is just a rare situation in general. If I had used 100% Impetus potency instead of 90%, then we'd get +5% more crit rate and Blistering Sallet +1 would fall off even faster. Using much lower Impetus potency may result in a larger window for Blistering Sallet +1 to beat Mpaca.
What Impetus stats you are using exactly? How much crit rate/damage?
Izanami said: »
Adhemar Bonnet +1B "caps" at 53638 damage around 1600 TP it seems
This doesn't make sense at all. Adhemar bonnet should have by far highest ceiling at capped crit rate. When crit rate is capped, Adhemar is 6% crit damage and 4%TA vs 10STR and 3%DA on Blistering. It's not even close, Adhemar should be like 5% ahead with capped crit rate.
[+]
SimonSes said: »
Izanami said: »
Adhemar Bonnet +1B "caps" at 53638 damage around 1600 TP it seems
This doesn't make sense at all. Adhemar bonnet should have by far highest ceiling at capped crit rate. When crit rate is capped, Adhemar is 6% crit damage and 4%TA vs 10STR and 3%DA on Blistering. It's not even close, Adhemar should be like 5% ahead with capped crit rate.
Thanks for the note. It seems my gear.py file is giving Adhemar Bonnet +1A Crit Damage, but Path B is getting Crit Rate. This must be an old typo since it should've been added when the code only supported NIN.
Fixing this shows Adhemar Bonnet +1B has an average damage cap of 55857 at around 1900 TP. Blistering Sallet +1 is still showing up as best in the given situation, though (53402 vs 52583 against Adhemar +1B).
SimonSes said: »
What Impetus stats you are using exactly? How much crit rate/damage?
90% max potency:
+90 Attack
+45% Crit Rate
+45% Crit Damage (Empy Body +3 equipped)
[+]
Izanami said: »
That's a lot of BST discussion for a Ninja/Python project thread. Interestingly, Ninja appears to be part of the discussion in the BST thread right now too. I recommend staying on topic. I don't think many users are going to search for BST testing results in the Ninja forums.
The code supports all jobs now if you wanted to download it and play with SAM and MNK. SAM sets are especially boring, though.
I'm slowly working on another FFXIAH node post that will contain sets for all jobs. So far I've created mid/high buff sets for DRK, DRG, MNK, SAM, BRD, WHM, and BST. This is a slow process since making FFXIAH item sets is not quick. If anyone knows of a way to semi-automate this, then please let me know.
See below for the copy/paste SAM and MNK sets taken from my node draft. The formatting is still a work in progress, but I think I'll settle for something similar to this anyway.
Edit: Looks like the far-right part of the sets gets cut off, though. At least you can mouse-over the partially visible item and see what it is.
Edit2: I've simply moved the cut-off sets down a row so they're visible in this spoiler format.
Bismarck.Nekhekh said: »
I wonder what the python method would look like for mnk and sam? So far this has helped me with my nin but still having trouble getting Gere ring from Odin.
The code supports all jobs now if you wanted to download it and play with SAM and MNK. SAM sets are especially boring, though.
I'm slowly working on another FFXIAH node post that will contain sets for all jobs. So far I've created mid/high buff sets for DRK, DRG, MNK, SAM, BRD, WHM, and BST. This is a slow process since making FFXIAH item sets is not quick. If anyone knows of a way to semi-automate this, then please let me know.
See below for the copy/paste SAM and MNK sets taken from my node draft. The formatting is still a work in progress, but I think I'll settle for something similar to this anyway.
Edit: Looks like the far-right part of the sets gets cut off, though. At least you can mouse-over the partially visible item and see what it is.
Edit2: I've simply moved the cut-off sets down a row so they're visible in this spoiler format.
AWESOME. I will look at what have as far as mnk and sam based on these. I am going to assume the same food "grape daifuku"?
Bismarck.Nekhekh said: »
Izanami said: »
AWESOME. I will look at what have as far as mnk and sam based on these. I am going to assume the same food "grape daifuku"?
Yeah. Those draft sets use Grape Daifuku against 1300 Eva, 1400 Def, 356 AGI, 267 INT, and 289 VIT.
Note: These sets all use R30 Odyssey gear. They may be slightly different with R25 gear instead.
Okay... then that's what the problem is at least for H.F. I seem to get higher damage with the following set:
Coiste
Mpaca's Cap
Monk's +2 (Aug)
Sherida
Moonshade AttTP
Tatenashi Haramaki +1
Herculean (4%TA,10STR,20acc/att)
Niqmaddu
Gere Ring
Segomo (STR+ DA)
Moonbow +1
Mpaca's Hose
Herculean (4%TA,10STR,20acc/att)
not even close to R25.
Coiste
Mpaca's Cap
Monk's +2 (Aug)
Sherida
Moonshade AttTP
Tatenashi Haramaki +1
Herculean (4%TA,10STR,20acc/att)
Niqmaddu
Gere Ring
Segomo (STR+ DA)
Moonbow +1
Mpaca's Hose
Herculean (4%TA,10STR,20acc/att)
not even close to R25.
Bismarck.Nekhekh said: »
Okay... then that's what the problem is at least for H.F. I seem to get higher damage with the following set:not even close to R25.
If you have access to R15 Mpaca, then I'm seeing the following two sets for Howling Fist. The Odyssey armor is R15, but I haven't added lower rank Odyssey accessories or weapons yet, so R30 Coiste shows up.
Edit: I do not include any augmented Herculean gear. You'll have to play with the code yourself to check how these compare to your herculean gear if you're curious.
Herculean is for sure better in many sets, but I also stopped even considering it whenever I can, because it's just too much to have different augment for everything and have so many toggles for different WS sets. like capped/uncapped attack, glasscannon/hybrid/normal, then you need to pair all of them. Maybe if you don't play many jobs and you have perfect timing and like to change toggle non stop during the fight to get 2% dps more, then it's worth to consider :)
I would never risk to wear that for like 2%(?) dps boost
Wearing that on Ngai, A3 or anything else from serious endgame is suicide. I guess maybe for something like A2 kite Henwen for first 50% when you are sure that you wont take hate, but I wouldnt even consider inventory space loss for something so niche.
Bismarck.Nekhekh said: »
Okay... then that's what the problem is at least for H.F. I seem to get higher damage with the following set:
Coiste
Mpaca's Cap
Monk's +2 (Aug)
Sherida
Moonshade AttTP
Tatenashi Haramaki +1
Herculean (4%TA,10STR,20acc/att)
Niqmaddu
Gere Ring
Segomo (STR+ DA)
Moonbow +1
Mpaca's Hose
Herculean (4%TA,10STR,20acc/att)
not even close to R25.
Coiste
Mpaca's Cap
Monk's +2 (Aug)
Sherida
Moonshade AttTP
Tatenashi Haramaki +1
Herculean (4%TA,10STR,20acc/att)
Niqmaddu
Gere Ring
Segomo (STR+ DA)
Moonbow +1
Mpaca's Hose
Herculean (4%TA,10STR,20acc/att)
not even close to R25.
I would never risk to wear that for like 2%(?) dps boost
Wearing that on Ngai, A3 or anything else from serious endgame is suicide. I guess maybe for something like A2 kite Henwen for first 50% when you are sure that you wont take hate, but I wouldnt even consider inventory space loss for something so niche.
An updated version of the code has been uploaded to the GitHub page. I've made the following changes over the last few weeks:
The gear.py file has 962 entries with a "Name" dictionary key, suggesting 962 items have been added so far. Let me know if you see any typos or missing stats in any of these items. Also let me know if you can't find an item that should be included. I'll work on adding all weapon skills soon.
Added Fell Cleave, Requiescat, and Detonator.
- Sequence provides +10% damage to Requiescat
Added various missing stats on some gear.
Added a handful of missing gear pieces such as Gerdr Belt, Odnowa Earring, Gelatinous Ring, and Defending Ring. I'll need to add more TP gear soon as well.
Added +1 JSE necks and +1 JSE earrings. These were mostly added because +2 earrings are a 4% drop rate from +1 cases, then another 1/22 chance to get the correct job. Additionally, most of the +2 necks are only worth ~1% more damage for ~30m more gil.
- +2 versions are selected by default when using the "Select All" buttons. You'll need to manually select +1s if you want to check them. Keep in mind that the code allows JSE Earrings in both ear slots, so you'll need to unselect the +2s from from both.
- JSE Earrings use the mid-value for augments. (Example: +2 earrings range from 7-15 STR, so the code will use +11 STR)
Added R15/20/25 odyssey accessories and weapons.
- Using the "Select All" buttons will only select the odyssey rank that you specify with the drop down menu.
Added DT, PDT, MDT, Evasion, Magic Evasion, and Magic Defense to gear and stats in preparation for TP sets. These stats can be viewed for each gear set in the "TP" section of the Outputs tab.
The outputs tab now shows Magic Burst II separately from Magic Burst Traits to give a better idea of the contribution from gear.
DRG weapon skill damage traits are now treated as a separate damage multiplier:
- (damage)*(1+wsd)*(1+ws_bonus)*(1+ws_trait)
Dead Aim (ranged crit damage) has been added.
Dead Aim has been removed (2023 January 05) since it does not apply to weapon skills.Added a checkbox for True Shot when RNG or COR is selected as main job.
- Disabling this will disable true shot bonuses entirely (no damage bonus for standing near the correct distance).
- Enabling this will provide the full damage bonus (plus gear) as if standing at the exact correct distance.
Updated ML20 base stats for all jobs except BST, BRD, and SMN.
- BST and BRD still need corrected stats since my BST and BRD are both Lv1.
- BLU has the correct stats, but still does not include any traits.
- BST and BRD still need corrected stats since my BST and BRD are both Lv1.
Added all jobs as subjobs.
- BST and BRD do not use correct stats since my BST and BRD are both Lv1.
All +3 Artifact gear now activates the Regal Ring set bonus. Previously, only NIN and SAM AF+3 would activate Regal Ring bonuses.
Corrected issue with handling accuracy from skill level (Katana Skill, etc). This change ended up being worth about -10 accuracy from skill.
Hovering your mouse pointer over the gear icons on the inputs tab now shows a summary of all stats used for that piece of gear. I'll update the format of this later.
Added a simple check for user screen resolution. If screen height is less than 900px, then use a horizontal layout. It doesn't look great, but it at least should be useable.
- This is a temporary fix until I get around to dealing with the UI scaling issues.
- The inputs tab should now show all content for displays with screen height of at least 600px
- Many of the stats in the outputs tab get cut off.
The gear.py file has 962 entries with a "Name" dictionary key, suggesting 962 items have been added so far. Let me know if you see any typos or missing stats in any of these items. Also let me know if you can't find an item that should be included. I'll work on adding all weapon skills soon.
[+]
Been following your work and wanted to drop in and say thank you for putting this together. And thank you to everyone testing it out and providing feedback.
All FFXI content and images © 2002-2024 SQUARE ENIX CO., LTD. FINAL
FANTASY is a registered trademark of Square Enix Co., Ltd.