This topic contains 33 replies, has 15 voices, and was last updated by 
Miqobot 3 years, 5 months ago.
-
Posts
-
Hey miqo here is the information with regards to the theory behind it, hopefully it helps with regards to potential fixing the rotation.
https://www.akhmorning.com/jobs/smn/guide/rotation/#bahamut-phase
Miqo currently uses ruin II during bahamut phase, not causing a wyrmweave. Previously you could cause wyrmweave with ruin II but this has been changed since early shadowbringers and never adjusted.
Without consuming too much development time perhaps it’s prudent to just fix the use of ruin II during bahamut. And for future optimization if the expansion does not change it saving up as many charges of ruin IV’s as possible pre-bahamut trigger (without overcapping).It’s not a huge priority for myself, but it’s one of the very clear rotational problems I’ve seen since early expansion handicapping summoner slightly.
Yes, this is one of the guides we have been using when implementing Summoner rotation.
Miqobot tries to follow it as best as she can while keeping the algorithm reactive and adjusting on the fly. It is impossible to follow it with 100% accuracy due to multiple factors, that’s why our goal was 90% which is currently achieved.Miqo currently uses ruin II during bahamut phase, not causing a wyrmweave. Previously you could cause wyrmweave with ruin II but this has been changed since early shadowbringers and never adjusted.
We have just verified this information and we can see Bahamut reacting to Ruin II with Wyrmwave correctly.
Would you please clarify, are there any additional steps to reproduce this issue?Without consuming too much development time perhaps it’s prudent to just fix the use of ruin II during bahamut. And for future optimization if the expansion does not change it saving up as many charges of ruin IV’s as possible pre-bahamut trigger (without overcapping).
Thank you for your request.
However, it is not very different from the previous formulation. We still have to understand how to handle the contradiction described earlier.
If not Ruin II, what should Miqobot cast for oGCD weaving while both Egi Assaults are on cooldown?
If using Ruin IV before Bahamut phase is not allowed, what spell should be used to weave Summon Bahamut?If not Ruin II, what should Miqobot cast for oGCD weaving while both Egi Assaults are on cooldown?
Ruin II, III or IV are not oGCD abilities, neither are the Egi Assaults.
If using Ruin IV before Bahamut phase is not allowed, what spell should be used to weave Summon Bahamut?
During Dreadwyrm stance, Ruin III becomes an instant cast ability allowing for 2 oGCD weaves. Meaning you can use Summon Bahamut as an oGCD right after a Ruin III while Dreadwyrm stance is up.
I have a picture of the rotation that Miqobot is using here:
Initial opener miqo uses 4x egi assault, then uses 2x Ruin IV as GCD, while triggering Summon Bahamut. There is no need to use 2x Ruin IV because Ruin III already becomes an instant cast ability.
Bahamut stance (Ruin III has a casttime now – 2.36 haste so not ideal set):
20s – 7 WyrmWeave, 2 AM
0.67s: Enkindle Bahamut
1.33s: Akh Morn
3.52s: Ruin III
3.56s: Wyrmwave
5.93s: Ruin III
5.97s: Wyrmwave
8.38s: Ruin III
8.42s: Wyrmwave
9.00s: Ruin IV
9.93s: Wyrmwave
11.36s: Ruin IV
11.45s: Wyrmwave
12.03s: Enkindle Bahamut
12.69s: Akh Morn
15.50s: Ruin III
15.54s: Wyrmwave
16.13s: Ruin II
17.06s: WyrmwaveNow if we had saved up 4x Ruin IV, we can get those used as instant casts during Bahamut without need to ever use Ruin II. Ideally we also use the final Ruin IV at the end of Bahamut phase to get our 8th Wyrmweave in, but this might be too min-max than really needed.
I just have an issue with the current implementation of using Ruin IV for no reason pre-bahamut right at the opener while Ruin III is already instant cast, and using Ruin II as filler perhaps because of this issue?
Hope this helps.
-
This reply was modified 3 years, 6 months ago by
miko.
Ruin II, III or IV are not oGCD abilities, neither are the Egi Assaults.
Of course they are not. The question was what GCD spell to use to weave oGCD abilities.
During Dreadwyrm stance, Ruin III becomes an instant cast ability allowing for 2 oGCD weaves. Meaning you can use Summon Bahamut as an oGCD right after a Ruin III while Dreadwyrm stance is up.
Now if we had saved up 4x Ruin IV, we can get those used as instant casts during Bahamut without need to ever use Ruin II. Ideally we also use the final Ruin IV at the end of Bahamut phase to get our 8th Wyrmweave in, but this might be too min-max than really needed.
All right, these suggestions are more specific and this is indeed something we can work with.
Thank you for providing the detailed feedback.While working on Summoner rotation we already tried implementing it in a similar way, but we encountered multiple issues such as Summon Bahamut not triggering in time for second weave, GCD clipping, overcharging Egi Assaults, Egi Assaults going to waste due to Bahamut erasing pets, and several others. According to the preliminary development report, it could take 1-2 weeks to fix these issues and the overall gain from implementing the perfect rotation would be +1% DPS. This is the exponential cost that was mentioned earlier.
We can try implementing them again, but first we have to understand whether this is indeed what the community needs. We have recently discussed a similar problem with Machinist imperfection and Drill drift. So now we have to ask this question again.
Should we delay the release of next version for 2 weeks for Summoner optimizations?
And if we go this way with Summoner, why shouldn’t we do the same for other jobs?Delaying releases to implement small changes should not be decided by single or small group of users in my opinion. With regards to my feedback in particular, I’m happy at least its acknowledged that the rotation has some minor deficiencies that could be shored up for a future release whenever/if you have time to adjust it.
I’ll quote my previous post here I’ve posted over a year ago that had some positive support:
I think this part of miqo is bit untapped on how great it can work if rotations are kept up to date.
Considering the amount of classes keeping them all in top notch performance is perhaps too much to ask but I’d be very interested if we could be provided with some classes with more current rotations:– 1 melee [sam is least positional intensive]
– 1 ranged [mch is already top notch]
– 1 caster [smn is in a great spot already but needs rotation tweaks for 5.0]
– 1 healer [whm most likely considering popularity and rotation]
– 1 tank [any of the 4 works]If about 5 classes could be kept bit more up to date every few months that would be pretty amazing.
I’m pretty sure your schedule and plans for 6.0 are already filled. If you could keep this feedback in consideration and perhaps make a poll for certain rotation priority considering we’ll be having 19 classes in total that would be great. This way we can get certain rotations faster implemented and the workload lessened on the developers.
-
This reply was modified 3 years, 6 months ago by miko.
Of course, we will consider your feedback.
The main problem is that implementing even a single perfect rotation may easily occupy two years of development, due to unlimited optimization potential provided by the combat system of Final Fantasy XIV. This is why complex theorycrafting guides exist in the first place.
However, we believe that this system was not designed for an AI to beat it. It was designed for human players who enjoy optimizing and adapting their rotation for each raid encounter. And if we go too far down this path, we risk drawing unwanted attention and decreasing the value of the game itself.If our users decide that it must be done, we will of course follow the demand.
But we sincerely hope that this will never become a common use case for Assist Mode in Miqobot community.In any case, we appreciate the detailed analysis and constructive feedback.
Thank you very much!Curious if the implementation of custom rotations is going to be included in this update? It would make sense as it’s been stated that combat is not going to be worked on much again for a very long time after, but with the ability for custom rotations to be made it would it not allow the community to upkeep this feature with little need for dev resources (only if new skills are added)?
As it was announced, this update will include level 80 healer rotation.
Rotation Customization tools can not implemented in such little time. At the very least, they will require a dedicated development cycle of 4-8 weeks.We haven’t stated that the Combat system will not be improved anymore. We have stated that rotations are currently optimized to achieve 90% performance and the cost of every additional improvement after this grows exponentially.
Our development plans are driven by the community demand, so if our users decide that it must be done then we will continue working on it. After the current development cycle is complete, there will be a voting period during which the community will select the next feature pack for development.
It appears so. We were advised 3-5 days for the healer changes and an additional 7 for the additional improvements. We were told this on the 4th and it should have been completed by the 16th. Perhaps it will be released this weekend? Regardless, an additional update on the status would be appreciated.
Yes, as it always happens, we encountered some unexpected problems during development.
Please give us a few more days. Thank you for your patience very much.Thank you for all of your hard work very much 🙂
I think it has recently surfaced with all the new threads talking about combat assistance that this is one of the more popular features that people want. Maybe the dev team should focus a bit more on it. Not just rotational AI but more QoL features that can allow customization by users. (Multiple rotations, weight sliders on logical decisions) I think these features would serve the most people just by their inherent customizable nature. Also the team won’t always be begged to update rotations or what not as users themselves can do this. I think its a good long term investment. Disclaimer, not claiming to know anything about software development. Just my educated opinion from cursory glance.
Yes, Rotation Customization tools have already been requested and included in the future development plans. At the very least, they will require a dedicated development cycle of 4-8 weeks.
At the moment the combat performance of Miqobot has reached 90% and the community has been asking us to shift the development focus towards other features. After the current development cycle is complete, there will be a voting period during which the community will select the next feature pack for development.
-
This reply was modified 3 years, 6 months ago by
The topic ‘Status update on healer rotation changes’ is closed to new replies.