Kraken Club Offhand Hits Per Round Simulation

If anyone is curious how the NIN spreadsheet operates when determining WS/TP/Cycle probability, feel free to ask me.


... nicely.

And this is relevant to Kraken Club offhand viability, how exactly?

Gonna slip this in since it was mentioned.

I found that the hardest part of modeling Raetic in a sheet wasn't so much the inclusion of the FUA's mechanic, i.e. its impact on strike attempts when wielded, but rather the the fact that the HQs introduce variation into the TP gain per strike.

When you're already modeling X attempted strikes at Y probability per hand, with Z TP gain per connected hit, this spreads the probability space out over a whole new table of X, Y, and Z values at which the altered TP gain must be considered.

The arrayformula in google to do this was something like 10,000 characters. Pure simulation is likely a lot simpler, and more practical for this kind of thing, generally - Simulation doesn't have to encapsulate these things in a probability space. My google sheets autism is kind of off-the-charts...so I made it happen.

Edit*: Language in last paragraph.

I wonder if this is entering into the problem space where the query function is what we really want to be using, have you considered leveraging it to simplify things? Or is it not very applicable to how you approached the problem?

I exposed the sheet from the solution here, if you want to see it. I leverage the query function, in most of my google stuff, rather extensively.

It's one of the few areas of the solution where I extensively documented the approach (due to the complexity). Feel free to leverage:
https://docs.google.com/spreadsheets/d/10LpRCPET1mYTV2rcNIZZvCLJeXHuWePKmNmSDOHIgjM/edit#gid=1611426922

Edit*: I favor fewer and more dynamic formulas. Sometimes this means that they get big. Google has some unfortunate limitations.

I'll check it out, thanks for posting it up!

Do any of the spreadsheets account for either the initial delay in engage , I’m talking the game delay not me being old and slow to hit attack, or the delay from WS itself? I’ve seen some add in job ability delay but I don’t see them add in this standard delay which throws off many comparisons and is always present and therefore overstates dps especially as you approach relatively small amounts of time to WS.

Yes. *In the NIN sheet*, an additional 120 (2 Sec.) is added for WS Delay.

It's simpler to include WS lockout into the DPS formulation and the starting TP as your anticipated weighted-average TP return than it is to have a separate iteration where you "start from the moment of engagement".

If your combined delay is less than 2 seconds, which it is in most haste-capped dual wield scenarios, then this method is actually more conservative (hurts your DPS).

The net, impact of this depends on the implemented measure of time and is negligible, imo.

So the reason I asked is actually somewhat simple In nature but to which is never accounted for in people’s napkin math. While a 2 second delay sounds minimal when you achieve 1k tp in 3 seconds or less that engage/WS delay actually becomes a huge part of the overall delay and performance of a job in general. So I believe we should account for it as accurately as possible especially among jobs that use multiple weapon types or setups where total time between actions can vary somewhat significantly.

It heavily depends on your use for the dps spreadsheet.

If your use for it is purely just "which piece of gear is better dps?" The ws/engage delay would impact both scenarios pretty close to equally as far as I am aware, and thus have little to no impact on the decision.

I appreciate the work by Austar and don't wish my post to undermine their effort in any way. I've always been more interested in theory rather than experiment so this may just be a point of ideological difference. I just can't see the value in a random number based iterative simulation. Any useful data must theoretically lie at a point of convergence which is provable by mathematical model alone rather than simulation. What use is it to simulate this value rather than determine it mathematically? If it is to show variance or demonstrate the marginal comparison between two similar options, couldn't variance instead be determined mathematically? I see how this replicates what is done in game but I can't help but feel we've spent years trying to isolate the the theory from the game just so we can make those objective comparisons between similar choices, however insignificant they are. On the other hand I could see how it may be too complex to formulate a mathematical model, and I'm not familiar enough with the variables to make that distinction. In that case I've missed the point entirely and I apologize.

dynamic behavior. being able to properly perform things such as starting at 0 tp and building up to 3K for aftermath. Performing aftermath and using a different WS for damage if desired. Saving TP for when it's wearing off to re-apply. Unique effects like Karambit, Raetic, Taruet, Hachimonji. I'm sure you can find a way to make some of those work, but would require a lot more effort.

It also reflects realistic variability. With the amount of variables involved in this game, calculating it would take a lot more than just running a simulation several thousand times and just adding one line of code to print the variance, IQR, mean, median, histograms, literally anything I need with the data is easily done. I just don't see why people are so against the fact that something that emulates actual game play. Do you just not understand what to do with the results?

Thanks for clarifying! It did seem as though it would be much more difficult to make a general model rather than having the simulation brute force the data for you. I didn't understand if the simulation was trying to give us something more valuable than a theoretical model or if it was just simpler to simulate, yet still adequate.

I'm not against it, thanks for your contribution to the community!

I think I would rather claw my eyes out than attempt to extend the mathematical/statistical model to the breadth of scenarios Austar is able to simulate. Screw that.

I havent seen anyone in this entire thread "against it", not even I. Do you feel like someone is "against it"?


I mean all of that sounds like excellent additions to make the data more meaningful and that sounds pretty great! I'm sure people would really appreciate it.

I want to be clear I dont think your program is bad, nor am I against it, nor do I perceive anyone else here is "against it", I think you just need to be very clear on exactly what you're goal, your commander's intent, of the program is.

When you say "DPS simulation", I think traditionally most people who play RPGs expect the simulation to put out the exact same result every single time it runs, with extremely low variance, so that they can simulate it with ever so slight tweaks so they can min/max their parameters.

Im not saying your program has to do that, but Im just saying you should be clear about the intended use of the program.

Something along the lines of "Here's how you can use this as a tool" in your post.

I think the fact multiple people responded in confusion and are asking pretty much the same questions should indicate to you theirs a gap between you and your end user.

Nothing wrong with that but hey, I bet people would really appreciate you clarifying "This is how you use it in a meaningful way, here's an example!"

Cheers :)

No simulation does that. If that was the case, you wouldn't need one in the first place.

This isn't the game to play if you want extremely low variance.

It really isn't that complicated. You run it with whatever settings and variables you need and get all the data you'd need to make a decision. For example, do I want to spend 80M on epam ring? Well, probably not since the difference you'll get in real world applications is non-existent. Most people don't know what to think of the data it provides and I've had maybe two people attempt to understand.

Not every RPG has an easy way to track DPS, many dont have chat logs or ways to track damage dealt, so individuals created simulators of DPS that run for exceptionally long times, allowing you to test gear.

And those DPS simulators have extremely low variance where even 1% differences in gear are noticeable.

A great example of this is the Diablo 3 gear simulator that is still maintained today: https://www.d3planner.com/

Another example would be Path of Exile's "Path of Building" application, which has unbelievably complex depth to its calculations, but is extremely accurate.

https://github.com/Openarl/PathOfBuilding

I think most people who play a lot of RPGs think of a tool like this when they hear "DPS Simulator", so thats where the confusion comes from.


Yet you have multiple people responding to your thread confused about the application of it. I mean you dont have to tell people how to use it, but the fact you get a bunch of people scratching their head over the value of your program kind of should make you pause and go "Hmm"

I mean or just keep being antagonistic to everyone, if you want.

Neither of those games have the amount of RNG FFXI has, and they still have RNG. If they didn't, there would be no point to having a simulation to begin with, since it would be the same thing every time.

Its literally in the title and very obvious what it does. If you can't figure out what it does based on that, well there is no helping you.

If you're talking about full DPS simulations, nobody has asked me what the data is.

Path of Exile community would love to hear that, haha, I encourage you to go post that opinion over on their community, I bet it will go over super well, hahaha

well your poe one isn't a simulation for starters, it's more like a spreadsheet, only fancier.

Its pretty clear you don't know what exactly a simulation is or what they're good for.

Why is it that narcissistic people can't tell they are narcissistic?

Just wondering this for NO relevant reason.



I don't mean to be a complete wet blanket, but the purpose of this simulator was never to determine damage. Austar has already stated as much, but just because something doesn't immediately give you the value for the exact variable you want does not mean it does not produce valuable data.

1) It makes perfect sense to assume you would be using the Kraken club on an offhand.

2) You can calculate your average WS damage based on your set either through personally testing it, or running it through another sim that already exists.

3) You act as though this data that the simulator produces is wildly inaccurate when that just can't be argued in good faith.

Im not acting as if any of those things are that way at all.

Maybe read what I wrote, actually read it, before responding, I was pretty clear in my post.

I am merely pointing out "Hey you know there's a pretty good reason numerous people responded to your program with confusion about its application, you should be clear regarding the intent of your program, like I dunno, add an explanation, perhaps an example of how to interpret the output, etc..."

Responding with "Well they should just know!" demonstrates you dont actually care about people using your program, which proceeds to beg the question of why even post it then. If you act openly hostile towards anyone who is like "Why is this the way it is?" rather than going "Okay hold on let me add some more info to my post", then that just makes people scratch their heads about what you even are doing, bothering to post stuff here. Like whats your target audience here exactly? Its ffxiah, not stack overflow lol.

Reminds me a lot of this dude who would often post their songs up way back in the day on this forum, and he'd upload the songs and literally every single person who voiced basically any type of opinion about what they got from the song, or asked questions, etc, he would just respond super passive aggressively or openly hostile, saying ***like "Well you should just get it" or "Clearly the song wasn't intended for you then" and ***like that, which quickly just boiled down to everyone just making fun of him whenever he uploaded another song.

Then eventually he just rage quit, deleted every single one of his songs and peaced out.

Like, if you can't remotely handle people wanting information about <thing> you have uploaded and instantly get super defensive and passive aggressive over something you've made, and get extremely hostile towards anyone who voices anything remotely close to an opinion or feedback, Im gonna be real.

Maybe you just aren't cut out for the internet...

dude still?

why not just make your better sim if this is such a big thing?

many grow up hearing "Those who can, do. Those who can't, teach." Well, in my family, it was "Those who can, do. Those with passion, teach. And those who can't, write about the achievements of their betters."

Less talky, more scripty. You wanna show everyone and their mother how someone else's work is wrong- then just post what the right one would be? Otherwise, keep peer review to a constructive way to grow and learn vs a constant NO U!!!!

Did you not read a word I wrote? It seems you literally read not a single word of my post.

Also: The irony is palpable, go back and read the thread.

as a person who has since day 1 openly admitted he knows jack ***about these kinds of things in terms of writing them, you're fun to read. Its interesting to watch your posts "devolve" starting around 7pm pacific and then all through the night.

A person needn't understand the topic when the emotions of one are so blatantly obvious.

So no, you didnt read my post then. You should read peoples posts before responding, I know you think you came up with some witty ***but everything you wrote is completely orthogonal to anything I said. It almost seems comes across like you accidentally posted it on the wrong thread it is so completely off base from literally anything I said.

Go back, maybe actually read.

how long does it take for you to google a $5 word for every post you make?

Hi. What seems to be the trouble here?

The usual trouble of "my idea/opinion/math/dad is better than your idea/opinion/math/dad!"