Stats and how they work.

[Drexel]

Despite everything being a couple of months or less from blowing up, I continued to do some updates on the Weapon Comparison sheet I made.

I cleaned it up so hopefully it's clearer to the user as to where to enter data. Basically Orange cells, put in the Weapon's Stat's in the weapon area and your two primary stats in the Character Stats area.

Silvano made the point that I wasn't taking into consideration the weapon's full potential because I wasn't looking at the stat caps by weapon - so he added in those calculations.

I also added a section for you to put your own two primary stats and see how that effects weights against the stat caps. This will also change the comparison for the weapon itself, the only time attribute stats on a weapon are calculated are when your stats are entered (since they are relatively and you could be at cap).


Stat caps were calculated using a logarithmic function based off of all the stat caps I knew, it's close but not perfect - it'll have to do until someone figures out the actual stat-cap by weapon formula (from every cap I know, I'm more inclined to believe it's a set value / hidden stat but within certain ranges). I thought about adding a place for user over-ride for stat caps but for general usage it's simply not worth it - using the curve gives a close enough value.

There seems to be some pretty large errors in your stat cap prediction equations, for instance if you put in the 36.74 DPS value of the Ifrits bow your spread sheet predicts a 303 PIE cap and a 283 DEX. We know from extensive testing that these values are incorrect. Furthermore inputting a DPS value of 20 results in predicted stat caps of -14 Primary stat and 111 Secondary stat. I suppose the negative number can be explained by an unknown floor so I suppose this does not innately disprove the function with respect to DPS values end gamers actually care about.

I’m also curious as to why you use a logarithmic function to predict stat caps when A) fitting a linear trend line to know data produces an R squared value of > 0.99 and B) everything else we know about stats in this game tends to show they relate to each other in a linear fashion. That being said I do applaud your work in testing and sharing your spread sheets with the public. I have made some of my own however have not shared them with anyone outside my LS, I too am still struggling with the missing link of how DPS affects stat caps.

[Izodius]

There seems to be some pretty large errors in your stat cap prediction equations, for instance if you put in the 36.74 DPS value of the Ifrits bow your spread sheet predicts a 303 PIE cap and a 283 DEX. We know from extensive testing that these values are incorrect. Furthermore inputting a DPS value of 20 results in predicted stat caps of -14 Primary stat and 111 Secondary stat. I suppose the negative number can be explained by an unknown floor so I suppose this does not innately disprove the function with respect to DPS values end gamers actually care about.

I’m also curious as to why you use a logarithmic function to predict stat caps when A) fitting a linear trend line to know data produces an R squared value of > 0.99 and B) everything else we know about stats in this game tends to show they relate to each other in a linear fashion. That being said I do applaud your work in testing and sharing your spread sheets with the public. I have made some of my own however have not shared them with anyone outside my LS, I too am still struggling with the missing link of how DPS affects stat caps.

Honestly for advanced users the stat caps should be manually entered; I just did that for the masses.

I used a logarithmic curve because I found the linear curve to be wildly inaccurate based on testing I had of Relic weapons and Garuda weapons stat caps (from a third party) - especially in regards to secondary stat caps. It also doesn't fit THAT well with Mog vs Ifrit caps.

I say on the sheet that Bows are not accurately accounted for - they can be used if you input the stat caps yourself, but it's still probably not the best - bows have some oddities about them relative to melee weapons and I simply haven't played with them enough to make better calcs for them. From all the data I've seen and played with recently - I'm lead to believe stat caps are calculated differently based on weapon type; I also have some small evidence that your arrow for bows plays into your stat cap as well - but I don't really intend on exploring that more than what my limited conjuncture and anecdotal evidence suggests.

I've updated the curves recently - they are far from perfect and until we have more tests on caps with more weapons I probably won't readdress them; I spent a fair amount of time trying to come up with a better formula but nothing I was doing was working.

TLDR: Basically I did what I could for the general masses involving stat caps with the data I had - I'd fully expect advanced users to manually enter the caps.

Thanks for the Kudos - a lot of people seem to be using the sheet but no one has seemed to throw out any thanks for the time I put into it. I don't expect it to be perfect but it's useful in it's own way.

[Churchill]

iirc bows are based on highest level arrow for the bow, so all relevant bows damage is based on cobalt

[Raldo]

Just posting to confirm the above statement. From the 1.22 patch notes:

[dev1380] The average amount of damage a weapon is capable of dealing per second (DPS) has been added to help text.

The DPS of a bow is calculated in the following manner:
DPS = (Damage + Damage value of the strongest ammunition at the bow's required level [excluding HQ]) / Delay

[Izodius]

iirc bows are based on highest level arrow for the bow, so all relevant bows damage is based on cobalt

That appears to be correct, the issue is moreso that Bows Stat Caps are on a completely different curve / line than Lance/Axe/Fists. All 4 actually appear to be on slightly different curves/lines.

[Meatdawg]

I want to say thank you for your work Izodius

You should convert the sheet into a calculator. You can make it web based or write a VB/java program for people to dl. Let them enter their stat caps, pick their appropriate job and you then list the top 5-10 weapons for their class.

You could turn on advance settings where they add melds to meldable gear, or they pick their GC or even let them alter your table of weapons. Shoot it over to ffxiv pads for the masses and if you have a page you can take donations for all your hard work.

I could help with a app but I think a webpage would be the easiest.

[Izodius]

I want to say thank you for your work Izodius

You should convert the sheet into a calculator. You can make it web based or write a VB/java program for people to dl. Let them enter their stat caps, pick their appropriate job and you then list the top 5-10 weapons for their class.

You could turn on advance settings where they add melds to meldable gear, or they pick their GC or even let them alter your table of weapons. Shoot it over to ffxiv pads for the masses and if you have a page you can take donations for all your hard work.

I could help with a app but I think a webpage would be the easiest.

Ha, thanks in a world where I was younger, more motivated, had more time on my hands, and there was no 2.0 looming - I might do that. I don't need donations - it was simple work anyone could have done and I had fun doing it, and it didn't take much time really. Silvano, Kaeko, and Seikin are the ones who really deserve credit - their tests allowed me to write simple calculations to compare.

I never expected people to actually use it, I made it mostly for myself originally - which explains why there's still a lot of holes in it.

I'll probably move the sheet to a downloadable excel file on a share site rather than Googledocs - with so many people hitting it daily now I'm having to go through every night and fix small things people have done (thankfully the calculations are protected and I have multiple backups).

I'll probably release my Enmity, Gear Set, and Item slot comparisions to the public this week as well - they're VERY simple but they may have some usefulness to someone prior to 2.0.

When 2.0 hits, I might go back and redo this with some of those suggestions. I've thought about them and they shouldn't be THAT hard to do realistically. It'd probably require Excel over Googledocs, since Excel is a bit friendlier and I could lock cells with it.

[Meatdawg]

2.0 is still 5 months out (assuming the latest interview was referring to the PC and not the ps3) so I am sure it will get plenty of use till then.

You are correct thought that everything will be up in the air again once 2.0 hit. Hopefully some closed beta testers can leak some parse results from stat min/maxing. Will be fun to calculate everyone over again.

[Kurokikaze]

I'm more hoping that one of the great people behind our current parsers make an addon during 2.0 beta. That is assuming that addon functionality is allowed to be tested then.

[Churchill]

If I input values for Ifrit's Axe and Ifrit's Axe

Then set player stats to say, 270/270

Then modify AP to 150 for one axe, and +60STR/30VIT in the other set

It shows the AP setup to be vastly better than capping your stats, yet everything we've seen from testing seems to indicate otherwise. Unless I'm not understanding something, doing what I did should replicate the difference between being stat capped and not. Is this an error in calculation or am I misunderstanding the capabilities this is supposed to produce?

[Churchill]

actually just went out and tested and attack power shits on dex for brd lol

[Izodius]

If I input values for Ifrit's Axe and Ifrit's Axe

Then set player stats to say, 270/270

Then modify AP to 150 for one axe, and +60STR/30VIT in the other set

It shows the AP setup to be vastly better than capping your stats, yet everything we've seen from testing seems to indicate otherwise. Unless I'm not understanding something, doing what I did should replicate the difference between being stat capped and not. Is this an error in calculation or am I misunderstanding the capabilities this is supposed to produce?

You're doing something wrong or you're not looking at the right column.

At 270/270 stats, Ifrit's Axe with 150 AP is 1,079.9.

With 90 STR and 30 VIT it's 1,130.1. As expected Primary and Secondary stats under the cap are significantly stronger.

This is a sheet with those exact numbers entered:

https://docs.google.com/spreadsheet/...GRVcDl6N2F4cFE

AP's strength comes from the fact that it increases both WS damage (buy up to +.42 according to Silvano's testing) and it also effects Auto-Attack damage which as a community we always write off as being nothing, but it's generally at least 30% of your damage; for BRD/ARC it's closer to 45% of your damage.

That said, AP will always be behind the primary stat until the cap (especially primary stat STR). The secondary stat is can be weaker than AP but it's usually not. I would recommend you ensure you're using the sheet correctly before drawing any conclusions.

You can also change the weights of the stats based on your own, or other's testing (such as raising AP to .42 to be more consistent with Silvano's testing).

Currently the weights are at:

0.38 Attack Power (AA+WS)
0.85 STR (Under WS cap WS only)
0.66 2ND Stat (Under WS stat cap WS only)

One of the (many) problems with using the sheet for ARC/BRD is that STR is not their primary stat, and as such their primary stat gets overweighted slightly if you use the sheet - because the sheet assumes your primary stat is STR and thus has to account for the AP gained from STR.

[Churchill]

Makes a bit more sense then. Damage still favors AP over secondary stats when I played with those numbers. What does the far right column represent exactly?

[Izodius]

Makes a bit more sense then. Damage still favors AP over secondary stats when I played with those numbers. What does the far right column represent exactly?

The far right column (V) is the weapon's weight with your entered primary/secondary stats. It's the best indicator of the actual weighted value of the weapon (while Column T is the weapons highest value in a capped stats from other gear situation). The reason why AP looks so strong in Column O (which is just the weapon's weight) is that primary stats are only considered when they relative to your stats because there are caps on them and their value is relative. They are not considered in column O (except for the AP gain from STR gain) - AP also effects auto-attack weight (calculated under all weights) since it increases AA damage, which makes it inherently stronger in column O.

I've added notation to the sheet to make that clearer.

[Therin]

What's the consensus on BRD stats? I assumed it was all DEX/PIE but I'm hearing that there should be some focus on STR once DEX/PIE are past 300. Any truth to that? Mine is 319 DEX, 310 PIE right now.

[Raldo]

STR will only help you in that it will grant you some minor amount of attack power (I believe the going theory is 2atk per 3 STR). If you are capped primary/secondary, then you're better off just going for attack power rather than STR.

[Therin]

Gotcha, that makes sense. How do I know the caps for DEX/PIE exactly?

[Churchill]

well it's PIE/DEX, so you have those backwards. I went and did some brief testing last night to compare my set vs someone elses

My stats were 330pie/290 dex and 505AP
His stats were 271pie/271dex and 650AP

Obviously my WS damage was much higher, that's a given with capped PIE, but his light shot damage was much much higher. I've changed my gear now to where it's sitting at 330PIE/270DEX and 590AP and will see how that parses on Garuda vs what I was doing previously

[miokomioko]

Your stat caps depend on the bow you use, once you know those and legitimately hit them, sure STR would be another option (assuming you can't put other stats there like pure Attack Power and/or Accuracy).

[Therin]

well it's PIE/DEX, so you have those backwards. I went and did some brief testing last night to compare my set vs someone elses

My stats were 330pie/290 dex and 505AP
His stats were 271pie/271dex and 650AP

Obviously my WS damage was much higher, that's a given with capped PIE, but his light shot damage was much much higher. I've changed my gear now to where it's sitting at 330PIE/270DEX and 590AP and will see how that parses on Garuda vs what I was doing previously

So I should throw another PIE ring on then, that's an easy fix. 590-650 AP is pretty high, mine is at 509 right now. Thanks for the info, I'll take that into account and make some gear changes.