[ QUOTE ]
Of course, the reason its a spreadsheet and not just a table of numbers is because I'm assuming that most of the people interested in such a thing would not hesitate to modify the numbers to reflect their own analysis assumptions.
[/ QUOTE ]
This may be true for some people who like to dig in and explore a bit, but I wouldn't be surprised in the least if some download it and take what's presented at face value. More dangerous is those same people trying to propagate it as fact.
There's not much (if anything) you can do to prevent that, of course.
[edit: Yeah, I know, that's a little cynical.]
[ QUOTE ]
Just noticed...
The summary table shows the values for nonpositional Psi, not the relative composite values.
[/ QUOTE ]
I noticed that as well, while repositioning things. Also, while trying to see what a chart of those numbers might look like (I'm going to probably have an "experimental" chart in the second tab of the workbook just for giggles). Should be uploaded sometime late tonight.
The reason that sort of error keeps coming up as I move things around has to do with the fact that the VLOOKUP functions correctly update cell references when rows and columns move, but do not update offset values commensurately. A danger of using lookup tables in general.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
[ QUOTE ]
[ QUOTE ]
Of course, the reason its a spreadsheet and not just a table of numbers is because I'm assuming that most of the people interested in such a thing would not hesitate to modify the numbers to reflect their own analysis assumptions.
[/ QUOTE ]
This may be true for some people who like to dig in and explore a bit, but I wouldn't be surprised in the least if some download it and take what's presented at face value. More dangerous is those same people trying to propagate it as fact.
There's not much (if anything) you can do to prevent that, of course.
[edit: Yeah, I know, that's a little cynical.]
[/ QUOTE ]
I considered not releasing this version at all, because as I mentioned I really haven't had the time to carefully proof it for errors, and this version has dramatic changes from the previous version (which had some minor errors to be sure) which were almost guaranteed to generate large goofs.
Because people asked, and because I figured more eyeballs would quickly correct any errors faster, I elected to release with warnings. Generally speaking, peer review is the fastest way to generate an error-free (or nearly so) version of things like this, and so far I haven't been disappointed there.
Now, separate from that, even if the numbers and formulas in the sheet were precisely correct, it *still* can be abused in a lot of ways. Implicit in all damage mitigation calculations is a basic assumption: the calculations consider all reasonably objective and mathematically analyzable factors, and in a way that its reasonable to assume that all other factors are either relatively equal, or very obviously set aside.
We presume, for example, that people know that Electric Armor's true value is its calculated passive damage mitigation *plus* whatever mitigation that endurance drain is providing, which means its some degree higher than calculated. We presume that while quick recovery offers no direct damage mitigation benefit, its nonetheless a very strong qualitative strength of any powerset that contains it, separate from calculated damage mitigation.
The spreadsheet doesn't say which sets are "better." It attempts to quantify what the relative value is of the things it counts. Is that useful? Well, in head to head comparisons, it very often is. If I decide to compare the relative performance of Invuln and Willpower, I can say that its very likely that if Willpower has higher calculated performance than Invuln, that its very likely that the non-mitigative benefit of quick recovery is probably going to have a much greater benefit that the tohit buff in Invincibility. That's a qualitative evaluation in and of itself, but its not difficult to defend. The computed numbers reduce the size of the playing field that qualitative judgement must play in.
One last thing: if I were a game designer (technically, I'm not sure I can make that negative hypothetical statement anymore) I would recognize that my computational analysis tools had limits, and therefore I would have to be cogniscent of what those limits were. But given that, I would strive to ensure that what I *could* quantify was reasonably balanced, so that the remaining parts I couldn't quantify could at least be controlled. Whether the devs actually do that, I still consider it a useful point of reference to judge how well they designed and balanced the different powersets.
In other words, suppose I was going to make a new powerset called "shields" and it was going to be primarily defensive in nature, but it was going to have some interesting powers that were not going to have easy to compute benefits. I would still want to *know* what the difference in performance was between a comparable powerset, say SR, and what I *calculate* for this hypothetical "shields" powerset, so I knew how much gap in performance I was dealing with. I could then at least have a *target* for how powerful these new gizmo powers could and should be.
It would be useful to know if "shields" was 50% of the strength of SR, or 75%, or 90% - or even 125% - to inform my intuition as to whether the unquantifiable parts of the powerset were strong enough without being overpowered. So even when you can't quantify everything, its still useful to quantify what you can.
But that's a subtle point not everyone fully appreciates. Just because the numbers are accurate, doesn't make them complete. But just because the numbers are not complete, doesn't make them unreliable either. The numbers said Willpower was going to be quite powerful, and it turned out to be exactly that. They also said Shields would be significantly weaker than peers, separate from its offensive abilities, and that turned out to be on the mark as well. With sufficient analysis care, the numbers tend to be dead on bullseyes when predicting relative performance.
The best I can do is state the above, so that for every person that quotes my calculations, someone else can quote my disclaimers in context.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
[ QUOTE ]
This may be true for some people who like to dig in and explore a bit, but I wouldn't be surprised in the least if some download it and take what's presented at face value. More dangerous is those same people trying to propagate it as fact.
[/ QUOTE ]
You've been reading the Inv thread in the Training Room.
[Edit for clarity.]
Kosmos
Global: @Calorie
MA Arcs in 4-star purgatory: Four in a Row (#2198) - Hostile Takeover (#69714) - Red Harvest (#268305)
[ QUOTE ]
The reason that sort of error keeps coming up as I move things around has to do with the fact that the VLOOKUP functions correctly update cell references when rows and columns move, but do not update offset values commensurately. A danger of using lookup tables in general.
[/ QUOTE ]
Haven't looked at the spreadsheet but you might want to consider switching from VLOOKUP (or HLOOKUP) to an INDEX / MATCH combination to avoid these kind of problems.
The (Shields) OwtS column (AI) hasn't been updated to the latest build on test:
<ul type="square">[*] The Defense Scales (AI21-23) are using the old +0.1 Scale per teammate to 7. They are also pointing to the wrong cell for the number of teammates (AD$16 instead of $D$15).[*] The Scale 0.5 PF base Def is missing (cell AI31).[*] The Resists are set to Scale 3.0 (S/L) and 2.0 (F/C/E/N/T/P). It should be Scale 4.5 S/L and Scale 3.0 (F/C/E/N/T) and 0 for Psi.[/list]
Also, the (Stone) Granite column (AU) is missing Stone Skin: cells AU58 and AU59 should have Scale 6.0 Res in them.
Kosmos
Global: @Calorie
MA Arcs in 4-star purgatory: Four in a Row (#2198) - Hostile Takeover (#69714) - Red Harvest (#268305)
Actually I think she's been reading the ea thread too.
I have long been a firm believer that The 'unqualifiable' factors are far more useful in powerset choices than the quantifiable ones. Quantifiable factors can, for the most part, be repaired with IO's and power pool choices, but no amount of IO slotting will give invulnerability an effective endurance drain, or willpower a damage shield.
[ QUOTE ]
[ QUOTE ]
This may be true for some people who like to dig in and explore a bit, but I wouldn't be surprised in the least if some download it and take what's presented at face value. More dangerous is those same people trying to propagate it as fact.
[/ QUOTE ]
You've been reading the Inv thread in the Training Room.
[Edit for clarity.]
[/ QUOTE ]
Yes, but I've seen it done before that thread.
Thanks for the disclaimer, I think that will be quite useful.
Of course... *looks at the length of the disclaimer again* damn! It's almost as if the spreadsheet as it's own EULA now! Hah! 
[QR]
Just a quick note: some of the latest changes I promised I haven't had a chance to make and upload yet: I haven't forgotten about them, but I've been heavily tied up with work. Usually, work affords me lots of time to keep up with the forums. At the moment, the reverse has been true. Hopefully, I'll have time to return to this (and other things) soon.
On the subject of threads, except for when I'm time-crunched I try to keep up with all the damage mitigation related ones, or at least the ones that don't go completely over the cliff into absurdity. I used to be able to say I read practically all of them, but for the last six months I've been slightly more selective due to time constraints.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Long overdue, I've updated the damage spreadsheet with the Shields corrections posted to the thread. I've also added a completely worthless sheet that charts the performance of the sets based on the main sheet's settings, and another slightly less worthless sheet that will calculate the diminshing returns of any attribute's Cur, Res, or Str for any archetype which I put practically zero time into making either pretty or pretty useful. It just works, at least as far as I could check. I'm afraid that's the best I can do at the moment: time constraints and general laziness combine to make it difficult to add more features for now.
I'm still thinking about updating this behemoth with some new ideas about how to measure survivability more accurately. The other day, I was thinking about attempting to quantify the true proportions of both the damage types and the attack types, but that got very complex very quickly - not because of the huge amount of data, that would just be tedious. More precisely the complexity of how attacks actually work: you might have cold defense and resistance, but there aren't just pure cold attacks. There are Ice Blasts which can be typed both smashing and cold as an attack, and have both smashing and cold damage. The tricky part is that while each damage component is resisted separately, which means the resistance numbers are independent, the entire attack is defended by the best of your smashing or cold defense, which means you don't just get smashing defense + smashing resistance and cold defense + cold resistance: you actually get the better of smashing or cold defense + smashing resistance, and the better of smashing or cold defense + cold resistance. You're usually a little better off than the component-by-component analysis suggests; sometimes a little more, sometimes a lot more (this affects sets like Willpower and especially Ice that have differing typed defense and resistance).
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
[ QUOTE ]
Just noticed...
The summary table shows the values for nonpositional Psi, not the relative composite values.
[/ QUOTE ]
I just noticed this, and it explains why Dark Armor was looking so uber (with its high Psi resist).
Interesting is that Shields winds up performing right around Invuln, and Electric is bottom of the barrel by significant factor.
Global @Diellan - 5M2M
Mids' Hero/Villain Designer Lead
Virtue Server
Redside: Lorenzo Mondavi
Blueside: Alex Rabinovich
Got a Mids suggestion? Want to report a Mids bug?