[ QUOTE ]
Could be anything from QA testing aids to predictive analysis to a nasty bug in the compiler. While I doubt Cryptic runs by Zen programming, it could be any one of a thousand reasons.
[/ QUOTE ]
Right, there are a number of commands that normal users don't have access to, that probably use those tables for certain debug related things. That said, it doesn't mean that the tables are necessarily kept up to date, at all. In fact, I've seen some games that I'm known for having "intervened upon" that included similar tables clientside for similar reasons, and eventually use of those tables was abandoned, as was updating them, but they were left around for legacy reasons.
I think the tables are nice, but one should always question their use. Its like this... if god came down from the heavens and handed scientists a complete map of DNA strands for all living organisms on Earth, do you think the scientists would trust it outright or verify it against what they already know?
[ QUOTE ]
I think the tables are nice, but one should always question their use.
[/ QUOTE ] I agree; and so far, the client data has shown to be pretty darn accurate. I've yet to come across a case where it was flat out wrong.
[ QUOTE ]
Its like this... if god came down from the heavens and handed scientists a complete map of DNA strands for all living organisms on Earth, do you think the scientists would trust it outright or verify it against what they already know?
[/ QUOTE ]
Hmmm...so since I'm giving all the forum-goers (scientists) all the game mechanics data with my Power Data Standardization (complete map of DNA), does that make me God? 
....or maybe just God's translator... 
Actually I was referencing the client as god - don't let it go to your head 
That's a valid point - Trust, but Verify.
But (and point me the discrepency if I missed it) until we find a testable - and well outside of margin of error - issue between Iakona's numbers and yours, we might as well look at them.
I mean, I don't think the scientists would throw away a Diety's code to the human DNA simply because it uses funky labels.
[ QUOTE ]
[ QUOTE ]
So there's no reason to believe that player tohit and critter tohit are necessarily coupled.
[/ QUOTE ]Except that they both have identical combatmod tables. Yes, they MIGHT operate differently, or use some other hidden, server-side tables, or any number of other explanations; however, the existence of the identical tables is sufficient for me to believe that they may be coupled. Also, the spreadsheet lists player and critter values as identical, rather than following pohsyb's numbers for player to-hit from +1 to +4.
[/ QUOTE ]
Now that I think about it, my recollection of Circeus' spreadsheets (don't have them handy at the moment) was that they only showed villain tohit, not player tohit at all, because they were damage comparison sheets (there was no need to know player tohit for those calculations at all). They scaled villain tohit for even con to +4 villains to show the effects of higher tohit against defense. Are we thinking of the same spreadsheets?
[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.)
The spreadsheet I'm referring to is identical to the one Circeus linked to in this thread.
[ QUOTE ]
In fact, the very fact that the tables exist suggests that they might used additively: tables are often used to "precalculate" multiplicative formulas. Sometimes they are used in true multiplicative scalers (when there is little choice), and sometimes not.
[/ QUOTE ]
Well, okay I wanted to go back and comment on this a second. I'm a programmer. In various applications I've worked on (specifically I can think of printing and fingerprint inspection) I've used similar tables to the ones iakona is posting. The thing is that in they're text form they are not the table that's used. They are generally used to populate the actual LUTs used when the program is running.
Why? Because there's no fast, quick way to look up somethign in a textual table programatically. So they generally get read in and stored as LUTs and/or Hash tables.
Why then have textual tables? Simple, ease of update. If you program (hard code) all those tables by hand, then any changes will require a recompile of your code. Keep your tables in text files, and there's no need for that.
The fun part then becomes this. Tables when read in only need be guidelines, only part of them need be used, or their numbers can be formulated when populating a LUT, etc. The sky is the limit there.
This is again, why I say the tables can't always be taken at face value. And that in terms of how they're used, there can be a higher formulaic truth to it - the numbers need not be used verbatim.
As for the (2 - X) formula to arrive at the 0 to +10 numbers? That was a simplification of (1 + (1 - X)) and I probably should have stated that because conceptually it demonstrates that the table is somewhat a reflection of itself around zero. To me that speaks volume because to me both Math and Programming are art, and that sort of use of numbers is artistic. And that feeling leads me to feel that its right - especially since it matches numbers that were previously used.
[ QUOTE ]
But (and point me the discrepency if I missed it) until we find a testable - and well outside of margin of error - issue between Iakona's numbers and yours, we might as well look at them.
[/ QUOTE ]
So then you're agreeing with me? Because that's what I said.
[ QUOTE ]
Why then have textual tables? Simple, ease of update. If you program (hard code) all those tables by hand, then any changes will require a recompile of your code. Keep your tables in text files, and there's no need for that.
[/ QUOTE ]
Or, you could come from old school Unix, where its almost genetically imprinted on you to store values in text form, and convert them into more appropriate forms at runtime. There's a reason why SMTP, HTTP, IMAP, and all the old school protocols are human-readable, and it has nothing to do with updating per se, and everything to do with making things look transparent to human eyeballs.
If it was just about keeping the values in a direct binary form, you could always do the equivalent of storing them in external flatfile databases (i.e. dbms) or numeric literals, neither of which would be directly human readable. I notice this being one of the major differences in style between programmers, and there is no middle ground.
[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.)
Argh...I'm trying to write a guide for Energy/Energy Brutes, and I want to demonstrate how the Defense scaling change will affect them.
I was planning to use a Level +0 boss and level +4 boss with a hypothetical attack against which the defender had a 30% Defense, under both current and I7 conditions...but I can't hack the math in the guide. When I did the 65% base to hit minus 30% Defense thing, I got 35% chance for the boss to hit you...but when I did the numbers given in one of the responses here, I got 26%, and I thought that the numbers for level +0 bosses were supposed to be remaining the same. I'm confused.
Could someone perhaps do those four calculations for me, with a note as to what each figure is?
[ QUOTE ]
... I thought that the numbers for level +0 bosses were supposed to be remaining the same. I'm confused.
[/ QUOTE ]
The numbers only stay the same for people with zero defense. An even level boss will have a +50% accuracy buff instead of a +25% To Hit buff. Which means a +0 boss currently hits someone with 30% defense ((50+25)-30=75-30=) 45% percent of the time, but after I7 he'll hit ((50-30)*1.5=20*1.5=) 30% of the time. Someone with no defense will still get hit 75% of the time.
[ QUOTE ]
The numbers only stay the same for people with zero defense. An even level boss will have a +50% accuracy buff instead of a +25% To Hit buff. Which means a +0 boss currently hits someone with 30% defense ((50+25)-30=75-30=) 45% percent of the time, but after I7 he'll hit ((50-30)*1.5=20*1.5=) 30% of the time. Someone with no defense will still get hit 75% of the time.
[/ QUOTE ]
Wait, what? Arcana's guide says that +0 bosses have a 65% chance to hit, not 75%.
75% is the old number, Pre-I5
[ QUOTE ]
75% is the old number, Pre-I5
[/ QUOTE ]
My bad. I got the formulas right, though.
So Bosses exchange their +15% to hit buff for a +30% accuracy buff. Which means the boss goes from an I6 (50+15)-30=65-30=35% chance to hit to an I7 (50-30)*1.3=20*1.3=26% chance to hit against R_M's proposed 30% Defense character.
So to R_M, yeah, you were doing everything right.
Well, looks like my post on 03/07/06 matches up exactly with Statesman's post today regarding accuracy/tohit. On Live servers, the first CombatMod table is multiplied by AttribBase for ToHit to find the BaseToHit by relative level and class. The second CombatMod table is multiplied by the power effect to find the effectiveness by relative level (for things like debuffs, damage, mezzes, etc.).
<font class="small">Code:[/color]<hr /><pre>BaseToHit, by Relative Level of the Target
Rel Lvl TH Mod Min BTH Lt BTH BossBTH AV BTH
14 0.01 0.5% 0.6% 0.7% 0.8%
13 0.02 1.0% 1.2% 1.3% 1.5%
12 0.03 1.5% 1.7% 2.0% 2.3%
11 0.04 2.0% 2.3% 2.6% 3.0%
10 0.05 2.5% 2.9% 3.3% 3.8%
9 0.08 4.0% 4.6% 5.2% 6.0%
8 0.15 7.5% 8.6% 9.8% 11.3%
7 0.33 16.5% 19.0% 21.5% 24.8%
6 0.45 22.5% 25.9% 29.3% 33.8%
5 0.55 27.5% 31.6% 35.8% 41.3%
4 0.64 32.0% 36.8% 41.6% 48.0%
3 0.73 36.5% 42.0% 47.5% 54.8%
2 0.81 40.5% 46.6% 52.7% 60.8%
1 0.9 45.0% 51.8% 58.5% 67.5%
0 1 50.0% 57.5% 65.0% 75.0%
-1 1.05 52.5% 60.4% 68.3% 78.8%
-2 1.1 55.0% 63.3% 71.5% 82.5%
-3 1.15 57.5% 66.2% 74.8% 86.3%
-4 1.2 60.0% 69.0% 78.0% 90.0%
-5 1.25 62.5% 71.9% 81.3% 93.8%
-6 1.3 65.0% 74.8% 84.5% 97.5%
-7 1.35 67.5% 77.7% 87.8% 101.3%
-8 1.41 70.5% 81.1% 91.7% 105.8%
-9 1.48 74.0% 85.1% 96.2% 111.0%
-10 1.68 84.0% 96.7% 109.2% 126.0%
-11 1.8 90.0% 103.6% 117.0% 135.0%
-12 1.95 97.5% 112.2% 126.8% 146.3%
-13 2.1 105.0% 120.8% 136.5% 157.5%
-14 2.2 110.0% 126.6% 143.0% 165.0% </pre><hr />
<font class="small">Code:[/color]<hr /><pre>Power Effectiveness, by Relative Level of the Target
Rel Lvl Eff Mod
14 0.01
13 0.01
12 0.01
11 0.02
10 0.03
9 0.04
8 0.05
7 0.08
6 0.15
5 0.3
4 0.48
3 0.65
2 0.8
1 0.9
0 1
-1 1.11
-2 1.22
-3 1.33
-4 1.44
-5 1.55
-6 1.66
-7 1.77
-8 1.88
-9 2
-10 2.1
-11 2.2
-12 2.3
-13 2.4
-14 2.5 </pre><hr />
If you want something to do all those messy I7 accuracy calculations for you, look no further than my sig for The Sheet!
Note: The sheet that is currently available has not yet been modified for this
[ QUOTE ]
In addition, higher ranks have an inherent resistance to To Hit Debuffs (.1 for Lts., .2 for Bosses, .3 for AVs).
[/ QUOTE ]
new piece of information. All defense calculations will be correct, but ToHit debuff calculations for LTs, Bosses, and AVs will be incorrect. This will hopefully be rectified by the end of the week (4/9/06).
Yeah, I know it's a shameless plug, but it's a relevant and useful shameless plug.
[ QUOTE ]
Well, looks like my post on 03/07/06 matches up exactly with Statesman's post today regarding accuracy/tohit. On Live servers, the first CombatMod table is multiplied by AttribBase for ToHit to find the BaseToHit by relative level and class. The second CombatMod table is multiplied by the power effect to find the effectiveness by relative level (for things like debuffs, damage, mezzes, etc.).
[/ QUOTE ]
Hold that thought.
[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.)
Looking back on Statesman's post, his numbers for the current tohit calculations don't match up with his explanation. Using the numbers he posted, it looks like ToHit is
<font class="small">Code:[/color]<hr /><pre> RankToHit + (CombatMod - 1) </pre><hr /> for a higher level target. This still doesn't match up with the tested values in your guide, though.
[ QUOTE ]
Looking back on Statesman's post, his numbers for the current tohit calculations don't match up with his explanation. Using the numbers he posted, it looks like ToHit is
<font class="small">Code:[/color]<hr /><pre> RankToHit + (CombatMod - 1) </pre><hr /> for a higher level target. This still doesn't match up with the tested values in your guide, though.
[/ QUOTE ]
Yes, this is the correct formula. Its corrected in my Guide, but for a specific reason I haven't posted the updated version yet.
Rule #1 is that if it boosts all your attacks, its a tohit buff/debuff, and if it alters a single power, its an accuracy buff/debuff.
Rule #2 is that all tohit buffs are additive, and all accuracy buffs are multiplicative.
The level scaler appeared to be a contradiction in that rule, so a while ago I investigated around the edges, and it turns out the multiplicative rule we've all (or mostly) assumed is wrong - its additive just like all other tohit adjustments, and as you've discovered.
There is a natural conclusion that comes out of this: if the I6 level scaler is additive (tohit), and the I7 one is multiplicative (acc), then the previously operative statement about the I7 change:
"Players with zero Defense will see no change from I6 to I7"
is not *exactly* correct. That's only true for minions. Its not true for any higher rank villain: higher ranked villains will be very slightly more accurate in I7 than in I6, against players with zero defense.
[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.)
Hmm, how to put this...
My current testing of I7 is not inconsistent with critters getting +10% accuracy per level, to +5, and then holding at +50% acc and getting tohit buffs from +6 higher.
Its also not inconsistent with LTs, Bosses, and AVs getting 1.15 (+15%) 1.3, and 1.5 accuracy bonuses on top of that.
And its not inconsistent with accuracy bonuses being multiplicative: a +2 Boss would have net accuracy of 1.2 x 1.3 = 1.56
And it seems Iakona is going to have difficulty confirming those numbers.
That is all for now.
[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.)
I await your collective discoveries eagerly. I have been assembling this math into a set of programming libraries so that I can quickly produce numbers for arbitrary situations. I'm looking forward to being able to "finalize" the tables and math involved in a useful way.
Blue
American Steele: 50 BS/Inv
Nightfall: 50 DDD
Sable Slayer: 50 DM/Rgn
Fortune's Shadow: 50 Dark/Psi
WinterStrike: 47 Ice/Dev
Quantum Well: 43 Inv/EM
Twilit Destiny: 43 MA/DA
Red
Shadowslip: 50 DDC
Final Rest: 50 MA/Rgn
Abyssal Frost: 50 Ice/Dark
Golden Ember: 50 SM/FA
[ QUOTE ]
I await your collective discoveries eagerly. I have been assembling this math into a set of programming libraries so that I can quickly produce numbers for arbitrary situations. I'm looking forward to being able to "finalize" the tables and math involved in a useful way.
[/ QUOTE ]
Last night, I discovered how long a level 38 Dark Armor brute gets to stay alive in a full team of level 40s doing a Mayhem mission full of 42s: about 20 seconds at a time.
[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.)
SR defenses are typed melee, ranged, and AoE. This means they each protect only against attacks typed as such. The assumption was that (except for Hamidon), all attacks have at least one of those three types, and therefore SR defenses work against all possible attacks.
Apparently not.
Note: appears to be a PvP issue only, because it only seems to have occured for player versions of the powers in question.
[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.)
Could you please clarify how power and enhancement accuracy combine? Do they add or multiply?
For example, Dominate has 1.2 accuracy. If you add two white accuracy SOs, is the final to-hit (1.2 x 1.67) = 200% of base, or is it (1.2 + 0.67) = 187% of base?
EDIT: Here's why I'm confused about this. The defense formula is (BaseAcc) * (AccuBuffs) * (other stuff). The explanation of the formula first states that power and enhancement accuracy are both "accuracy buffs" (AccuBuffs?) but then later it refers to power accuracy as "base accuracy" (BaseAcc?).
Let P = power accuracy (e.g., 1.2 for Dominate), E = enhancement accuracy (e.g., +67% for two SOs), S = the subtotal after other modifiers, and H = net to-hit. Which is the correct formula?
H = (P + E) x S
H = P x (1 + E) x S
RE-EDIT: OK, I scoured the thread again, and example of a Martial Arts scrapper with 2 SOs confirms the (P x (1 + E) x S) formula. Assuming the example is correct, could you please update the next version of the guide with an example or a better explanation? (Also, note that Castle recently confirmed that MA attacks only have 1.05 accuracy, not 1.1.)
[ QUOTE ]
I can't see any good reason for the client to need some of of the data that it has. Maybe some Dev was being lazy, and didn't bother to split the data into server-only and public parts. Who knows.
[/ QUOTE ]
Maybe so people like _Castle_ don't need a full server install just to dig through a couple variables?
Could be anything from QA testing aids to predictive analysis to a nasty bug in the compiler. While I doubt Cryptic runs by Zen programming, it could be any one of a thousand reasons.