Impact of Offensive AAXP on DPS

[Xheng]

Here's another useful way to look at relative impact of AA spendout.

The following evaluation assumes the following baseline values for damage output, when comparing relative impact of abilities. If your damage output is relatively more or less than this, then adjust accordingly. The percentage impact of abilities will still be as shown in the table, but the absolute DPS increase will be different.

Assumes:
. DoT Base Damage = 350 dps
. DD Base Damage = 150 dps
. Pet Base Damage = 100 dps

Index:
. DPS/AA = Actual numerical gain in damage per second, per AA spent. Higher values are better.
. AA = Number of AA required to buy this ability
. DOT/DD/Pet = Percentage increase to base DPS for that damage medium.

Critical Affliction * * *DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.CA1 * * * * * * * * * * *3.50 * ** *3 * * * 3.0 * * * * * * *
.CA2 * * * * * * * * * * *1.75 * ** *6 * * * 3.0 * * * * * * *
.CA3 * * * * * * * * * * *1.17 ** * *9 * * * 3.0 * * * * * * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Imp.Critical Affliction *DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.CA4 * * * * * * * * * * *3.50 * ** *3 * * * 3.0 * * * * * * *
.CA5 * * * * * * * * * * *1.75 * ** *6 * * * 3.0 * * * * * * *
.CA6 * * * * * * * * * * *1.17 ** * *9 * * * 3.0 * * * * * * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Spell Casting Fury * * * DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.SCF1 * * * * * * * * * * 1.50 * ** *2 * * * * * * * 2.0 * * *
.SCF2 * * * * * * * * * * 0.75 ** * *4 * * * * * * * 2.0 * * *
.SCF3 * * * * * * * * * * 0.75 ** * *6 * * * * * * * 3.0 * * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Fury of Magic * * * * * *DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.FoM1 * * * * * * * * * * 1.00 ** * *3 * * * * * * * 2.0 * * *
.FoM2 * * * * * * * * * * 0.50 ** * *6 * * * * * * * 2.0 * * *
.FoM3 * * * * * * * * * * 0.33* * * *9 * * * * * * * 2.0 * * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Fury of Magic Mastery * *DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.FoMM1 * * * * * * * * * *0.50 * ** *3 * * * * * * * 1.0 * * *
.FoMM2 * * * * * * * * * *0.25 * ** *6 * * * * * * * 1.0 * * *
.FoMM3 * * * * * * * * * *0.17 ** * *9 * * * * * * * 1.0 * * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Destructive Fury * * * * DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.DestF1 * * * * * * * * * 0.35 * ** *3 * * * * * * * 0.7 * *
.DestF2 * * * * * * * * * 0.20 ** * *6 * * * * * * * 0.8 * *
.DestF3 * * * * * * * * * 0.15 ** * *9 * * * * * * * 0.9 * *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Death's Fury * * * * * * DPS/AA * * *AA * * *DoT * * *DD * * *Pet
.DF1 * * * * * * * * * * *0.22 * ** *3 * * * * * * * * * * * 0.65
.DF2 * * * * * * * * * * *0.22 ** * *3 * * * * * * * * * * * 0.65
.DF3 * * * * * * * * * * *0.22 ** * *3 * * * * * * * * * * * 0.65
.DF4 * * * * * * * * * * *0.22 ** * *3 * * * * * * * * * * * 0.65
.DF5 * * * * * * * * * * *0.22 ** * *3 * * * * * * * * * * * 0.65
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Quickening of Death * * *DPS/AA * * *AA * * *DoT ** * DD * * *Pet
.QoD1 * * * * * * * * * * 0.32 ** * *5 * * * * * * * * * * * 1.60
.QoD2 * * * * * * * * * * 0.32 ** * *5 * * * * * * * * * * * 1.60
.QoD3 * * * * * * * * * * 0.32 ** * *5 * * * * * * * * * * * 1.60
.QoD4 * * * * * * * * * * 0.32 ** * *5 * * * * * * * * * * * 1.60
.QoD5 * * * * * * * * * * 0.32 ** * *5 * * * * * * * * * * * 1.60


In general, if you're looking for the most bang for the buck, go for Critical Affliction first, at least through level 5. Then go for DD aa's through about FoMM2 unless you really don't use them much for dps. If you use pets more for DPS, max out QoD before considering Death's Fury.

Not included in this list are the Burst Damage Abilities: Lifeburn, Swarm of Death (level 3), and Army of the Dead (level 1). All of these are nice burst damage, and personally I would put them roughly after FoM1 in terms of overall damage impact.


This way of analyzing data was inspired by an old post on necro.eqclasses.com by someone whose name I now unfortunately forget. Information on AA costs and percentage impact taken from various places around the net, and may not be precisely correct. Please post if you see errors.


Edit 11-Jun-2005: Corrected cost for FoMM series. Original costs, taken from Graffe's, were too low.

[Xislaben]

I'm not sure I really follow you here, buts since I use dots almost all the time I figured doing CA would give me the most bang for the buck, though when clearing trash on raids I'm lucky to get a single nuke in anyways, so dots on raid mobs do even less unless it's a high hp mob like the raid target itself.

[Xheng]

Originally posted by Xislaben@Jun 6 2005, 02:28 AM
I'm not sure I really follow you here
Here's how the calculation works. Using Critical Affliction 2 as an example...

DoT Base Damage = 350 dmg/sec
Critical Affliction 2 modifier = 3.0 % (that's the incremental improvement from this AA... brings it to a total of 6%)
AA Cost of CA2 = 6

DPS Increase from buying CA2 = (350) * (0.03) = 10.5 dmg/sec
DPS/AA = 10.5 / 6 = 1.75 DPS/AA

Essentially this says that if you purchase the ability, and your base DoT damage was 350 dps, then you'll gain another 10.5 dps from CA2 for all future fights. On average, of course. Thus every AA you spend for CA2 delivers 1.75 dps on every future fight you have.

If you, instead, spent those 6 AA on FoM2, then on average you would be adding only 0.50 dps on future fights, per AA spent. So CA2 is a better choice.

It's just a way of comparing 'bang for the buck' numerically.


If your playstyle was such that typical damage was much higher loaded towards nukes than DoTs for some reason, then you could come up with your own personalized DPS/AA factors, were you so inclined. The percentage gains in the table won't change, but the relative value will.

[Brahman]

all of your numbers are wrong because you are basing them on the assumption that the gains are equally split between the tiers. IE CA1 = 3% CA2 = 6% CA 3 = 9% when the numbers just don't work like that. (if my memory serves me correctly the numbers are 2, 5, 9)

So, unless you have someplace with the parses for all the exact numbers just throw throw out that chart.

EDIT::: looking closer it seems perhaps you have the actual numbers for the DD AAs so that data may be correct. but i know crit affliction, and, i'm pretty damn sure, all the pet AAs are NOT straightline upgrades.

[lesk]

I think his numbers are right. I hear 3-6-9 from most people but I think that comes from the secret area of the farming guide site and I think it was mathematically shown, not parsed.

I'm parsing 3% from CA1 (slightly more actually) but I'm not completely confident that my method is accurate. The only thing that bothers me is Splurt. I ignore it completely and I think that because it is effectively removed from the sample it shouldn't be a problem. Since I'm not a statitition (see, I can't even spell it) it gives me doubt. I'm going to go back and handle Splurt properly anyway....at least I would feel better.

I'm not too sure about sample size either.


Total dot ticks: 8295
Total crit ticks: 271
Percent occurance: 3.26702833031947
Total dot damage: 1877000
Crit damage: 61980
Percent damage increase: 3.30207778369739


I know that doesn't answer the gains per level question, but maybe soon.

[Felicite]

I just want to thank you for the analysis, it has always been very helpful in making AA spending decisions.

[chabeksis]

Originally posted by Brahman@Jun 6 2005, 06:24 AM
all of your numbers are wrong because you are basing them on the assumption that the gains are equally split between the tiers. IE CA1 = 3% CA2 = 6% CA 3 = 9% when the numbers just don't work like that. (if my memory serves me correctly the numbers are 2, 5, 9)

So, unless you have someplace with the parses for all the exact numbers just throw throw out that chart.

EDIT::: looking closer it seems perhaps you have the actual numbers for the DD AAs so that data may be correct. but i know crit affliction, and, i'm pretty damn sure, all the pet AAs are NOT straightline upgrades.
Even if the numbers are slightly off, the method of determining the value of the AA still seems to be valid. I ran the numbers according to the 2, 5, 9 that you suggested and CA is still the best offensive AA for necros by far assuming the 350 DPS figure. However that being said, the parses that I have run show CA2 to be really close to 6%, but I still need a larger sample to be sure. If anyone is interested tell me the parser you want me to use and what you think is an adequate sample size and I will post my results as I move from CA2 to CA6. I will be doing this in the next two weeks so it might be a good experiment.

I would love to see more parses on this site, for SCM, CA, and the DD AAs, I think that is something that all of us should work towards. Then we would have an objective body of knowledge that would help a lot of people make better decisions about what AAs to get first.

[Xislaben]

Originally posted by lesk@Jun 6 2005, 04:06 PM
I think his numbers are right. I hear 3-6-9 from most people but I think that comes from the secret area of the farming guide site and I think it was mathematically shown, not parsed.

I'm parsing 3% from CA1 (slightly more actually)


Total dot ticks: 8295
Total crit ticks: 271
Percent occurance: 3.26702833031947
Total dot damage: 1877000
Crit damage: 61980
Percent damage increase: 3.30207778369739


Nice, I've got CA2 atm, and would be happy to parse it, what did you use? I've got Yalp and Jalp, but haven't played much with the former.

[aeriform]

I could have sworn I posted something like this already, but maybe only on the defunct forums.

"Here's a start for the chart you are wanting, using the parse values which are agreed upon by most.. assuming your nuke and dot and pet dps are all the same, which they probably wouldn't be.
Keep in mind that 1% of 300dps from dots is a lot more points than 1% of 200dps from nuking, or 1% of 70dps from a pet..

The number in bold is the % increase in dps per AA point spent for each rank: aka the bang for your buck (the higher the number, the better, if it wasn't obvious). The numbers are for the total AA costs/dps increases since each AA must be done in rank and in order for DD and DoT AA's." -- what I posted in response to the person who started the thread, but I don't remember who started the thread...

redone DD crit changes for spell casting fury, since that wasn't in effect when this thread was active

DoT AA
Critical Affliction
Level 1: 1.00, 3AA, +3% Crit Rate.
Level 2: 0.49, 6AA, +3% Crit Rate.
Level 3: 0.31, 9AA, +3% Crit Rate. (+9% total DoT dps increase for all ranks for 18AA)
Advanced Critical Affliction
Level 1: 0.92, 3AA, +3% Crit Rate.
Level 2: 0.45, 6AA, +3% Crit Rate.
Level 3: 0.29, 9AA, +3% Crit Rate. (+18% total DoT dps increase for all ranks for 36AA)


Direct Damage AA
Spell Casting Fury
Level 1: 1.00, 2AA, +2% Crit Rate, +100% crit amount. (since change from +33%)
Level 2: 0.49, 4AA, +2% Crit Rate, +100% crit amount. (since change from +66%)
Level 3: 0.48, 6AA, +3% Crit Rate, +100% crit amount. (+7% total Nuke dps increase for all ranks for 12AA)
Fury of Magic
Level 1: 0.62, 3AA, +2% Crit Rate.
Level 2: 0.31, 6AA, +2% Crit Rate.
Level 3: 0.20, 9AA, +2% Crit Rate. (+13% total Nuke dps increase for all ranks for 30AA)
Fury of Magic Mastery
Level 1: 0.29, 3AA, +1% Crit Rate.
Level 2: 0.15, 6AA, +1% Crit Rate.
Level 3: 0.10, 9AA, +1% Crit Rate. (+16% total Nuke dps increase for all ranks for 48AA)
Destructive Fury
Level 1: 0.32, 3AA, +7% Improved Crits. +1.12% dps.
Level 2: 0.18, 6AA, +8% Improved Crits. +1.28% dps.
Level 3: 0.15, 9AA, +10% Improved Crits. +1.60% dps. (+25% higher nuke crits, +4% nuke dps for all ranks of DF for 18 AA)
(+20% total Nuke dps increase for all ranks of DD AA for 66 AA total)

Pet AA
Death's Fury
Level 1: 0.22, 3AA, +0.65% dps.
Level 2: 0.22, 3AA, +0.65% dps.
Level 3: 0.21, 3AA, +0.65% dps.
Level 4: 0.21, 3AA, +0.65% dps.
Level 5: 0.21, 3AA, +0.65% dps. (+3.25% total Pet dps increase for all ranks of DF for 15AA from magician parses)
Quickening of Death
Level 1: 0.32, 5AA, +1.6% dps.
Level 2: 0.31, 5AA, +1.6% dps.
Level 3: 0.31, 5AA, +1.6% dps.
Level 4: 0.31, 5AA, +1.6% dps.
Level 5: 0.30, 5AA, +1.6% dps. (+8% total Pet dps increase for all ranks of QoD for 25AA from magician parses)
(+11.5% total Pet dps increase for all ranks of Pet AA's for 40AA)

Spell Casting Fury being the only AA that you get more for your money from on a AA/dps perspective by completing all ranks of it as opposed to just the first rank.

Assuming your dps is split evenly between your pet DoT's and DD, one best path might be like:

CA2 ==> SCF3 ==> FoM1 ==> CA3 ==> ACA2 ==> QoD5 == ACA3 ==> FoM3 ==> FoMM1 ==> DeathF5 ==> FoMM3 ==> DestF3

--------------------------------------------------------------------------
The previous assumes that you get dps from every type equally.
Let's say your dps for the situation you're most in (or concerned about) for solo kiting is broken down into:
200 dps from DoT
0 dps from DD
80 dps from Pet

Implies dps gain per AA spent of:

DoT AA
Critical Affliction
Level 1: 2.00
Level 2: 0.98
Level 3: 0.62
Advanced Critical Affliction
Level 1: 1.84
Level 2: 0.90
Level 3: 0.58


Direct Damage AA
All 0 dps gain

Pet AA
Death's Fury
Level 1: 0.18
Level 2: 0.18
Level 3: 0.17
Level 4: 0.17
Level 5: 0.17
Quickening of Death
Level 1: 0.26
Level 2: 0.25
Level 3: 0.25
Level 4: 0.25
Level 5: 0.24

In this particular case I'd go:
CA3 ==> ACA3 ==> QoD5 ==> DeathF5

[Felicite]

Originally posted by aeriform@Jun 6 2005, 12:17 PM
I could have sworn I posted something like this already, but maybe only on the defunct forums.
Wheeee: http://www.necrotalk.com/index.php?showtop...t=0&#entry31453 (http://www.necrotalk.com/index.php?showtopic=3515&st=0&#entry31453)

[aeriform]

Total dot ticks: 8295
Total crit ticks: 271
Percent occurance: 3.26702833031947

If that's correct, then that shows a 95% confidence interval with min of 2.89 and max of 3.65 . I am willing to believe that result because I've seen parses posted for 2%, 3%, and 4%, so 3% is pretty likely.

And someone can use search better than me... /poke

[Xislaben]

Originally posted by aeriform@Jun 6 2005, 08:17 PM
I could have sworn I posted something like this already, but maybe only on the defunct forums.
Are you taking into account that the first level of both CA and SCF do 200% now when they crit?

[Xheng]

Aeriform's data is the stuff that inspired my crunching... remembered seeing it on the old eqclasses forums that disappeared into the mists, but not who posted it.

One other thing to add was that I did some parsing of critical ticks way back when when I had only Critical Affliction 2. I removed all splurts because they're well nigh impossible to count as crit or not, in the early stages. This was done manually, using Excel. After going through it, it's clear to me there's gotta be an easier way... cause I'm not going through that again ;)

Spell * * * * * * * * * *Crit Ticks * Total Ticks
Blood of Thule * * * * * * * *12 * * * *254
Dark Plague * * * * * * * * * 19 * * * *259
Embracing Darkness * * * * * *15 * * ** 344
Funeral Pyre of Kelador * ** 106 * ** *1616
Horror * * * * * * * * * * * *62 * * **1062
Pyrocruor * * * * * * * * * * 30 * * * *380
Saryrn's Kiss * * * * * ** * *21 * * ** 284
Zevfeer's Theft of Vitae * * * 1 * * ** *10
total * * * * * * * * * ** * 266 * ** *4209 = 0.063

Though not definitive, that (combined with Occam's Razor about simple being probable) leads me to believe that the generally accepted 6% number is correct for it.

[lesk]

I had a slow day last week and threw together some quick and dirty perl. I'm not a coder by trade or even hobby really, so go easy on me. I'm sure there's a windows interpreter somewhere.

You'll need to adjust it to fit, of course.


#!/usr/bin/perl -wT
#
# parse crits from eqlog
# for purposes of this simple script, a crit is anything > 2*unfocused
####

use strict;

# basedmg hash
# this is the base damage per tick as reported by lucy
# be careful of level ranges, format should be obvious,
# capitalization counts.
my %basedmg = ( "Embracing Darkness" => 70,
"Funeral Pyre of Kelador" => 310,
"Pyrocruor" => 166,
"Dark Plague" => 190,
"Saryrn's Kiss" => 200,
"Blood of Thule" => 354
);

# random variables that may or may not be used in this thing..
my $total_dotticks = 0;
my $total_crits = 0;
my $total_dmg = 0;
my $crit_dmg = 0;
my $lasttick;
my $thistick;
my $dotname;
my $damage;
my $crit_pct;
my $dps_increase;

open(THELOG,"eqlog_Lesk_tunare.txt") or die "Couldn't open log: $!\n";
while(my $line = readline(THELOG)) {
chomp($line);

if ($line =~ /..:..:(..).*taken (\d+) damage from your ([\w\s\']*)/) {
$thistick = $1;
$damage = $2;
$dotname = $3;
} else {
next;
}

next if ($dotname eq "Splurt"); #for now...

if (!exists($basedmg{$dotname})) {
print "Base damage not known for $dotname.\n";
print "Please add it to the basedmg hash.\n";
close THELOG;
exit;
}

if ($damage >= ($basedmg{$dotname}*2)) {
$total_crits++;
$crit_dmg += ($damage/2);
}

$total_dmg += $damage;
$total_dotticks++;
}

close THELOG;

$crit_pct = $total_crits / $total_dotticks * 100;
$dps_increase = $crit_dmg / $total_dmg * 100;

print "Total dot ticks: $total_dotticks\n";
print "Total crit ticks: $total_crits\n";
print "Percent occurance: $crit_pct\n";
print "Total dot damage: $total_dmg\n";
print "Crit damage: $crit_dmg\n";
print "Percent damage increase: $dps_increase\n";

exit;

[Xislaben]

Here's a little outing from last night with CA2:

http://img18.echo.cx/img18/5227/crits5gv.jpg

[Lifleviya Life'Taka]

Other than Yalp and Jalp are there any good parsers?

I tend to get mistakes on both, one only records my DoT damge while the other, "misses" fights.

[Xislaben]

Originally posted by Lifleviya Rahl@Jun 7 2005, 09:07 PM
Other than Yalp and Jalp are there any good parsers?

I tend to get mistakes on both, one only records my DoT damge while the other, "misses" fights.
I didn't like the information for crit reporting on the parsers I've found so I wrote my own in vb6 for specifically crit info.

[aeriform]

Originally posted by Xislaben+Jun 6 2005, 09:20 PM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (Xislaben @ Jun 6 2005, 09:20 PM)</td></tr><tr><td id='QUOTE'> <!--QuoteBegin-aeriform@Jun 6 2005, 08:17 PM
I could have sworn I posted something like this already, but maybe only on the defunct forums.
Are you taking into account that the first level of both CA and SCF do 200% now when they crit? [/b][/quote]
In the one I just posted I did make that update. CA has always been 200% though, so one only needs to change the SCF levels 1 and 2.

DoT parsing is pretty simple to do compared to a lot of other stuff, one thing to be careful of though are parses from places like MPG where you might be DoT'ing mobs that have innate DoT sheilding.

[Xheng]

Originally posted by Lifleviya Rahl@Jun 7 2005, 09:07 PM
Other than Yalp and Jalp are there any good parsers?

I tend to get mistakes on both, one only records my DoT damge while the other, "misses" fights.
JALP can record all of your damage, not just DoTs, if you set it up to do so:

1) Once per logfile, say the following out loud, or to any channel: <EQPARSE-Register Name>-> YourNameHere

2) When you summon a new pet, say the following: <EQPARSE-Register Pet>-> %M You can do it in say, any chat channel, or go back into your log file and manually put it in later... though if you do that you obviously need to type out the pet name instead of using the %M shortcut.

With these commands, JALP will very nicely lump in your pet damage and DD damage with your DoTs, for an overall DPS figure. All that I'm aware of it missing are:
. pet lifetaps,
. pet toy damage,
. pet DoT damage (if you have clicky pet from epic),
. pet 'punches' if you raise an Army of the Dead on players' corpses

I still like JALP overall. It's biggest problem for me is when you hit mobs that do odd forms of AE damage, like the boulder crash thing from the final boss in Ikk4. Those send it into an infinite loop unless you manually remove those offending lines from your logfile.

[Xislaben]

Originally posted by aeriform@Jun 8 2005, 04:36 AM
DoT parsing is pretty simple to do compared to a lot of other stuff, one thing to be careful of though are parses from places like MPG where you might be DoT'ing mobs that have innate DoT sheilding.
Oh yeah that would kind of suck. There's no way to tell if your dot critted other than looking at 2x or more over base damage since there's no message, so yeah if a mob mitigated that you may get false negatives which would artificially lower your crit numbers.

More CA2 parsing. If it's 6% at CA2 it sure does like to parse low:

http://img231.echo.cx/img231/7985/critsca26xd.jpg

[Xislaben]

CA2's looking more and more like 6% now:

http://img60.echo.cx/img60/8076/critsca20ha.jpg

[Brahman]

If nothing else, i'm glad my post pushed some parsing out :)

[chabeksis]

Originally posted by Brahman@Jun 17 2005, 04:08 AM
If nothing else, i'm glad my post pushed some parsing out :)
I need to post mine, but it is almost identical to what Xislaben has shown. I am almost to CA3 now so I will definitely post that parse as soon as I get CA3t, and I get a large enough sample.

BTW Brahman I am glad too, I would love to see more parsing.

[Xislaben]

Originally posted by Brahman@Jun 17 2005, 04:08 AM
If nothing else, i'm glad my post pushed some parsing out :)
:)

Either the RNG is being kind to me (and when does that happen?), or the odds may not be exactly 6%

http://img146.echo.cx/img146/7162/critsca2c1ep.jpg

[Xislaben]

I am stepped in blood so far....

So I thought I'd keep doing this, since I'd written a p[arser and the numebrs weren't what I expetced. Now with over 14k ticks, the crit % is climbing, is it possible it's a streaky rng, or is CA2 really closer to 6.15% ? If so that makes it even more valuable as a dps/aa than previously thought.

http://img235.echo.cx/img235/2653/critsca2d7cb.jpg

[aeriform]

Originally posted by Xislaben@Jun 22 2005, 01:53 PM
I am stepped in blood so far....

So I thought I'd keep doing this, since I'd written a p[arser and the numebrs weren't what I expetced. Now with over 14k ticks, the crit % is climbing, is it possible it's a streaky rng, or is CA2 really closer to 6.15% ? If so that makes it even more valuable as a dps/aa than previously thought.

http://img235.echo.cx/img235/2653/critsca2d7cb.jpg
Step out of the blood...

it is more than just possible that you are slightly better than average, heck there's a 50% chance of that for anyone...

You are using statistics. There is virtually nothing you can "prove" with statistics. The best you can do with "random" numbers in the real world is provide an estimate on how likely something is occuring.

With only 15,000 samples and 6.19% average from an infinite population, there is a 95% chance that the actual percentage to crit is in the interval of 6.19% plus or minus 0.39% (which includes 6.00%) or an interval of [5.80,6.58]%

Similarly, there is a 99% chance that the actual percentage to crit is in the interval of 6.19% plus or minus 0.51% (which includes 6.00%) or an interval of [5.68,6.70]%

If you wanted to be 95% certain that the crit percentage was not 6% and had an average of 6.19%, you would need a confidence range of less than 0.19% which would require more than 60000 samples (ie. 4 times more than what you've collected so far). This wouldn't be conclusive however since again, you wouldn't be 100% certain, only 95% certain. Welcome to statistics...

Question: What's the chance that Sony used an integer for the percent chance to crit?

[GnekroeGnomicon]

Welcome to statistics...Ssssshhhhoooooosshhhhhh! I will have none of that here! :blink:

[Xislaben]

Originally posted by aeriform@Jun 22 2005, 05:27 PM
If you wanted to be 95% certain that the crit percentage was not 6% and had an average of 6.19%, you would need a confidence range of less than 0.19% which would require more than 60000 samples (ie. 4 times more than what you've collected so far). This wouldn't be conclusive however since again, you wouldn't be 100% certain, only 95% certain. Welcome to statistics...
I'll settle for 95% confidence, but I doubt I'll still be at CA2 in 45k more ticks.

[Saetanis]

It all makes sense to me - statistics, calculus, trig - whatever!

My momma didn't raise no fool - present me the facts...

and can easily see that, in this case, 1 + 1 always equals 3.


- Saetanis

[GnekroeGnomicon]

in this case, 1 + 1 always equals 3.With a 95% confidence that is...

[Xislaben]

Originally posted by aeriform@Jun 22 2005, 05:27 PM
Question: What's the chance that Sony used an integer for the percent chance to crit?


For tradeskilling passes they've used percent chances to the ten's place, so they may have done that here.

http://img59.echo.cx/img59/7989/critsca2f7yx.jpg

[Jebasiz]

You have 20k ticks total but most of those dots weren't on mobs for alot of that time.

It's still nice to see, and it may even be valid since the numbers look fairly consistent through several thousand ticks of dotting.

I'd be interested to see how often pets flurry and crit..or backstab and stun for that matter.

[Xislaben]

Originally posted by Jebasiz@Jun 27 2005, 02:20 PM
You have 20k ticks total but most of those dots weren't on mobs for alot of that time.

It's still nice to see, and it may even be valid since the numbers look fairly consistent through several thousand ticks of dotting.

I'd be interested to see how often pets flurry and crit..or backstab and stun for that matter.
Dots weren't on mobs for alot of that time? I don't understand what you mean by this. I'm confident that it's much closer to 6% than 5% for CA2, but the jury's still out as to it being exactly 6.00% or perhaps 6.05% or 6.10% or some such other number.

I'd like to see the pet parses too, but I know that I'm frequently out of melee range required to catch them, so I'd be missing quite a few, especially for the rog damage which would be really interesting to know. You could have a second toon shadow your pet purely to parse it as you agro kite, but that's a bit more work than I'm willing to commit to. Unless all hits were viewable from the same distance I'm not sure it would be valid to accept only the hits you could see for statistical purposes. Crits for instance may be something you could see from farther away, and if so this would possibly bias the data towards higher critting as normal hits and misses may be missed.

That and i don't have any of the pet aa's.

[aeriform]

Originally posted by Jebasiz@Jun 27 2005, 02:20 PM
You have 20k ticks total but most of those dots weren't on mobs for alot of that time.
You're misunderstanding the way Xislaben is representing his info (which looks perfectly fine to me). When the picture says you have 19566 ticks that's actually much less than 20K times 6 seconds of time parsed. What that number is is the total number of chances for each individual DoT to crit, ie. add up all the "# Ticks" for each individual DoT.

No where in the picture shown is there any information directly related to the actual time or actual number of ticks that the parses were taken over. That info really isn't necessary or useful and is rightly not included because DoT's that are running do not all crit on the same tick. Each chance to crit is based on the individual DoT AND each individual tick, rather than just each individual tick.

[Siddar]

So makeing rough ball park estimate based on model of a average necro soloing where he does damage as follows from DoT 66% Pet 30% DD 4% I come out with rough estimate of.

36 AA in DoT crits nets me a 16% increase in damage with above model.

40 AA in Pet crits and flurrys gives me a 3.3% increase in damage with same model.

60 AA in DD crits give about a 1% increase useing same model.

Guess ill be putting DD crit AA on back burner for awhile.

[Xislaben]

Here's the final parse of CA2, cause I just bought CA3. Since I need only 3 aa's for the next level, ICA1, I will have even less data for CA3. Siddar, I rarely use dd's when not raiding. If I had lots of aa's spent in crit heals and crit dd's I'd probably tap more, but I don't so... I'd say my damage from dots far far exceeds dd damage.

BTW all parses here are with glowing stone of expulsion, ba5, and the warped choker of enslavement for damage focii.

http://img148.imageshack.us/img148/1036/critsca2g2gx.jpg

[hughman]

shameless bump as I found this usefull, again!

Loketh on Prexus.

[UsulDaNeriak]

actually i like mathematics too,

but my stratgey was pretty simple for every new expansion:

1st get the real functionally uber aa, like 9th spell slot,.....
2nd go for everything raising your dps. after crit dots, the others, in whatever order
3rd go defense, dmg avoidance like precognition first, ........
4th go for the rest

just my 2cp.

Usul