Some preliminary conclusions concerning avoidance in 3.2
Moderators: Fridmarr, Worldie, Aergis, theckhd
Re: Some preliminary conclusions concerning avoidance in 3.2
Wanted to see how changing some dodge to parry worked in action for my character with the new formulae, so I played with swapping out a 34 dodge gem with a 34 parry gem on the PTR.
I noticed that unbuffed (108 agi) I had a net loss of avoidance of ~0.03%. With Kings + Mark of the Wild, the difference between the dodge and parry gems was 0.01% avoidance. Add in Horn/SoE, and I think I'd see a net gain. Very small differences, but it seems like this fine tuning actually works.
Of course, with the gear I have on PTR atm I only have one dodge gem to swap out, the rest is all on my gear, and it's barely worth paying for the gem for just a few fractions of a percent, but it means in future I'll be considering parry (or parry/def) in red slots.
I noticed that unbuffed (108 agi) I had a net loss of avoidance of ~0.03%. With Kings + Mark of the Wild, the difference between the dodge and parry gems was 0.01% avoidance. Add in Horn/SoE, and I think I'd see a net gain. Very small differences, but it seems like this fine tuning actually works.
Of course, with the gear I have on PTR atm I only have one dodge gem to swap out, the rest is all on my gear, and it's barely worth paying for the gem for just a few fractions of a percent, but it means in future I'll be considering parry (or parry/def) in red slots.
 KysenMurrin
 Posts: 6827
 Joined: Thu Jun 26, 2008 6:37 am
 Location: UK
Re: Some preliminary conclusions concerning avoidance in 3.2
Theckhd,
Looking over it the data you posted is pretty exciting. Some of your results for the first stage matched mine as well (the point before you start stacking defense again). I found the dodge/parry ratio to be 1.875 if you keep defense constant, which is close to your 1.75 found through iteration.
Using differential analysis, I would have found that 6.3% parry matched with 11.8125% dodge roughly, which is extremely close to your 11.94% gotten through iteration, so those results seem to match up well. I haven't put in the time to see how stacking all 3 at once would work, but it looks like you used the same method to get your ratios, so that seems pretty sound.
Out of curiosity, have you though about how the results look when taking defense in integer steps? Converting from defense rating to defense skill seems to happen in integer values (at least from what we have seen so far).
A poster over at tankspot has gone through that and seems to have found that it is ultimately better not to stack defense at all due to the effects of the stairstep effect of the conversion. I will be honest and say I haven't looked through it thoroughly myself, but it might be something to consider and at least check out. The link for the first post is http://www.tankspot.com/forums/f14/5240 ... post251383 but there are other posts after that too. I pointed him over here as well.
So my question to you is if you can compare your proposed ratio, but take into consideration the stair step effect and see how it compares to simply keeping your dodge=1.875parry while not stacking defense at all.
Looking over it the data you posted is pretty exciting. Some of your results for the first stage matched mine as well (the point before you start stacking defense again). I found the dodge/parry ratio to be 1.875 if you keep defense constant, which is close to your 1.75 found through iteration.
Using differential analysis, I would have found that 6.3% parry matched with 11.8125% dodge roughly, which is extremely close to your 11.94% gotten through iteration, so those results seem to match up well. I haven't put in the time to see how stacking all 3 at once would work, but it looks like you used the same method to get your ratios, so that seems pretty sound.
Out of curiosity, have you though about how the results look when taking defense in integer steps? Converting from defense rating to defense skill seems to happen in integer values (at least from what we have seen so far).
A poster over at tankspot has gone through that and seems to have found that it is ultimately better not to stack defense at all due to the effects of the stairstep effect of the conversion. I will be honest and say I haven't looked through it thoroughly myself, but it might be something to consider and at least check out. The link for the first post is http://www.tankspot.com/forums/f14/5240 ... post251383 but there are other posts after that too. I pointed him over here as well.
So my question to you is if you can compare your proposed ratio, but take into consideration the stair step effect and see how it compares to simply keeping your dodge=1.875parry while not stacking defense at all.

jere  Posts: 2958
 Joined: Fri May 11, 2007 5:12 pm
Re: Some preliminary conclusions concerning avoidance in 3.2
Okay, I feel a little stupid now, but if I'm understanding it correctly, when i do the substracting:theckhd wrote:
 Take current ratings and subtract out 689 def, 96 dodge, and 64 parry
 Figure out your current ratio of dodge:def:parry
 if one of these is higher than the ideal ratio of 2.4:1.8:1, swap out some of the rating that's above the target ratio for a rating that's below or at the target value.
 Recalculate and lather, rinse, repeat.
 Code: Select all
def =759689 =70
dodge=71696 =620
parry=15864 =94
I end up at some 620:70:94, now that's not nearly even similar to the ratios desired for 3.2
Does that mean I will have to drop some two hundreds of dodge(very roughly) to get some efficiency out of these stats?
 MrDuck
 Posts: 433
 Joined: Wed Jun 04, 2008 4:27 pm
 Location: Obsidian Sanctum
Re: Some preliminary conclusions concerning avoidance in 3.2
MrDuck wrote:Okay, I feel a little stupid now, but if I'm understanding it correctly, when i do the substracting:theckhd wrote:
 Take current ratings and subtract out 689 def, 96 dodge, and 64 parry
 Figure out your current ratio of dodge:def:parry
 if one of these is higher than the ideal ratio of 2.4:1.8:1, swap out some of the rating that's above the target ratio for a rating that's below or at the target value.
 Recalculate and lather, rinse, repeat.
 Code: Select all
def =759689 =70
dodge=71696 =620
parry=15864 =94
I end up at some 620:70:94, now that's not nearly even similar to the ratios desired for 3.2
Does that mean I will have to drop some two hundreds of dodge(very roughly) to get some efficiency out of these stats?
so your ratios are:
6.6:0.7:1
If Theck's ratios and analysis are correct that means:
6.6 is higher than 2.4, so go to the next one.
0.7 is lower than 1.8, so stack more defense until you have a 1.8:1 ratio with parry
Once that is done, stack both def and parry in a 1.8 ratio until your dodge is down to 2.4:1 with parry
Stack all three at 2.4:1.8:1
Assuming I understood Theck's post correctly.

jere  Posts: 2958
 Joined: Fri May 11, 2007 5:12 pm
Re: Some preliminary conclusions concerning avoidance in 3.2
Theck, looking over some of the intial numbers, I have some questions.
At one point, you state when you have 689 defense rating and 64 parry rating, you get 6.3% parry. Now in 3.1, 64 parry rating would give you 1.3% parry, but 1.3% + 5.6% (from 689 def rating) will be 6.9%, which doesn't match your 6.3% figure. Furthermore that's the 3.1 rating. The 3.2 rating would make 64 parry = 1.41% parry giving 7.01% parry when adding in the 5.6% from defense.
What am I missing there? I can't get the numbers to come out correctly. I am sure I am doing something silly.
At one point, you state when you have 689 defense rating and 64 parry rating, you get 6.3% parry. Now in 3.1, 64 parry rating would give you 1.3% parry, but 1.3% + 5.6% (from 689 def rating) will be 6.9%, which doesn't match your 6.3% figure. Furthermore that's the 3.1 rating. The 3.2 rating would make 64 parry = 1.41% parry giving 7.01% parry when adding in the 5.6% from defense.
What am I missing there? I can't get the numbers to come out correctly. I am sure I am doing something silly.

jere  Posts: 2958
 Joined: Fri May 11, 2007 5:12 pm
Re: Some preliminary conclusions concerning avoidance in 3.2
Diminishing returns
Amirya wrote:some bizarre lovechild of Hawking, Einstein, and Theck

majiben  Moderator
 Posts: 6999
 Joined: Fri Aug 22, 2008 4:37 pm
 Location: Retired
Re: Some preliminary conclusions concerning avoidance in 3.2
Decided to pop over into this thread  I'm the one who was doing all the math on Tankspot.com. From what I discovered using bruteforce calculation of the best possible avoidance given an arbitrary number of points to spread between dodge/defense/parry rating (with 304 bonus agility as a paladin), the following rules should govern gearing for best avoidance in 3.2:
1) Stack defense rating until you reach a number that has a minimum of wasted in the defense rating > defense conversion. (728,787,846,910 and 969 are the best possible ones to hit).
2) Stack dodge/parry rating from there so that the postDR portions of your Dodge% and Parry% are in a 1.875:1 ratio
3) If your total dodge+parry rating is getting anywhere near your defense rating, try to hit the next good number of defense rating (while adjusting your dodge/parry rating to maintain #2) as the diminishing returns on miss will be such that you will notice a slight difference.
4) If you can't hit the next good number of defense rating, don't worry  keep stacking dodge/parry in the aforementioned ratio and you'll be fine.
You can look at the tankspot.com post which Jere posted above to get to the plots, but if you want to look at the raw data, Here is a google docs spreadsheet of the maximum possible avoidance at each total of rating points (calculated by checking the total avoidance of every combination possible). Aside from listing just the ratings and total avoidance, I also made a column with the postDR percentages of dodge/parry/miss, and the ratio of postDR dodge% to postDR parry%. You'll notice that every single ratio hovers around 1.875:1.
Also, in regards to 1), I am not advocating going out of your way to hit those five exact numbers. Any defense rating amount with <10% wasted defense skill should be indistinguishable from those to two decimal places. (see the tankspot.com post for a plot and further explanation.) Also, once epic gems come out and we're gemming in 10defense rating increments, you won't be able to use gems to shift toward a better defense number easily by gemming since the steps up a defense rating line are almost all in 5point increments.
I believe if you run the numbers on your own simulation, though, you'll see that the closer your dodgeparry ratio is to 1.875:1, the greater your avoidance will be for any given point spread.
1) Stack defense rating until you reach a number that has a minimum of wasted in the defense rating > defense conversion. (728,787,846,910 and 969 are the best possible ones to hit).
2) Stack dodge/parry rating from there so that the postDR portions of your Dodge% and Parry% are in a 1.875:1 ratio
3) If your total dodge+parry rating is getting anywhere near your defense rating, try to hit the next good number of defense rating (while adjusting your dodge/parry rating to maintain #2) as the diminishing returns on miss will be such that you will notice a slight difference.
4) If you can't hit the next good number of defense rating, don't worry  keep stacking dodge/parry in the aforementioned ratio and you'll be fine.
You can look at the tankspot.com post which Jere posted above to get to the plots, but if you want to look at the raw data, Here is a google docs spreadsheet of the maximum possible avoidance at each total of rating points (calculated by checking the total avoidance of every combination possible). Aside from listing just the ratings and total avoidance, I also made a column with the postDR percentages of dodge/parry/miss, and the ratio of postDR dodge% to postDR parry%. You'll notice that every single ratio hovers around 1.875:1.
Also, in regards to 1), I am not advocating going out of your way to hit those five exact numbers. Any defense rating amount with <10% wasted defense skill should be indistinguishable from those to two decimal places. (see the tankspot.com post for a plot and further explanation.) Also, once epic gems come out and we're gemming in 10defense rating increments, you won't be able to use gems to shift toward a better defense number easily by gemming since the steps up a defense rating line are almost all in 5point increments.
I believe if you run the numbers on your own simulation, though, you'll see that the closer your dodgeparry ratio is to 1.875:1, the greater your avoidance will be for any given point spread.

Xenix  Posts: 244
 Joined: Thu Jun 25, 2009 4:56 am
Re: Some preliminary conclusions concerning avoidance in 3.2
So I get
3 defense 336 Dodge 271 Parry
This is subtracting 689/96/64 from my ratings of 692/432/335
Now how do I get my ratio and from my ratio where do I go?
3 defense 336 Dodge 271 Parry
This is subtracting 689/96/64 from my ratings of 692/432/335
Now how do I get my ratio and from my ratio where do I go?
 towelliee
 Posts: 1160
 Joined: Tue May 27, 2008 8:09 am
Re: Some preliminary conclusions concerning avoidance in 3.2
towelliee wrote:So I get
3 defense 336 Dodge 271 Parry
This is subtracting 689/96/64 from my ratings of 692/432/335
Now how do I get my ratio and from my ratio where do I go?
336:3:271
1.23 : 0.01 : 1

honorshammer  Moderator
 Posts: 2147
 Joined: Tue Apr 10, 2007 9:35 am
 Location: Charleston, SC
Re: Some preliminary conclusions concerning avoidance in 3.2
Considering that aatm have a total of three gems where I socket defense or agilitiy, I wonder if fine tuning is actually possible in a way that matters. In 3.2 it´s six sockets as JC gems won´t be prismatice any longer. Don´t think I will get many more sockets, so we talk about hm perhaps 80 rating or 100 for a blacksmith? Most of the stats are dependant on gear and selection choices are limited to tier, a few off set pieces and so on. Some of them will probably itemised horribly.
So the whole thing will rather be interesting for the gear lists.
So the whole thing will rather be interesting for the gear lists.
 Dantriges
 Posts: 1252
 Joined: Wed Jan 30, 2008 12:39 am
Re: Some preliminary conclusions concerning avoidance in 3.2
So now what do I do with
1.23 : 0.01 : 1 What will tell me about the best possible manner in which to gem my gear or itemize.
1.23 : 0.01 : 1 What will tell me about the best possible manner in which to gem my gear or itemize.
 towelliee
 Posts: 1160
 Joined: Tue May 27, 2008 8:09 am
Re: Some preliminary conclusions concerning avoidance in 3.2
It means that to maximize your avoidance, you'll need a LOT more defense and less parry.

culhag  Maintankadonor
 Posts: 1728
 Joined: Wed Aug 06, 2008 7:50 am
 Location: France
Re: Some preliminary conclusions concerning avoidance in 3.2
Makes sense thats from my EH set my AV set has 31% d 21 parry so I have to log in and switch gears
 towelliee
 Posts: 1160
 Joined: Tue May 27, 2008 8:09 am
Re: Some preliminary conclusions concerning avoidance in 3.2
I hate blizz
Every 2 month you have to change your theorycraft and your gear because they stupidly change the rules of optimization
Definitively a thing that upset me ..
no wait ... I really hate blizz
edit: I'll try to share my own results when I will have them
Every 2 month you have to change your theorycraft and your gear because they stupidly change the rules of optimization
Definitively a thing that upset me ..
no wait ... I really hate blizz
edit: I'll try to share my own results when I will have them

Cema  Posts: 426
 Joined: Wed Oct 01, 2008 8:46 am
Return to Advanced Theorycraft and Calculations
Who is online
Users browsing this forum: Yahoo [Bot] and 1 guest