Nerfed block chance on Baleroc encounter?

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Nerfed block chance on Baleroc encounter?

Postby Rhiannon » Sun Jul 10, 2011 3:40 pm

Edit: To summarise this lengthy thread, it appears that for whatever reason you require 103.4% CTC to be capped against his white swings. To be capped against inferno strike, you require considerably more, people have reported inferno strike hits even with 108.4%+. From a gearing point of view you're most likely best off going with 103.4% and just using CDs other than holy shield for inferno blade rather than trying to cap vs inferno blade.

I'm not sure whether I'm just being incredibly stupid at the moment or if Blizzard have actually changed something (or possibly both).

I've been using a block-capped set for several fights in Firelands, and I still seem to be taking unblocked hits. My raid buffed character sheet dodge is 14.3, parry 14.6, and block 69.05. 5 + 14.3 + 14.6 + 69.05 = 102.95% (I was at 102.45% but switched a couple of gems after noticing unblocked hits, but still see them).

Now I've noticed the odd unblocked hit showing up in logs and put it down to me possibly screwing up in some way and getting hit in the back (when moving hatchlings around worms on Alys, or getting aggro on spinners on beth and not really caring, for example), but now we're working on Baleroc which is a fight that's pretty impossible to let the boss hit you in the back on.

Looking at tonight's attempts, over all tries I took:

1814 melee swings, of which:

Hit: 125 (6.8908%)
Block: 962 (53.0320%)
Absorb: 112 (9.5369%)
Dodge: 230 (12.6792%)
Miss: 56 (3.0871%)
Parry: 268 (14.7740%)

(edit: fixed numbers)

Even assuming all the absorbs were blocked hits, I'm about 6.5% (edit: not 9%) short on my block for some reason.

Example of unblocked hit:

[18:50:49.593] Baleroc hits Rhia 184257

Example of full absorb:

[18:57:30.048] Baleroc hits Rhia Absorb (75753)

So, anyone got any ideas?

Edit: as Marblehead pointed out, world of logs counts absorbs as hits. So if you have the same number of hits as absorbs, that means you didn't take any unblocked hits (or if you did it was fully absorbed, we can't know for sure). If # of hits - # of absorbs > 0, then you have probably taken unblocked hits.
Last edited by Rhiannon on Sat Oct 08, 2011 10:31 am, edited 6 times in total.
Rhiannon
 
Posts: 1061
Joined: Mon Mar 03, 2008 8:17 am

Re: Nerfed block chance on Baleroc encounter?

Postby Koatanga » Sun Jul 10, 2011 5:04 pm

Some times, for those tracking it down.

Code: Select all
[18:50:49.593] Baleroc hits Rhia 184257

[21:33:33.915] Baleroc hits Rhia 33533

[21:47:45.010] Baleroc hits Rhia 78715

[21:48:41.170] Baleroc hits Rhia 315182


I don't see any pattern from it - it looks very much like random full hits, which I agree should not exist with full CTC.
Retired. Koatanga, Shapely, Sultry, Doominatrix of Greenstone - Dath'Remar
Koatanga
 
Posts: 1995
Joined: Mon Nov 17, 2008 12:46 pm

Re: Nerfed block chance on Baleroc encounter?

Postby rodos » Sun Jul 10, 2011 6:01 pm

Maybe Ghostcrawler came good on his promise to give bosses expertise in higher tiers.

We'd need to see a lot more parses to make any conclusions, but your (dodge+parry) from character sheet also comes out 0.5% higher that what you saw land during the fights.

Does this happen on any other boss?
User avatar
rodos
 
Posts: 1120
Joined: Mon Sep 24, 2007 8:20 pm

Re: Nerfed block chance on Baleroc encounter?

Postby Rhiannon » Sun Jul 10, 2011 6:07 pm

Well that'd be expected, character sheet dodge/parry of 14.3/14.6 would translate to 13.7/14.0 vs 88s due to level difference. But anyway due the small sample size a 0.5-1.0% deviation doesn't bother me. an ~8% deviation from what's expected and the presence of full hits that shouldn't actually be possible is much more concerning.

And yeah, I've noticed it vs other bosses but it hadn't contributed to a tank-death related wipe up until today so I didn't really investigate/put it down to me being bad. You can find plenty more examples from our beth and alys logs though, both of which I had the same gear set for.
Rhiannon
 
Posts: 1061
Joined: Mon Mar 03, 2008 8:17 am

Re: Nerfed block chance on Baleroc encounter?

Postby theckhd » Sun Jul 10, 2011 6:16 pm

I get slightly different numbers because those don't add up to 1753, they add up to 1814:

Hit: 6.8908%
Block: 53.0320%
Absorb: 9.5369%
Dodge: 12.6792%
Miss: 3.0871%
Parry: 14.7740%

Nonetheless, that's really strange. If they had decided to give these bosses extra "expertise," I would have expected them to tell us officially somehow. And if they did so, they'd probably do it through a straight nerf to dodge, not block, so that it would affect all tanks equally. You only seem to be "down" about 1.62% dodge, which isn't very much - certainly not enough to make a huge difference, and not enough to account for a nearly 7% block-cap discrepancy.

Nor does a two-roll system work out right. With a 33.9% chance to avoid and 69.05% to block, we'd expect an (1-0.339-0.006*3)*(0.6905-0.006) = 44.01% observed block rate and a (1-0.339-0.006*3)*(1-0.6905+0.006)= 20.29% hit rate. So it's not rolling once for avoidance and a second time for block.

However, there's something very strange going on either way. Can you give us a link to the exact gear/buff setup you used to reach block cap? I can't imagine you made any mistake that would be 6% or more off, but we have to rule out the obvious bases first before we can start making statements about hidden boss mechanics.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7970
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Nerfed block chance on Baleroc encounter?

Postby Jeremoot » Sun Jul 10, 2011 6:18 pm

Chance to miss is 4.4%. Still doesn't bring you below full CTC. You don't need to translate 14.3/14.6 to an 88 because you're already accounting for that by using the 102.4 figure. It's 100% + 0.2% block, miss, dodge, and parry per level.

This is a serious issue if it's not intended, and an even bigger issue if it was an intended stealth mechanic.
User avatar
Jeremoot
 
Posts: 434
Joined: Tue Feb 08, 2011 5:36 pm

Re: Nerfed block chance on Baleroc encounter?

Postby Rhiannon » Sun Jul 10, 2011 6:26 pm

Oops, calculator fail - editing first post for total # of hits.

Armory, with 315 mastery from elixir/food, plus additional parry from horn of winter + kings.

Jeremoot wrote:Chance to miss is 4.4%. Still doesn't bring you below full CTC. You don't need to translate 14.3/14.6 to an 88 because you're already accounting for that by using the 102.4 figure. It's 100% + 0.2% block, miss, dodge, and parry per level.


I know, I was just saying to Rodos that 14.3% dodge on character sheet equates to expected 13.7% dodge vs 88 due to the level based suppression.
Rhiannon
 
Posts: 1061
Joined: Mon Mar 03, 2008 8:17 am

Re: Nerfed block chance on Baleroc encounter?

Postby lythac » Mon Jul 11, 2011 3:27 am

It is almost as if the DW penalty bosses used to suffer from has been reversed twice but for OH only.

MH -5 -19 +19 = -5
OH -5 -19 + 19 + 19 = +14

If one of the weapons is at 14%, you would see a hit rate of 7% and it would push block of the table.


Guessing is fun \o/

Edit - Do you have to do Chess Event in Kara to get to Prince Malchezaar?
Ryshad / Lythac of <Heretic> Nagrand-EU
User avatar
lythac
Moderator
 
Posts: 2685
Joined: Wed Sep 24, 2008 8:10 am

Re: Nerfed block chance on Baleroc encounter?

Postby Chicken » Mon Jul 11, 2011 3:47 am

lythac wrote:Edit - Do you have to do Chess Event in Kara to get to Prince Malchezaar?
Yes, it's soloable though. I'd link a video guide but most websites I could get that from are blocked from work, so...
Image
User avatar
Chicken
 
Posts: 1597
Joined: Fri Jun 26, 2009 2:19 pm

Re: Nerfed block chance on Baleroc encounter?

Postby Marblehead » Mon Jul 11, 2011 5:26 am

When a hit is fully absorbed, WoL counts it also as a full hit.

According to the log, you had 125 hits and 112 full absorbs. Using a query on the expression editor to find out how many unblocked hits you've got, returns (guess what) 13 hits. I'd bet that you were missing some kind of buff during these.
Image
Life is not difficult, people are.
User avatar
Marblehead
 
Posts: 202
Joined: Sun Aug 02, 2009 9:28 pm
Location: Bloodhoof (EU)

Re: Nerfed block chance on Baleroc encounter?

Postby Rhiannon » Mon Jul 11, 2011 6:00 am

Hm, ok that's a much smaller number. However the only buffs beyond gear I get are motw, battle shout, elixir and food. Looking at the attempt where I took two swings:

[21:47:45.010] Baleroc hits Rhia 78715
[21:48:41.170] Baleroc hits Rhia 315182

When I died:

[21:48:59.870] Aurelya's Mark of the Wild fades from Rhia
[21:48:59.870] Eija's Battle Shout fades from Rhia
[21:48:59.870] Rhia's Elixir of the Master fades from Rhia
[21:48:59.870] Rhia's Well Fed fades from Rhia (and it's 90 mastery, not 90 dodge from using feast by mistake)

Checking uptime on that try for battle shout it was up for both hits, and I had motw from the start.

Edit: there's also pretty much no chance I could have accidentally turned my back to the boss, as I don't move after the first 5 seconds of the pull, but just in case, looking at the last hit I blocked both the previous and following hits.

Double checking I'm not somehow misremembering:

Image

That's with kings, horn of winter, elixir and food.

So discounting the absorbs now, I took 1641 total swings, of which 13 were full hits. That's 0.79%.
Last edited by Rhiannon on Mon Jul 11, 2011 6:24 am, edited 1 time in total.
Rhiannon
 
Posts: 1061
Joined: Mon Mar 03, 2008 8:17 am

Re: Nerfed block chance on Baleroc encounter?

Postby theckhd » Mon Jul 11, 2011 6:22 am

Armory audit checks out. When taking your numbers and hand-calculating everything (well, MATLAB-calculating anyway) I get exactly your raid-buffed avoidance/block numbers:

Parry: 14.60%
Dodge: 14.30%
Block: 69.05%

102.95% total CTC.


Looking at the numbers again, if we assume all of those full absorbs would have been blocks, we get 62.46% block. With the 6.89% hits represented, that makes 69.46%, close to our expected block value. That could mean that we're seeing some sort of flat 7% block suppression, though of course there are other possible explanations as well.

I'm puzzled by Lythac's suggestion about the Dual-wield penalty. The numbers work out, but it's counter-intuitive based on our knowledge of how the combat table works. For example, if we write out the combat table for a generic weapon:

Code: Select all
00.01-04.40 = miss  (4.4%)
4.41-18.10 = dodge (13.70%)
18.11-32.10 = parry (14.00%)
32.11-100.0 = block (67.9%) (0.55% spillover)
----------- = hit   (0.0%)


The DW penalty should manifest itself as a "free" 19% CTC in the form of miss - i.e. instead of 0.01-04.40, it would cover 0.01-23.40. Removing that penalty once would drop it back to 4.4%, removing it a second time should just null out the remaining 4.4% and leave us with

Code: Select all
00.00-00.00 = miss  (0.0%)
0.01-13.70 = dodge (13.70%)
13.71-27.70 = parry (14.00%)
27.71-96.75 = block (69.05%)
96.76-100.0 = hit   (3.25%)


Which should show up as a 3.25/2 = 1.63% hit rate in the logs. If it wasn't being zero-limited properly, I'd expect the table to start eating up the next piece, which would be dodge. That would nullify another 9.6%, giving a total miss chance of 12.85% for that weapon and an overall chance of 12.85/2=6.43%. That's actually pretty close to what we're seeing. But it's really bizarre behavior unless we assume that the combat table is programmed funny.

On the other hand, looking through the log, there were 0 unblocked hits during Beth'tilac attempts. Interestingly, WoL reports 26 melee hits in Rhia's breakdown, but they're completely absent from the expression editor breakdown.

We see full hits from hatchlings on Alysrazor, but those could be movement-related issues.

You were also getting some full hits on Shannox, though again those could be movement related. They seem to still occur when he's got his spear (and you're presumably not moving too much), which could suggest that it's not a DW issue. I'm not sure if you were wearing the same gear set that day though.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7970
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Nerfed block chance on Baleroc encounter?

Postby theckhd » Mon Jul 11, 2011 6:24 am

Marblehead wrote:When a hit is fully absorbed, WoL counts it also as a full hit.

According to the log, you had 125 hits and 112 full absorbs. Using a query on the expression editor to find out how many unblocked hits you've got, returns (guess what) 13 hits. I'd bet that you were missing some kind of buff during these.


This explains some of the disagreement I was seeing between the compiled data and my expression editor queries. 13 hits out of ~1800 swings is less than 1%. The real question is whether we can explain those away based on transient effects, or whether there's still a ~1% discrepancy with our understanding of the combat table.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7970
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Nerfed block chance on Baleroc encounter?

Postby Rhiannon » Mon Jul 11, 2011 6:28 am

Don't honestly remember about Shannox, I know I wasn't wearing the same gear but I don't recall if I was in a block-capped set or not.
Rhiannon
 
Posts: 1061
Joined: Mon Mar 03, 2008 8:17 am

Re: Nerfed block chance on Baleroc encounter?

Postby theckhd » Mon Jul 11, 2011 7:53 am

Interesting notes:

Hits don't seem to be clustered around the Decimation Blade proc, so it's unlikely to be an interaction with that (presumably they wouldn't need a special mechanic to prevent blocks on DB since it's a simple "if it connects, deal 90% HP damage thing, but worth checking anyhow).

Hits do seem to be limited to periods where Windwalk is not active.

The Windwalk proc gives you about 2.5% extra dodge. If we assume that it's a DW correction bug, then that would drop the miss chance in this data set to 0.75% for the OH or 0.375% overall while Windwalk is active. However, the time-averaged expectation based on WW uptime is still around 1.34%, which is higher than observed.

It's also possible they buffed him by 20% instead of 19% to compensate for DW - that would give a fairly consistent 1% error, which would lead to an observed error of around 0.77% based on WW uptime. That's pretty close to the 13/1800 we're observing.

<edit> Future tests (if anyone gets a chance):
1) Gear for 103.4% CTC or higher on Baleroc - see if any full hits are observed. If so, it's not a simple 1% error.
2) Gear for 106.8% CTC or higher - see if any full hits are observed. If so, this rules out the "double DW compensation" error.

Not sure if these are completely reasonable to achieve yet, but it might be possible using lower-ilvl pieces with more mastery.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7970
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Next

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest

Who is online

In total there is 1 user online :: 0 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: No registered users and 1 guest