Jadawin

Blood Tap or Runic Corruption?

22 posts in this topic

So while in the previous thread we discussed the difference between the DS/min and the resource gain from RC vs BT you were quite adamant that the real issue at hand was the consistency that RC gave you in terms of every single proc moved you closer to your next DS thus every single proc mattered. You are absolutely 100% correct in this statement. Undeniable that when you proc 100% of the time it does get you closer to your next DS. The only problem is such. Just because you proc and you move closer to your next DS does not mean that in all situations that a RC proc would have gotten you closer to your next DS faster than BT.

We can say that any scenario involving a comparative analysis of the two abilities would be based upon 3 static factors and a % chance.

3 Static factors.

Current Rune Postion(such as if you were at 40% of your runes for DS ect

Current RP and thus number of RS able to be executed

Current Blood Charges

Obviously the % chance would be the 45% chance on RC to actually proc.

So we can say in the millions of possible iterations of comparative analysis there are only 3 possible outcomes with all possible scenarios.

a. BT gets you to your next DS faster

b. RC has the opportunity to get you to your next DS faster

c. They are equal(I would imagine this would be quite a small % of the whole)

So you could say that a, b, and c all had a % of 100 of happening based upon the millions upon millions of iterations.

So while no one is going to do the variance breakdown of the infinite possible scenarios we can make a couple of very ironclad statements.

1. When BT has the opportunity to be the ability to get to your next DS faster it is 100% of the time being as a no variance ability.

2. When RC has the opportunity to be the ability to get to your next DS faster it is only the best ability a % of the time based upon the number of Runestrikes for example.

a. For any iteration of 1 Runestrike 55% of the time RC will be even with BT and 45% it would be faster

b. For any iteration of 2 Runestrikes 30.3% of the time RC will be even with BT and 69.7% it would be faster

3. Any % of iterations where RC does not actually proc only puts more of the % into the even or BT probablity of being better at achieving your next DS faster.

While the numbers here are not perfect they prove my point perfectly. So lets say in the infinite iterations of having 30 RP saved up and you have 1 RS to execute to move you closer to your next DS. We will say for posterity sake that the split is 50/50(please note I dont attest to this figure im just using it as a reference to show why the talents are not equal.

So if 50% of the time BT was better and 50% of the time RC had the opportunity to be better 55% of the 50% of the time RC was better you would not proc with the single Rune Strike thus making RC completely equal to BT because you would move to your next DS at the exact same rate of 8.33 seconds(modified by haste of course) This would mean that.

50% of the time BT was better

27.5% of the time they would be equal BT would be barely ahead

22.5% of the time RC would be better

So while yes I agree with the fact that every proc of RC means a DS sooner if you really think about the probability over the course of all possible scenarios just based upon the 45% proc chance of RC there are vastly more scenarios where BT will get you to your next DS faster than RC.

I hope you see this isnt a statement of resources gained per rune strike or DS/min this is simple a comparative analysis of which talent will more often provide you with your next DS faster when you needed it.

Edited by Jadawin

Share this post


Link to post
Share on other sites

lol, one thread gets locked, lets make a new one ^^

I have been using RC the most sofar, and recently I started changing my approach towards it.

lets say I have 90 Runic Power (wich doesn't and shouldn't happen too often) This gives me 3 rune strikes.

When I use the all three after eachother and the second and the third proc, this is not optimal use of RC. (This is how I've been doing it for quite some time)

When I runestrike I now always check if RC procced and let the proc run out before I do another runestrike. it may only matter a second of double regeneration, but every second counts towards the deathstrike total we all seem to be striving towards.

Furthermore I must say that (And I realise this can be done wiht BT too) I like RC because of the regenration of bloodrunes aswell. when there are just a couple of adds that spawned I can easily try to proc RC and do some boiling to get their attention. using blood tap this would cost me a deathrune I would probably rather use for deathstrikes.

thats all.

Shine

Share this post


Link to post
Share on other sites

When you stop looking at the throughput, and instead look at "how often will these talents give me benefit, regardless of how much that benefit is".

Does 10 stacks of BT give you more than 1 proc of RC? of course, but what doesn't show up well on paper (and this is what i've been trying to get you to see) is that while in terms of throughput and overall resources BT will be ahead always, the gaps between DS have a chance to be bigger with BT on occasion. If you are left with 1 FU pair and 0 charges of BT (had just finished using them) you will have to wait for the entire remaining duration of the other recharging FU pair before you get your next FU pair. Now look at RC, as long as you get a SINGLE proc (which requires at minimum 1 runes strike as opposed to the 5 needed for BT) you will already be ahead because no matter how much regen RC gives you the gap will be less than the full duration wait of that situation in BT.

That is RC's strength: frequency, not potency. Notice how I didn't use any algebra or resource analysis to find RC's strength, because some things in theorycrafting aren't just "ok punch the numbers".

Now that's it's strength, and it's what makes RC a viable and equally optimal t75 talent. That doesn't mean it's the only optimal t75 talent, and BT does have it's own strengths, those being potency compared to RC as well as control (and working well with t16 4pc bonus assuming there are any free death strikes left after they are done nerfing it to the ground) and that's why they are both viable and optimal because they are equal but different. Unless there is a talon rake/thrash/impale/fatal strike/triple puncture-esque mechanic in a fight, which will always favor BT (for the ability to frontload 3 death strikes in front of that massive hit as opposed to 2 with RC).

Share this post


Link to post
Share on other sites

So I wanted to make an example of a situation that was you could statistically understand which ability would get you to your next DS faster. Now you could do this for any situation and extrapolate the outcome.

For this example we will say you are at

0.3U 0.3F 0.0B

30RP

We will do all Blood Charge options of 0-12

45% proc chance on RC is static

So in this particular situation we will have basically 7 different scenarios that will change based upon the proc chance of RC.

Iteration 1 with 0 Blood Charges

45% you will proc making RC better

55% you will not proc making them even

Iteration 2 with 2 Blood Charges

45% you will proc making RC better

55% you will not proc making them even

Iteration 3 with 4 Blood Charges

45% you will proc making RC better

55% you will not proc making them even

Iteration 4 with 6 Blood charges

45% you will proc making RC better

55% you will proc making then even

Iteration 5 with 8 Blood Charges

100% BT will be better Even if you proc you will be at 10 Charges after 1 GCD

Iteration 6 with 10 Blood Charges

100% BT will be better

Iteration 7 with 12 Blood Charges

100% BT will be better

So you can very simple see that in 4 of the possible 7 situations here RC actually has the opportunity to be the best ability, however due to the 45% proc nature of the ability it only ends up being the best in 180/700 iterations while the two talents are equal 220/700 and BT is the best in 300/700.

So you could say that these were the values of which ability would be better in this particular situation.

RC would be the best talent 25.7% of the time

Both talents would be equal 31.4% of the time

BT would be the best talent 42.8% of the time

So you can see that even in a situation where BT only has 3/7 possible iterations to be the best its actually quite a good bit better than RC.

Here is another example

0.25U 0.25F 0.0B

60RP

0-12 Blood Charges

45% proc chance(going to modify this to 69.7% for the probablity that you will proc 1 out of 2

Iteration 1 with 0 Blood Charges

69.7% you will proc at least 1 RC will be better

30.3% you will not proc and they will be equal

Iteration 2 with 2 Blood Charges

69.7% you will proc at least 1 RC will be better

30.3% you will not proc and they will be equal

Iteration 3 with 4 Blood charges

69.7% you will proc at least 1 RC will be better

30.3% you will not proc and they will be equal

Iteration 4 with 6 Blood charges

100% BT will be better

Iteration 5 with 8 Blood Charges

100% BT will be better

Iteration 6 with 10 Blood Charges

100% BT will be better

Iteration 7 with 12 Blood Charges

100% BT will be better

So in this example RC actually only had the chance to be the best talent in 3 out of 7 possible scenarios however it did benifit from having 2 Rune strikes thus increasing its chance of procing 1 of those to 69.7% . This let it be the best talent in 209.1/700 while in the two were equal in 90.9/700 and BT was the best in 400/700.

So for this example we could say these were the %'s

29.87% RC was the best talent

12.98% the two were equal

57.14% BT was better

Now while I have often spoken about resources and which one generates more ect. the main arguement that people have brought to my attention is that RC will get you to your next DS consistently because every proc moves you closer. You are actually correct there but you using that basis to completely ignore all the other possible outcomes by simply stating that when it does occur its better.(this isnt even actually true there are many situations where even if you did proc your RC BT would still be better for example if i have 8+ Blood charges and we are 5 seconds from filling our DS runeset even if you proc BT will still be better)

Ive often heard someone say something like "well if you only have 1 RS what is going to lower the interval of your next DS?" I think I quite simply showed the results of a situation just like that in my first example of 30 RP where BT was actually quite a bit better than RC.

I very firmly believe that we can say that BT is vastly superior to RC in terms of being able to provide you with your next DS more consitantly for a couple of very key reasons.

1. Even in situations where RC has the opportunity to be the best talent its only that by a factor of 45% when you have 2 proc chances its 69.7% ect. However its not 100% of the time.

2. When BT has the opportunity to be the best talent it is 100% of the time.

3. Even when you proc RC it does not even mean that it will be the best talent 100% of the time.

You can take out a ton of these scenario breakdowns and see that in a vast majority of these situations and you will see that BT is superior in terms of the probablity that it will cut the interval of your next DS.

You made this statment.

That is RC's strength: frequency, not potency. Notice how I didn't use any algebra or resource analysis to find RC's strength, because some things in theorycrafting aren't just "ok punch the numbers".

Thats just not true you are using 1 iteration of a scenario to make your point and you are absolutely just plugging in your numbers let me show you how.

Your quote

If you are left with 1 FU pair and 0 charges of BT (had just finished using them) you will have to wait for the entire remaining duration of the other recharging FU pair before you get your next FU pair. Now look at RC, as long as you get a SINGLE proc (which requires at minimum 1 runes strike as opposed to the 5 needed for BT) you will already be ahead because no matter how much regen RC gives you the gap will be less than the full duration wait of that situation in BT.

Ok in this ONE iteration of this scenario and you are plugging in your numbers 100% and crunching them.

You are saying you are at

1.0U 1.0F 0.0B

0 Blood Charges

RP who knows you didn specify

45% proc chance

So you are saying in this 1 iteration of the scenario where you have 0 Blood charges and you are fortunate enough to Proc 1 RC that yes you will generate your next RC faster. I have never once contended that this was not true. However you cant just look at 1 possible iterations of an entire scenario where you make a 45% proc chance then pass that off as good statistical analysis. What about all the times you have 2/4/6/8/10/12 blood charges and the times you dont proc and so forth.

You keep hammering how the "when RC procs it cuts your interval" and you are right all im saying is that in any situation there are many many different outcomes based upon the factors and if you take 1 outcome based upon 1 set of factors and use that as your comparative opinion then you arent really looking at it properly because I could just as easily use the example that when I have 10 Blood charges it doesnt matter if you proc or not because im getting there before you anyway. The probablity of being at 10 stacks is exactly the same as being at 0 stacks so you see im doing the exact same thing as you are just on the polar opposite end so that it proves my point instead of yours.

What I did in these examples was show you the probablity of each ability cutting the interval of your next DS by more.

Let me repeat this:

Your statement isnt wrong but you arent looking at the overall picture you are looking at 1 small portion of the situation and using that as your basis for the ability being equal.

And my final statement that this isnt about resources generated or DS/min generated this is about in 1 situation where you needed to lower the interval of your next DS which ability would do that more consistently.

Edited by Jadawin

Share this post


Link to post
Share on other sites

Oh Jesus....

2 people like this

Share this post


Link to post
Share on other sites

Oh Jesus....

U found it tldr 2 didnt u
1 person likes this

Share this post


Link to post
Share on other sites

You're missing the point. There is one, and only one reason why some tanks choose Blood Tap. If you sit on your charges then you will always have 1 extra death strike exactly when you need it. It doesn't matter if RC proc's more death strikes per encounter, or statistically can sometimes proc your next death strike faster. You can not mathematically guarantee that RC will proc your next death strike at the exact moment where one more death strike is the difference between life and death. However, if you are sitting on BT charges then you CAN use BT for one more life saving death strike, 100% of the time.

That's it. One extra death strike that is there at the exact moment you need to save your life, once or twice per encounter. With BT you can guarantee it's there. with RC you can not.

Edited by Storm

Share this post


Link to post
Share on other sites

The issue with your attempt at a statistical model is that it splits all the iterations up, when in reality iteration 1 is connected to iteration 7. In other words, you have to get from 0 to 10 BT stacks in sequence, and it requires 5 rune strikes to do so. look at it from a "what does it take to get to the goal from here" point of view.

Iteration 1 with 0 Blood Charges

it will take 5 rune strikes for BT to get your DS hastened. with 5 RS you have a 99.993% chance of getting a proc at some point, and a 99.2% chance to get a proc before you would have gotten BT (so on the 4th or earlier RS)

Iteration 2 with 2 Blood Charges

it will take 4 rune strikes for BT to get your DS hastened. with 4 RS you have a 99.2% chance of getting a proc at some point, and a 90.85% chance to get a proc before you would have gotten BT (so on the 3rd or earlier RS)

Iteration 3 with 4 Blood Charges

it will take 3 rune strikes for BT to get your DS hastened. with 3 RS you have a 90.85% chance of getting a proc at some point, and a 69.75% chance to get a proc before you would have gotten BT (so on the 2nd or earlier RS)

Iteration 4 with 6 Blood charges

it will take 2 rune strikes for BT to get your DS hastened. with 2 RS you have a 69.75% chance of getting a proc at some point, and a 45% chance to get a proc before you would have gotten BT (so on the 1st RS)

Iteration 5 with 8 Blood Charges

it will take 1 rune strike for BT to get your DS hastened. with 1 RS you have a 45% chance of getting a proc.

Iteration 6 with 10 Blood Charges

BT is already at its goal

Iteration 7 with 12 Blood Charges

BT is already at its goal and will have 2 charges going into the next sequence of RS.

In this version, you can see that only 3 out of 7 iterations directly favor BT in terms of "quickening the return". You'll notice that it seems that i'm saying that 10 stacks of BT is equivalent to 1 RC proc, and that's exactly what i'm saying. Damage comes at you in bursts, so just because RC isn't instant doesn't mean that it's return is infinitely worse because it's not instant. RC proc time between bursts can be thought of as instant, which is why we look at "getting the benefit" as opposed to just "getting a full DS".

"There is one, and only one reason why some tanks choose Blood Tap. If you sit on your charges then you will always have 1 extra death strike exactly when you need it. It doesn't matter if RC proc's more death strikes per encounter, or statistically can sometimes proc your next death strike faster. You can not mathematically guarantee that RC will proc your next death strike at the exact moment where one more death strike is the difference between life and death. However, if you are sitting on BT charges then you CAN use BT for one more life saving death strike, 100% of the time.

That's it. One extra death strike that is there at the exact moment you need to save your life, once or twice per encounter. With BT you can guarantee it's there. with RC you can not. "

In situations where random burst damage is meaningful you can't guarantee BT will be there either. The reason BT is viable is the control it gives, not because it's a 100% guarantee, because it's not. You may need that saved charge before you get your other FU pair back up, which means you don't have anything saved.

BT is still very good, and the control it gives is great in all situations and mandatory in some (such as burst mechanic fights like horridon, jikun, sha of thrash, impale, ect.) because if you hold on to all FU pairs and save them for one moment, you can have 2 saved with RC used over 2 GCDS and 3 with BT used over 3 GCDs.

Share this post


Link to post
Share on other sites

The issue is still all about preference imho. Why would it matter which one is actually that small percentage better.

I'm quite good at math but I choose not to read all the calculation everyone is doing (in both this topic as the previous one) since I do not think it is that important.

WOW I have one more deathstrike per 5 mins because I use RC.

WOW You have your deathstrike on command because you use BT

Buy a lollipop and go celebrate.

I'm just gonna do what feels right.

1 person likes this

Share this post


Link to post
Share on other sites

"I'm just gonna do what feels right. "

believe it or not some people care very much what is optimal, and rightly so if they are pushing bleeding edge content.

1 person likes this

Share this post


Link to post
Share on other sites

"I'm just gonna do what feels right. "

believe it or not some people care very much what is optimal, and rightly so if they are pushing bleeding edge content.

I get that, but I've seen about 12 different ways of calculating it by now.

As a scientist this tells me two things.

- Everyone is right.

and

- Noone is right.

This may sound strange, but it is actually true. All of the math I've seen sofar has truth to it, but fact remains that we're dealing with two very different factors here. I still dont think any of the two is always better, I think it should be seen from a encounter to encounter perspective.

While RC may be better in FightA BT will be better for FightB.

Should we not be focussing more on when the one benefits more then the other instead of just make a steady stream of calculations noone is going to win anything with.

Please dont be offended, I might sound pissed, but I am a little. It is all just about a small difference, and I stick to my earliest made point:

- RC is better for fights with a steady flow of damage

- BT is better for soaking big bursts.

I dont need a chalkboard or two to know that.

You are right ofc. In a HC Raid setting, even the smallest numbers count towards the end.

I'm just sick of hoping someone has something new to tell and instead find more of the same.

Well I guess I'm just being bitchy about it, and I apologise for it.

Hope I didnt offend too many folks

Shine

Share this post


Link to post
Share on other sites

I'm not trying to say one is better. The issue at hand is whether RC is viable, and I'm simply arguing that the two are equally optimal so barring fight mechanics you can use either and be "optimal". I also gave quite a bit of abstract analyses without math in the other thread if you are truly tired of seeing numbers, and if you really don't care which is best for bleeding edge min/maxing then there really isn't going to be any value to you in this thread so there's no real need for you to follow it, especially if it's only making you angry to see the discussion.

Share this post


Link to post
Share on other sites

I guess you're right there, and I may have been jumping to comclusions a bit fast. I didnt even read half the posts in the end.

Goodluck figuring it out

Shine

Share this post


Link to post
Share on other sites

So I had to go back and read the previous posts to make sure that I was completely clear on how I wanted to approach this. So what I have gathered after reading the prior and current thread is that when we are talking about the comparison of BT and RC:

1. Both give the ability to save 1 full UF pair for reactive healing

Now since this is the case and we both agree that the total resources or DS/min is not the metric by which we should measure the survival of any of the talents then we are basically only left with 1 comparative scenario.

What ability will provide you with your next DS or cut the interval to your next DS more often or consistently. This doesnt have anything to do with total resources this has everything to do with taking any scenario wherein you took a spike without a full DS ready and you need your next DS now or where you just burned your reserve DS and you need to get back up to your healing pair pooled up either way thats what its about at its core.

Getting the next DS filled up, not how many over a fight its right now in this situation what talent has the better chance to get me there.

Now I put out some scenarios with multiple iterations based upon 4 very specific factors in any possible scenario.

1. Rune State

2. Runic Power

3. Blood Charges

4. RC proc %

Now my math was quite conclusive and you did some math of your own but you did a completely different comparison. I was comparing the probability of each talents ability to achieve my next DS faster. You did a comparison on the amount of RSs it would take to generate my next DS with BT in regards to the probability that RC would proc in that many RSs. You didnt take into account the state of your current runes or the amount of runic power you had at that time thus dictating the amount of RSs you were actually able to execute.

Perfect example is if we are 2.5 seconds away from filling our next DS and we are at 0 Blood charges Blood tap has a 0% chance to get you to your next DS faster than RC. However RC has a ~69.7% chance to achieve this with 60RP and a 45% chance to achieve this with 30 RP. Where you are getting lost is that in 3 globals the next DS will be here anyway what we are trying to establish is which talent will get us to that DS faster than 2.5 seconds, not how long it will take to generate a DS with blood charges.

Thats what this entire discussion is about at its very core. Which ability has the highest probability of giving you that next DS faster in a situation where you need it right now.

Your iterations are incomplete in the regard that they dont provide specific data as to how long it will be until your next DS would happen anyway.

We both agree on this 100% that measuring the power of an ability in terms of survival isnt about overall DS/min or resources its about very specific scenarios where you need that survival right now or as fast as possible or you need to have it ready because you know a Triple puncture, impale, talon rake ect. is incoming soon.

Now I could do thousands of iterations based upon the factors and make myself go all beautiful mind however I can undeniably prove that BT will achieve this more often than RC with a few statements.

1. ~12% of the time when you are less than or equal to 1 Global from your next DS only BT can be better in 28.57% of the iterations with 10 or 12 blood charges all other iterations the two talents are equal because obviously if you RS 1 time then the global fills your DS thus its a wash unless you have 10 or 12 Blood Charges then you can shorten the time before your next DS. So by this one statement we can say that:

a. BT will be the best in ~3.4284% of all iterations

b. They will be equal in ~8.5516% of all iterations

2. So we can say that we only really have 88% of all iterations left to make a comparison upon and of those 88% you can make a very specific statements.

a. In all scenarios of 10 or 12 Blood charges or ~ 28.57% of 88% or 25.1416% of all iterations BT is 100%

Now I think we could put both of these statements in the "Thats pretty obvious" column saying that anytime you have 10 or 12 Blood charges Blood tap is better or that you are less than 1 global away from your next DS then even casting 1 RS is actually extending the gap in your next DS rather than just waiting for it.

3. So we are now down to 62.8584 of all possible iterations left to compare and these are just the iterations of 0-8 blood charges that are more than 1 gcd away from filling up the next DS. Now this is where the math is far less accurate but you can still make some reasonably accurate statements.

a. In 20% of these iterations (8 rune charges) or 12.57% of all iterations ~72.72% of the time(with 30+RP) Blood tap will be better

1. In 9.1421% of all iterations BT is the best.

After these 3 statements making any kind of very specific statement about the % time of say 6 Blood charges with 60 RP gets much more complicated and drawn out due to multiple RC chances.

So lets recap in 3 statements I showed that BT was the best in 37.7121% of the time and they would be equal in 8.5516% of the time. So what we can take from this is that out of all iterations ~46.2637% of the time RC doesnt even have the opportunity to be the best talent.

So that leaves only ~53.7363% of all possible iterations that RC even has the CHANCE to be the best. Even in these scenarios there are going to be times with 6 blood charges and 60+ RP or 4 Blood charges and 90+ RP where BT will still pick up some % points in the equation.

There will also be a nice % of iterations where you have the abilities being equal when you have low Blood charges and you dont proc RC.

While I will not take this out to the thousands of possible iterations anyone can look at this and make this hypothesis:

Out of the remaining ~53.7363% iterations where there will be a mix of results where all 3 possibilities are true. While RC will be better than BT in a majority of situations there will be many situations where the 2 talents will be completely equal.

a. Example 30 RP 3 seconds from Next DS 0 Blood Charges 45% RC 55% even

For RC to be the better talent in a majority of scenarios greater than ~70.1799% of the remaining ~53.7363 iterations would have to prove/show RC to achieve your next DS faster than BT.

Coupled with that we would have to say that BT would cover 0% of those remaining which we know to not be true.

You can see that this is a statistical impossibility and if you were crazy enough to write a program to model this or if you were beautiful mind and did them all I believe you would find that on the average considering all possibilities that the % chance based upon all factors and scenarios possible would be something like this.

Blood Tap 45.7725%

Even 21.9856%

Runic Corruption 32.2417%

Now while I fully admit that these are guesses as to the total %'s and we could see some variance to these specific figures it will not be enough to tip the scales twards RC because if you look at the model that I went with. For the remaining ~53.7363 I went with 60% for RC which is I would say maybe even too generous of a %. Then 25% for even and 15% for BT.

Is this 100% accurate? no not at all. No one will do the programming work or the billion pieces of paper to actually figure this out but I would say this is reasonably accurate and for RC to gain enough to even be equal to BT would be an impossibility.

I can honestly say that with a very high probability that while my %'s of total are not 100% accurate that I have shown categorically that over the course of all possible scenarios the probability that BT will lower the interval of your next DS more consistently over RC is a statistical fact. Ergo if the probability that BT will achieve this more often than RC it would make BT the superior talent for survival.

Edited by Jadawin

Share this post


Link to post
Share on other sites

Your iterations are flawed because you are compartmentalizing everything, including the next DS. The reason I don't give a current rune state, or RP, is because rune state doesn't matter because any thing you get immediately before a DS will help you get the DS following the one you're currently working towards. You are already almost getting a pair back, and if you are that close your next goal should be the NEXT DS, so it really doesn't matter at all in terms of our RC vs BT analysis. This is on top of what i've already pointed out with you putting each rune strike in its own box. You say if you have 1 RS and 0 BT stacks now that BT would be better 55% of the time, but you're still 5 rune strikes away from a returned DS with BT as well and you're ignoring all the potential procs from those. To put it a bit differently, in your example of n BT stacks, you are assuming that you didnt proc RC for the n/2 rune strikes you used to get there, and the chances of that happening are slim for any n above 4. They both are equal there for that ONE rune strike if RC doesn't proc, but your sim is unrealistic because it doesnt take into account the following rune strikes and death strikes.

Share this post


Link to post
Share on other sites

Your iterations are flawed because you are compartmentalizing everything, including the next DS. The reason I don't give a current rune state, or RP, is because rune state doesn't matter because any thing you get immediately before a DS will help you get the DS following the one you're currently working towards. You are already almost getting a pair back, and if you are that close your next goal should be the NEXT DS, so it really doesn't matter at all in terms of our RC vs BT analysis. This is on top of what i've already pointed out with you putting each rune strike in its own box. You say if you have 1 RS and 0 BT stacks now that BT would be better 55% of the time, but you're still 5 rune strikes away from a returned DS with BT as well and you're ignoring all the potential procs from those. To put it a bit differently, in your example of n BT stacks, you are assuming that you didnt proc RC for the n/2 rune strikes you used to get there, and the chances of that happening are slim for any n above 4. They both are equal there for that ONE rune strike if RC doesn't proc, but your sim is unrealistic because it doesnt take into account the following rune strikes and death strikes.

Here is why you are incorrect at its most simplistic basis. You can not statistically show that just because something happens more frequently and at a lower amount of effectiveness that it accomplishes the same thing because it happens more often. You can state that every proc of RC does in fact reduce the interval of your next DS which is true but that in no way can be used as a correlation on its effectiveness compared to BT because it doesnt even statistically tell you anything other than every proc does something. I could say that every time I am at 10 Blood charges I can get a free DS. Im saying exactly the same thing as you which is absolutely nothing because that has no bearing in the statistical comparison of BT to RC.

You have said yourself many times that the entire point of taking RC over BT would be the frequency and consistency of your returns and I should stop worrying about the amount of the return. This is so completely ridiculous that it makes me want to explode my head. By that rationale any if you took RC and made it 100% proc chance but made it have a .1 second uptime then it would be better because regardless of the return you were 100% consistent and frequent in your application of the buff because the amount of the return was less important than the frequency of how often you benefited from it.

I am going to quote you here

When you stop looking at the throughput, and instead look at "how often will these talents give me benefit, regardless of how much that benefit is".

You can see that this is a completely silly statement because while frequency is important is is not the only means for measurement at all and you are using this as your gold standard defense and I cant believe that you cant admit that while you are right in your statement it in no way means anything about the comparison of the two abilities or my prior statement about 100% proc chance and 0.1 second uptime would be true.

Im going to quote you again

You'll notice that it seems that i'm saying that 10 stacks of BT is equivalent to 1 RC proc, and that's exactly what i'm saying. Damage comes at you in bursts, so just because RC isn't instant doesn't mean that it's return is infinitely worse because it's not instant. RC proc time between bursts can be thought of as instant, which is why we look at "getting the benefit" as opposed to just "getting a full DS".

Really? Come on now they arent even close to each other in value 10 stacks of BT is without a doubt so much more valuable than 1 proc of RC its just crazy to suggest otherwise. I any single iteration of any encounter 10BT is vastly more powerful than any 1 proc of RC. You could say that while it takes you 5 rune strikes to build 10 blood charges thus the build up makes it less valuable but we arent talking about that we are talking about an infinite number of scenarios, and in each scenario your only goal is to achieve that NEXT ds as fast as possible and which talent will get you there a larger percentage of the time.

What I really love about having discussions like this is that when you are working your brain to really get to the bottom of something you even start finding holes in your own reasoning and you have to start making adjustments to your calculations. I hope that the examples of frequency and return make you see the flaw in that line of thinking but im shocked that you didnt point out the HUGE alarming problem with my iterations. While I dont agree with your reason for them being wrong I left out maybe the biggest thing that we both should have picked up on really quickly and that is the % probability that you already have a RC proc running on you and the % of the buff that is currently left still ticking. I cant believe we both didnt pick up on this.

I will quote you again

Your iterations are flawed because you are compartmentalizing everything, including the next DS. The reason I don't give a current rune state, or RP, is because rune state doesn't matter because any thing you get immediately before a DS will help you get the DS following the one you're currently working towards

This is categorically not true this goes against the very basis of theory crafting because this isnt a race over multiple DS's if it was a race of building multiple DS over the course of multiple Rune strikes BT would decimate RC remember this entire arguement was that you are trying to state that while BT has the ability to generate ~48% more DS runes over the course of a fight that RC will be as effective as BT in terms of survival. Perfect example of this is if we started at 0.1U 0.2F 0.0B 30 RP and 10 Blood charges. So by your rational because I already executed my next DS is the proc of RC cutting the interval of the current RS or the next death strike? Cant really be doing the next one can it because you havent even filled this one. This is why you cant compare them in this manner or you turn it into a total DS resource race and BT will decimate RC in every discussion like this.

Again its really only about 1 thing and that is that there are "x" number of scenarios(infinite of course) and in all of those scenarios there are a number of determining factors that will dictate which talent in each scenario has the greater probability of filling your next DS the fastest. Now with this new insight I have about having RC already up when the snap shot starts will give me a ton more work to do in my comparative statistics because its going to change the math a pretty good bit.

Ill wait a couple of days so that if anyone has any other insight in to what I may have missed in terms of things that will dictate DS interval in a specific situation I can add them to my calculations.

Share this post


Link to post
Share on other sites

You're not accepting that RC's proc has value because it's not instant. Until you get over that, you will not get the correct answer. I'm sorry, that is just the case.

"This is categorically not true this goes against the very basis of theory crafting because this isnt a race over multiple DS's"

Lol ok if you want to let me know when the basis of theorycrafting was decided to be the next immediate death strike that'd be great. Theorycrafting is about practical and real situations, and in no situation are you literally only thinking about the very immediate death strike only. That doesn't mean you are thinking about all the death strikes in the whole fight at once either, if that were the case we would be using RE because of the throughput and you were already proven wrong on RE being viable. In reality you look at the next few death strikes. You're looking at the next 10-20 seconds and making sure you can time around what comes at you right now and a bit into the future. If you don't accept this and continue to only look at compartmentalized and flawed "statistics" then I'm not even going to try to counter your argument because your simply not accepting what i have to say. I've said everything there is as to why RC is viable, and if you choose to go back and piece it all together then maybe you'll learn something.

Keep in mind this isn't just my opinion, literally every other high end DK theorycrafter would tell you the EXACT same things I have.

Share this post


Link to post
Share on other sites

Keep in mind this isn't just my opinion, literally every other high end DK theorycrafter would tell you the EXACT same things I have.

Except the almighty Jadawin of course. It makes me wonder if Jadawin raids on a DK at all because his assumptions, arguments and posts seem to be completely devoid of all situational circumstances. I understand that your math is strong and backs up what you think is right, but it almost seems like you are trying to force this square into the circle hole. It slightly reminds me of the people who religiously looked at Simcraft single target sims when over half the fights in Throne have some sort of add/multiple add's mechanic. I think Reniant put is best when he said tanking is not about the immediate DS or Shield Block or ShoR or whatever. It's about the next 20 seconds of the fight. I mean just the fact that top DKs in the world argue that RC is as acceptable as BT AND use it on Heroic mode fights without issues should be enough for you to realize that its a solid talent. So what that the math shows over the course of a fight you get 3 more death strikes with BT or whatever. (And I made three up, please don't try and get all mathematical on a post that has no math)

Edited by Drafty53
1 person likes this

Share this post


Link to post
Share on other sites

"It makes me wonder if Jadawin raids on a DK at all"

Let's keep the discussion based on logic, not ad hominem arguments. He's been civil so far in this thread so there's no reason to be uncivil back. How much he raids shouldn't matter, just like how how much I raid shouldn't matter. A solid and wide experience of the game at a high level will certainly help shape someones viewpoint and help them think logically in context of raiding. Take the best starcraft theorycrafter, and have them theorycraft WoW and they will surely get certain some things wrong. Some things don't always make logical sense until you actually sit down and experience the reality in which the theorycrafting is being done for, but if you are familiar with that reality (which many people are even if they don't raid anymore or at a high level) then you have no disadvantage. Either way, progression doesn't dictate theorycrafting.

Share this post


Link to post
Share on other sites

Well, I see you appear to have lost none of your charm!

First of all, the "BT has the ability to generate ~48% more DS runes over the course of a fight" was annoying me. (Edit: Was saying something here along the lines of more variables and possibilities to improve RC return but then reread and saw your own 27% quote, though you made it sound like a not-so-often get lucky scenario. With the 10k iteration sim from later on it is well averaged out these cases and still gives about a 27% disparity).

I'd have to say a more accurate and simple test would be in simcraft to actually take everything in our regular tanking arsenal into account. Using the sim, the Death Strikes will be used as soon as available, which ensures no capped resources, but results in identical numbers to a player playing their runes correctly as a tank (keeping a set available but using them before the other pair caps) as no resources are wasted.

Setup: 450 second duration, no time deviance, 10000 iterations, and a few proirity list edits to better maximize proc gaming and such.

none: 47.0 ds, 6.27ds/min

rc: 66.4 ds, 8.85 ds/min (+19.4 ds)

bt: 71.6 ds, 9.55 ds/min (+24.6 ds)

re: 73.1 ds, 9.75 ds/min (+26.1 ds)

More like 27% than a 48% lead. I as well hold that ds/min is not the factor we care about, but just felt that we should be correct about that as it does make it a bit more favorable than you presented.

So in regards to the current debate, it looks like you're still cherrypicking the "death strikes that matter" to favor BT. Reniat already nailed it with the fact that without sitting on a 12 charge waiting for "that" death strike you need, a nice 10/12 charge BT will not always be guaranteed right when you need it for that unplanned spike damage. In the BT cycle of 0 to 10 charges, you will have had 5 RS to get there. In that time, you will have no chance of quickening any of those few deathstrikes in between it. What if one of those few death strikes that happen in between is "that" deathstrike you need to stay alive after some bad damage intake/avoidance rng streak? Look at RC though, by the 5th RS you have a:

~95% chance to proc at least 1

~75% chance to proc at least 2

~40% chance to proc at least 3

~13% chance to proc at least 4

~2% chance to proc all 5.

(Edit: Something I initially forgot to add, but the RC sample will have a higher RS use, meaning that in some cases the window of the 5 RS will be shorter or simply have a 6th RS, since both were simmed at the same time length. This adds a bit in favor of the chances to proc RC in the same time it would take to build 10 charges with BT.)

I didn't zoom in on individual stack counts and rune strikes or whatever, because that doesn't mean a thing to us when we're trying to demonstrate consistency. This 0 to 10 cycle is what we'd see over and over if we're playing with BT as you describe (saving to 10+ and using them for that full DS), and the numbers shown for RC look pretty good to me when I'm looking for the most possibilities (and only the most possibilites) to hasten a death strike in that window.

This is categorically not true this goes against the very basis of theory crafting because this isnt a race over multiple DS's if it was a race of building multiple DS over the course of multiple Rune strikes BT would decimate RC remember this entire arguement was that you are trying to state that while BT has the ability to generate ~48% (27%) more DS runes over the course of a fight that RC will be as effective as BT in terms of survival.

You are absolutely right here, though it didn't argue with what he said about RC either. It's definitely not a race about building the better number of DS, it is well understood that within the same number or RS you have invested more value in BT than the strictly quantitative value of RC. It's the qualitative property that takes a bit of familiarity with to understand it's worth that provides the survivability described in these windows of rune strikes.

That is the consistency and frequency we want when we take RC and say it can be on par for survivability. We have a much greater chance at hastening "that" death strike we need, and not the one down the road. Obviously it's not a full DS in one proc (don't try to tell me this, I know just well), but that's the tradeoff we're talking about: the burst BT gain vs the frequent/smaller RC gain.

Bill gets paid every other the day, and makes 25$ each payday starting on Monday and ending Friday. Jack gets paid 100$ on every Friday of the week. They both pay their bills on Friday and start the week at 0$. On Tuesday, they both realize they need a gallon of milk. Bill is good to go the very next day, while Jack waits till the end of the week to buy his milk, but by that time his wife has left him since he had no milk for their cereal. (lolwut I dunno).

Bill is of course RC and Jack BT. Of course we're not accounting for probability of him getting paid, because we already know the average cases and number of procs to expect, no matter how little the individual proc's worth. That gallon of milk is the nasty hit from whatever raid mechanic + bad rng boss hit that we as dk's have to worry about. The wife leaving is whatever the non-optimal outcome of the event might be, whether it's salvaged by a big healer cd and flash heals, or ends with a run back after you go down and wipe the attempt. Can you move needing the gallon of milk to Friday and have Jack surprise the whole family with 3 gallons of milk? Yes, and you can also move it to Saturday where by the next Friday, his wife would have left and Child Services came and took the kids away as well.

Edited by Thejazz
1 person likes this

Share this post


Link to post
Share on other sites

Edit: Got so caught up with trying to actually make this thing work before I realized I was even doing it wrong. I guess the main goal of this now is just to add to what was already said about the reasoning as to why these kinds of purely mathematical approaches don't work when we're talking about talents that are chosen based on it's qualitative properties. Before I came to this, I had tried a few things, and just for the sake of it:

...

Iteration 5 with 8 Blood Charges

100% BT will be better Even if you proc you will be at 10 Charges after 1 GCD

Iteration 6 with 10 Blood Charges

100% BT will be better

Iteration 7 with 12 Blood Charges

100% BT will be better

One was the: "100% BT will be better". You can't throw out the value of the 45% chance to proc on the last ones just because. Think it was already brought up, but since this should be to compare how the talents can benefit us, and more specifically making our argument that RC can be viable vs BT based on frequency, that we can't leave out chances for it to benefit. Basically this line should look more like:

(since BT is 100% here)

45% you will proc RC making them even

55% you will not proc making BT better

So then we have:

25.7% that RC is better

23.6% that BT is better

50.7% that they are even

My next issue was that from iteration 5 and onward, things get a bit too shaky. This may sound weird and I really am unsure how to explain it but think about this; after the RS's gcd on iteration 5, there is theoretically an infinite number of instances where BT can give a benefit and be "better" since it is off the gcd, and only 1 finite instant where RC can even compete with a chance to proc. At any possible time, and I mean any infinitesimal time offset from the gcd ending, you could use BT if a situation arose that needed it. Of course this sounds silly since no human could ever create such number of instances, but it's mathematically true. So taking 7 scenarios, adding the proc/benefit chances for each and averaging them out doesn't work anyway. I mean, obviously BT is not infinite times better then RC is it?

But yea as I said, scratch that. These issues (and so much more missing I was starting to see just thinking about it) add quite a complication to this approach and pretty much render this approach unusable. As I said, I was really trying to figure it out and make it work, but really... it can't be done, and it won't be it happening even with the additions and alterations you were setting out to make. Seeing as this iteration-by-iteration analysis was what most of the previous posts' info and assumptions were based off of, it just disturbs it all and I think we're back to square one on how to mathematically compare and analyze something that can only reasonably be justified by simple probability (in my 1st post) and that familiarity and gut feeling of what to expect as a blood dk.

Edited by Thejazz

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.