1. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by MouseDK
Could you not be turning a hit or block into a parry? (Forgive me if I am reading the attack table wrong)
That is what Lizana is saying. A block (mitigated hit) has the potential to become a parry and a hit (unmitigated) has the potential to turn into a block (mitigated hit).

2. I try
Join Date
Oct 2007
Posts
253
I'm trying to think of a way for you to understand how the RNG requires a statistical analysis and not the turning hit into block thing.

Lets try considering what the game actually does (or would have to do for your statement to be true). For it to "turn a hit into a block" the server would have to randomly generate its number, check what the result would be on the attack table then do another check of what buffs are up that effect the attack table then recalculate the attack table then apply the same initial number that it generated and then turn it into what the new result would be.

What I'm saying is that the game looks at your combat table once when it's instigating a swing. The only thing you can compare is what one combat table looks like (with buff) vs what another combat table looks like (without buff) at the moment the attack is made. Each swing that occurs has the full range of 100 options.There is no turning one outcome into another outcome, just apply random number vs apply random number.

3. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by Superspy23
I'm trying to think of a way for you to understand how the RNG requires a statistical analysis and not the turning hit into block thing.

Lets try considering what the game actually does (or would have to do for your statement to be true). For it to "turn a hit into a block" the server would have to randomly generate its number, check what the result would be on the attack table then do another check of what buffs are up that effect the attack table then recalculate the attack table then apply the same initial number that it generated and then turn it into what the new result would be.

What I'm saying is that the game looks at your combat table once when it's instigating a swing. The only thing you can compare is what one combat table looks like (with buff) vs what another combat table looks like (without buff) at the moment the attack is made. Each swing that occurs has the full range of 100 options.There is no turning one outcome into another outcome, just apply random number vs apply random number.
Super can you please supply your statistical analysis to validate this reasoning? Because what everyone else is saying is if the same number is rolled. Yes it is all RNG but what would that 80 look like with or without buff applied, 0r 61, the 50's and 20's would most likely still remain the same since bladeward doesn't affect those numbers.

4. Banned
Join Date
Apr 2008
Posts
1,399
And that right there shows you do not understand the basic premise of my posts. You can quite easily look at the combat table and see what the results if you had a buff up and if you didnt have the buff up. The values change by a set amount.

Now stay with me here, heres where you are totally missing the point, Say the game did its die roll and it rolled a 55. You can quite easily see where 55 was on one table, and then see where that same exact attack would have been on the other table.

An attack that would have been a regular hit, would not become a parry, dodge, miss with a blade ward proc. I am not saying the game does 2 different checks. I am saying you can easily see that that by adding in the buff on that attack. If you could magically have it added to the combat table, you will not turn an attack that would have been a hit into a parry. The attack you parried with blade ward up, if you didn't have the buff up would have been either a parry or a block. Never a miss or a dodge or a regular hit.

5. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Tarigar
Super can you please supply your statistical analysis to validate this reasoning? Because what everyone else is saying is if the same number is rolled. Yes it is all RNG but what would that 80 look like with or without buff applied, 0r 61, the 50's and 20's would most likely still remain the same since bladeward doesn't affect those numbers.
We're trying to compare what happens in the game when the buff is up vs when the buff isn't up. Those are not using the same random number in the game but they are using the combat table. Since we can know the combat table results of each scenario we have to simply compare the difference in tables as a whole, not just the picking one number and applying it to every situation. Basically we are looking at what that 80 looks like and the 61, all the 50's and the 20's as well as every other possible number equally. The net result is that you receive less hits and more avoidance in this situation. Hope that helps.

6. Banned
Join Date
Apr 2008
Posts
1,399
We're trying to compare what happens in the game when the buff is up vs when the buff isn't up
So then why use a different attack for when the buff is up and when it isnt. A true comparison of the buff would be to compare it to every possible outcome on the table and see what changes it would make for every possible attack. And when you do that you will see a hit becomes a block, a block becomes a parry, Hits never become parries.

7. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Lizana
Now stay with me here, heres where you are totally missing the point, Say the game did its die roll and it rolled a 55. You can quite easily see where 55 was on one table, and then see where that same exact attack would have been on the other table.
Show me an instance of the game using the same exact number in the next roll instead of generating a new number and I'll concede the point to you.

You can't call a random value fixed.

8. Banned
Join Date
Apr 2008
Posts
1,399
I am not talking about the next attack, i am talking about using your exact same attack. I dont care about the next attack, i care about the attack that i already took. I am looking at it after the attack has landed, I dont care what it has the possibility to do in the future, i want to know what it just did for me.

Show me one time the die roll changes after its rolled and i will worry about using different values

9. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Lizana
I am not talking about the next attack, i am talking about using your exact same attack. I dont care about the next attack, i care about the attack that i already took. I am looking at it after the attack has landed, I dont care what it has the possibility to do in the future, i want to know what it just did for me.

Show me one time the die roll changes after its rolled and i will worry about using different values
Thats only going to be of use for evaluating one hit and only one hit. This method cannot possibly come to a proper overall qualitative statistical analysis of all this enchant does.

Now lets say for arguments sake that reasoning just one and only hit was especially important. Is there some way to see which specific number was rolled by the server? Can we use this knowledge of one numbers result and negate all other results to evaluate overall usefulness in a raiding environment? Can this one result provide reliable statistics to every application of the buff? Get where I'm going with this?

10. Banned
Join Date
Apr 2008
Posts
1,399
No, but i can say without any doubt that you didnt turn a regular hit into any avoidance. And i can also say that as long as you have more block than blade ward provides in parry, that you will never have an attack that would have been a hit turned into avoidance. To state that your turning a hit into avoidance is a flat out 100% falsehood. Also i can say any parries you gained with the buff (parries you would not of had otherwise) would have been a block if you did not have the buff up. This means you are turning mitigated hits into avoidance and regular hits into mitigated hits. You at no point at all are turning regular hits into avoidance. Therefore you are not reducing damage by nearly as much as you think

As i have said, your still attacking me, instead of doing any work yourself to prove me wrong. If you want a "overall qualitative statistical analysis of all this enchant does" then make it. (as a side note i am pretty sure you mean quantitative instead, but thats just nitpicking)
Last edited by Lizana; 06-24-2009 at 12:35 PM.

11. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by Superspy23
Thats only going to be of use for evaluating one hit and only one hit. This method cannot possibly come to a proper overall qualitative statistical analysis of all this enchant does.

Now lets say for arguments sake that reasoning just one and only hit was especially important. Is there some way to see which specific number was rolled by the server? Can we use this knowledge of one numbers result and negate all other results to evaluate overall usefulness in a raiding environment? Can this one result provide reliable statistics to every application of the buff? Get where I'm going with this?
I know this. We will know the given range of what had occured. What we are doing essentially is shifting and minimizing/maximixing the range of different factors on the table. But the range/shift of other factors will remain unchanged. So for example:

Base Avoidance
001-012=miss
013-037=dodge
038-055=parry
056-075=block
076-100=hit

001-012=miss
013-037=dodge
038-058=parry
059-078=block
079-100=hit

So what happened is we decreased the range to be hit, shifted our range to block, and increased the range to parry.

So in summary there is the potential that 6 different rolls will be different from what would have occured from the base. But in reality 94 of rolls will be exactly identical to the previous.

So what is the real difference if Mongoose and Bladeward are the same? Well we are just shifting more figures if we use mongoose.

12. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Lizana
No, but i can say without any doubt that you didnt turn a regular hit into any avoidance. And i can also say that as long as you have more block than blade ward provides in parry, that you will never have an attack that would have been a hit turned into avoidance. To state that your turning a hit into avoidance is a flat out 100% falsehood. Also i can say any parries you gained with the buff (parries you would not of had otherwise) would have been a block if you did not have the buff up. This means you are turning mitigated hits into avoidance and regular hits into mitigated hits. You at no point at all are turning regular hits into avoidance. Therefore you are not reducing damage by nearly as much as you think
Ok say whatever you want then as long as its recognized that this has no bearing on the statistically based quality of the enchant in a raiding environment.

Do you realize that by saying "no" at the start of your post you agree that your method of analysis is worthless to evaluating the usefulness of this enchant?

You can take a past result, layer a bunch of buffs that effect the attack table then try to think of what the past result would have been given new modifiers but that doesn't work in WoW and won't do anything for evaluating how its likely to work in game.

13. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by Superspy23
You can take a past result, layer a bunch of buffs that effect the attack table then try to think of what the past result would have been given new modifiers but that doesn't work in WoW and won't do anything for evaluating how its likely to work in game.
Why not?

If you are decreasing a range of something ot occur isn't it less likely to be duplicated?

Hence the purpose between being hit capped and expertise capped? Are we not essentially pushing misses and dodges off our combat table?

You may not be able to say what would exactly would occur but you can say I had the potential to be struck in combat 3 less times.

14. Banned
Join Date
Apr 2008
Posts
1,399
Is there some way to see which specific number was rolled by the server? Get where I'm going with this?
That is what the No was in answer to.

You keep throwing phrases like "statistically based quality of the enchant". All of my numbers are based on stats, and they are about the quality of the enchant. Now are they measuring what you feel needs to be measured to be considered qualitative, no they are not, but the do meet whats required for quantitative data.

You can take a past result, layer a bunch of buffs that effect the attack table then try to think of what the past result would have been given new modifiers but that doesn't work in WoW and won't do anything for evaluating how its likely to work in game.
Actually i view that as a great way to view somethings effectiveness and judge its value. Looking at past results and changing them for new values is what theorycrafting is built on.

15. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Tarigar
So in summary there is the potential that 6 different rolls will be different from what would have occured from the base. But in reality 94 of rolls will be exactly identical to the previous.
Thats right. We can compare combat tables and see these changes. However that doesn't reflect probabilities well. We're focusing on 6 out of 100 numbers where each number has exactly equal probability of occurring. You can look back a past result and apply different values and call it turning it into something else but that's not what the game does and therefore has no place in evaluating probability based qualities of the buff where all options are possible each new swing.

16. Banned
Join Date
Apr 2008
Posts
1,399
What your not grasping with your probabilities, is only those 6 values have any change. So the effect of the buff will only be of use if you get one of those 6 values. Every other value has no benefit from the buff. If you dont get one of those 6 values, the buff did nothing at all for you.

And you can and should use past performance to predict future results. Were not saying the game will go back in time and change the results, but instead you can judge based on those past results changed with the new values a reasonable expectation of future results.

17. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by Superspy23
Thats right. We can compare combat tables and see these changes. However that doesn't reflect probabilities well. We're focusing on 6 out of 100 numbers where each number has exactly equal probability of occurring. You can look back a past result and apply different values and call it turning it into something else but that's not what the game does and therefore has no place in evaluating probability based qualities of the buff where all options are possible each new swing.
So essentially what I interpret from this.. "There is no effective way to measure enchants".

And like Lizana said changing the probability of something to occur for future value is statistics and theorycrafting.

If you look around you there is probability based decision making occurring everywhere in real life. And they are taking all of this from past data and skewing the results. By A shifting the values or adding new values.

18. I try
Join Date
Oct 2007
Posts
253
Originally Posted by Lizana
Actually i view that as a great way to view somethings effectiveness and judge its value. Looking at past results and changing them for new values is what theorycrafting is built on.
But the true value of the number is a range (its random and will be random every time the game uses it) to try to look back and not continue to evaluate that number as a range is faulty reasoning. Its not a proper application of the mechanics. Basically you'll only have a 1 in 100 chance that your backwards evaluation will still work in any other situation. I would prefer finding the numbers that are applicable every time.

19. Established Registrant
Join Date
Jun 2009
Location
Posts
367
Originally Posted by Superspy23
But the true value of the number is a range (its random and will be random every time the game uses it) to try to look back and not continue to evaluate that number as a range is faulty reasoning. Its not a proper application of the mechanics.
The problem with this statement is. Mechanics were brought into question. We know how the mechanics work. Since the game is all based on RNG it is all probability.

Since we know how the mechanics work... i.e. a combat table. We are changing the probability for that to happen again on that combat table. By increasing A to occur it will reduce the chance B will occur.

For example I can take a die and roll it until I get every number. I take each side and divide it by the number of total rolls and the values will automatically be assigned as a probability of something to occur. This is past data.

Then lets stick a magnet on the table and part of a magnet on one side of the die. Because we want that X number to increase in probability to be displayed.

Since the magnet is on one side it will automatically land on that side with the magnet more. Increasing the chance that the number on the opposite side to be displayed.

But if it does not land on that side with the magnet it will not automatically flip to the magnet side so I am not changing the probability of that value.

So by changing a given characterisitic in the object then I am decreasing the results for on thing to occur and increasing the results of another to occur.

Hence what we are doing here.

20. New Registrant
Join Date
Jun 2009
Posts
7
Personally with a fairly good parry chance already I have seen a bit of an uptake in my TPS with this enchant, However Mogoose has it's own set of nice bonuses because of it's consistancy, but its hard to judge one versus the other.
Where mongoose didn't have much effect on my dps, bladeward averages about 1% of my total damage over the course of a raid. at the same time I saw a slight drop in my average number of dodges with bladeward v mongoose I also saw my avg number of parries rise not alot maybe by 5 or so but thats something.

I don't think I could honestly tell someone one way or the other which way to go, but get whichever one you like.

With regards to what you guys are "arguing about" Statistics lie, if you base your decsion on enchants based on raw stats Bladeward would win every time, simply because a greater possible chance to avoid is possible and could occur however in practice this rarely happens. Even so a chance is still a chance and a greater chance could save your life. Comparatively i still have to say thanks for the cop out on any real tank enchants again, if a 70 enchant is even EQUAL to an 80 enchant you are doing something wrong.
my .02
Last edited by TokiWartooth; 06-24-2009 at 01:22 PM.