Defensive shield tank spec - the maths

[-naetha-]

This is a thread for everyone asking about paladin specs recently and just to satisfy my own curiosity

Most shield tanks want the best percentage chance to defend a melee attack (I guess anyway). I was playing about with numbers earlier and this is what I came up with - a good reasoning why NOT to go 50 shield. The maths might be a bit dodgy, but I'm pretty sure it works out Btw this is for a paladin (assuming they go 46 chants) but I'd have thought the same would go for other tanks.

With my spec of 42 shield, 24 parry (I got autotrained points) you have a 25.5% chance to block and 16.5% chance to parry - overall = 42% chance to defend. 50 shield spec = 29.5% chance to block, Parry is variable, you can go medium slash (34 say) and get 7 in parry, or low slash (29) and get 19 in parry. This gives you either 8% chance, or 14.5% to parry, overall = 37.5% or 44% chance to defend, but this SERIOUSLY reduces your damage output and you hit like a soggy baguette, just for 2% extra chance to defend.

And if that hasn't got you convinced, as points over 50 only count for 10% of what they count for under 50, 50+11shield = 31.05 chance to block, 42+11 shield = 29.65. So not very much difference at all

As I said, I'm not 100% sure on the maths, if you honestly think there is a better spec, then tell me!

Edit: sorry bad maths

 

[talen_sun]

Originally posted by -naetha-

With my spec of 42 shield, 24 parry (I got autotrained points) you have a 25.5% chance to block and 16.5% chance to parry - overall = 42% chance to defend.

Unfortunatly thats incorrect.

You don't have an overall chance of 42% chance to block.

First, it rolls on the chance to hit.
Then it rolls on the chance to parry (16.5%) if you DO NOT parry,
it then roll again to see if you block (25.5%)


Imagine it like this - you're playing a tabletop RPG - and you character has a percentage chance to block.

You roll 2 x 10 sided dice. The figure on the dice 1 is Tens. The figure shown on dice 2 is units.

If I have a 42% chance to block the highest i can roll in dice 1 is a 4 and on dice 2, 2. (42)

Because they are separate chances to block/parry the most I can roll on dice 1 for the parry is 1, and on dice 2, 6. (16)

Then I would roll for the shield block max of 2 and a 5 (25)

I have a greater chance of stopping the blow with 42% than what I do with a 16% then 25% chance.

The 2 chances to block cannot be added together to make a higher chance to block because they are separate actions.

 

[old.windforce]

i think you cannot add the numbers like that

when it its an independent change the math should be
(for easy number)
25% change to block
25% change to parry

change to defend:
100% - (75% x 75%) = 43,75% to defend

75% change to block
75% change to parry
100% - (25% x 25%) = 93,75% to defend, not 150%

for the best change to defend it is way to optimise the product of the 2 figures

3 x 3 = 9
2 x 4 = 8

so best to even out parry / shield then to put 1 skill significatant higher then the other.

This effect is furter strengthened by the higher cost of putting a skill higher then the other

you can get 50 shield 28 parry or 42 shield and 39 parry for the same amount of specpoints. So if you are not after a certain style (like butalise) it is best to spec 42/39 imho


Not sure if game mechanics works like this.

<NERF POSTING 1 SECOND before me>

 

[Wermin]

Hehe i think u forgot about one spec Naetha...
50 Shield, 39 Slash, 38 Chants and 18 parry.

I must say that the Brutalize style is much better then slam IMO coss even tho Pallys has end chant they get ooe very quick if they miss 1 or 2 slams and hit the 3rd.

Only thing i miss with 38 chants is better end chant, however second best seems to be enough.

 

[chi_]

Parry is gimped anyway, only works 1v1.

Shield >>> Parry at the moment

 

[Goryk]

The problem here is that parry and block have other things which take affect. With a alrge shield you can block 3 people attacking you without taking any penalties, whilst parry does not have that benefit. From what I've read, parry is bugged in that it reduces your chance to parry depending on the number of group mates your opponent has wehther or not they are physicaly in combat with you. For that reason, It makes sense to weigth your spec towards shield ratehr than parry. If that doesn't convinve you, the fact that you get shield styles, and no parry styles should. If parry was ever fixed, it might be worth looking at raising it but as it is, high parry only ever has much of an effect vs a lone opponent, and they tend to be the moniority of situations...

 

[chi_]

I think that if there are 8 people attacking your group, melee or not, your chance to parry is divided by 8, making your effective chance about 3-6%
As Goryk said, parry works wonders 1v1 and PvE, but is useless is almost every other situation.

 

[Goryk]

That math is still a little wrong or maybe just na little misleading, windforce as the two events aren't independant...


The actual chance to block is not the straight X% derived from the shield spec, but is the chance to fail to parry AND succeed the block so you are looking at the conditional probability of blocking given the fact that you did not parry. Shove that value for blocking into your equation and it should work I think

(Statistics has never been my strong point though)

 

[-naetha-]

Sorry I forgot to take into account that they're independent events and the statistics shenanigans (N in A-level statistics ) But as Windforce says its still better to have for example 42 shield 39 parry than 50 shield, 28 parry. And aren't they fixing the way parry works in 1.60?

 

[Javai]

You have also to bear in mind that with shield you get double defense cos you can also be guarding someone. So the maths only works if you are talking about 1 v 1

My Pally is specced 50 slash 42 shield 36 chants and 15 parry. She is flawless for pve and useful in rvr (I'm just bored of playing a tank and so she doesn't come out much).

She suffers from low hps and weapon skill (thanks to Mythics idea of what Paladin hit tables and stremgth cap should be) but she's one hell of a main tank.

 

[chi_]

Originally posted by Javai
My Pally is specced 50 slash

Why? You'll never actually do much damage?

EDIT: Forgot I was talking about a paladin/endurance

 

[talen_sun]

Originally posted by chi_
Why? You'll never actually do damage, and Amy --> Amy is better than Amy --> Diamond

How come? Show me proof please?

When I have a pally in the group, endurance is irrelevent.

With Amy -> Amy i do the same(ish) amount of damage twice.

With Amy -> Diamond i do more damage on the second hit.

So, why you say what you say??

 

[Danya]

Originally posted by Goryk
That math is still a little wrong or maybe just na little misleading, windforce as the two events aren't independant...

Yes they are. Failing (or succeeding) a parry does not affect your chance to block.

 

[Javai]

Originally posted by chi_
Why? You'll never actually do much damage?

EDIT: Forgot I was talking about a paladin/endurance

I went to 50 slash in order not to constantly miss rather than for the diamond slash style. Although since we Paladins got end chant I do use diamond slash before that the end/damage ratio wasn't worth it.

 

[decline]

Originally posted by -naetha-
And if that hasn't got you convinced, as points over 50 only count for 10% of what they count for under 50, 50+11shield = 31.05 chance to block, 42+11 shield = 29.65. So not very much difference at all

!

Defensive skills don't have dimishing returns post-50.

 

[Solid]

safe to say a thread full of misinformation?

 

[sorusi]

i have to say i see a big differance between ppl with 42 shield and 50 shield..

also there is no reason at all to take parry high imo, it doesnt work in rvr very good, but nice in pve tho...

 

[mirieth]

This thread is making my head hurt.

I was fairly certain I would go 42shield 48chants 39slash xxparry with my Paladin, but now I'm not sure.

50shield 34parry 48chants seems very tempting to me for two reasons... I'm going to hit like a wet fish anyway, so why waste more points in slash to do, say, 10 extra damage per swing? When I can do my primary job better: guarding and protecting my group and keeping them alive in general.

Help me.

 

[bracken_woodman]

Traditional defensive armsman spec is 50 slash, 42 shield, 39 parry, but I went for 50/45/36. Reason being that you get the same overall chance of preventing a blow hitting (give or take a little), but the balance more in favour of blocking - which is what you want in rvr (as the previous points made demonstrate) . Throw in the master of blocking/parry RAs to level 2 and seems to work well in rvr (given the known problems with block/parry that exist).

In pve against yellow mobs I block or parry around 75% of the time, often more but rarely less. Throw in some nice buffs and use of riposte and red con mobs become straight forward kills


Obviously doesnt help your pally decisions, but info all the same

 

[chretien]

I'm going 50 shield for exactly the reasons you state Mirieth. In my case as an Avalonian, my forte is in blocking anyway due to my higher Dex and my damage will be forever gimped because of my low strength. Therefore it makes sense to me to do what I do best rather than strike a poor average in everything.