Weaknesses of the presented metrics

The three presented survivability metrics are very fitting for several amounts of situations, however not every situation is described by the metrics in an adequate manner. Let’s review them shortly:

• MEHP describes the minimal survivability against single attacks from bosses well

• AEHP-Boss describes often hitting boss fights by using the statistical expected result over an infinite amount of hits.

• AEHP describes often hitting enemies by using the statistical expected result over an infinite amount of hits.

Now, two questions are catching the attention:

1. What is with ‘Deflect’ and the probability of surviving hits outside the minimal survivability that are ignored in the MEHP metric? May it be better to invest into such cases too instead of maximizing the MEHP?

2. What is with fights where the enemy hits more than one time but not too often either? Then the infinite amount of hits does not represent survivability very well! A typical approach would then be to use border numbers of survivability for different scenarios or by using a more fine grained statistical approach.

Now, let’s shortly shred more light into the questions.

To question 1)

As we already know, Deflect can not be increased to 100% and thus is not fully reliable. However, it can be increased to 90%, which creates a fair amount of reliability. At least when a certain dangerous situation does not happen too often! Abstractly said, one can depict three areas of survivability instead of the mentioned 2 in the MEHP metric:

• The area where one survives in any case

• The area where one has a chance to survive due to a deflected hit

• The area where one dies no matter what.

Let’s draw an example to better understand what this means: Let’s assume that the minimal survivability described by MEHP is calculated at 1 000 000. The Deflect value is set to 50% and the Deflection Severity to 50% too. The green area in the image depicted below is the area where one ever survives. The yellow area is the area where one may survive in 50% of all cases because the hit was deflected. And the red area describes where one gets killed in any case.

If we decrease the amount of stats that affect the MEHP formula and increase instead in Deflect and Deflect Severity, the bars are moving! For example if we reduce the value that the MEHP gives by 10% and add 20% in Deflect then the graphic looks as follows:

Or when we reduce the MEHP by 20% and add 20% Deflect and 40% Deflect Severity:

Now it is hard to say which of those is better! Ignoring the fact that stat values were just added in ‘made up’ amounts, the answer is: It depends on the situation! For example, if we know for sure that the boss hits between 900 000 and 1 000 000, then it is probably more wise to cover it by using the stats to maximize the MEHP formula. But when we know that the boss hits over 1 000 000 only a few times in the fight and cannot be mitigated in another way, then it is a gamble to survive those hits anyway. In this case it is probably (and only probably because of the other situations of lesser damaging hits) more wise to increase Deflect and decrease the green area. If we overdo it however - like in the third case, then it is a sure loss. Worth to note is, that in all other cases - where the boss hits are all well below 800 000 or well over 1 333 333 it does not matter, because the outcome is the same!

Thus the MEHP formula is content dependent and there are situations where Deflect on cost of other defensive stats is well worth investing into!

---

To Question 2 (more hits than one but way less than infinity):

These situation can be depicted in two ways:

1. The simple way would be to just add the few hits into one value and treat it like one hit by using the MEHP formula. In most cases this shall be sufficient, because the hits are in the same range of damage as a tank buster and thus it can be safely built for the case that no hits are deflected.

2. If, however, the damage added together is way higher than a tank buster (for example because 10 seconds or more the boss rages and attacks significantly more often) then a more fine grained statistical approach is needed. Neither the MEHP nor the AEHP formula would suffice. For a more fine approach, more variables are required. This may include

   1. healing per second

   2. difference of damage of the hits

   3. attack speed of the boss

   4. attack damages

   5. and some more content dependent variables

Thus, one can only use a fine grained approach when the environment is well defined. This goes out of scope of this guide, because there is no hard enough content currently available that justifies these approaches (and maybe shouldn't be either). At least it saves a ton of time by not diving into this!

Summarized, the error in the results for a few hits are ignored and instead either the MEHP or the AEHP metric is used.