Here, this is closer to reality :)

1. Double Strike is no different from Retaliation... In both cases, a coefficient of x2 will work...

Allow me to disagree! :)
Double Strike is no different from Retaliation only in the case when... well, you know when.
In a normal case, after the first strike, the stack size decreases, and consequently, the strength of the next strike decreases (we are considering not a single unit, but a stack, otherwise it wouldn't make sense). As follows from the general trend, the average damage fluctuates within 1/2-1/5 of the unit's HP, i.e., after the retaliatory strike, the stack size will decrease by the average value :) 1/3.5, i.e., by about 30%, so the second strike will be performed with a strength of 0.6-0.7 of the initial value.
Considering that in the case when... this coefficient is slightly increased, we get approximately 0.75. :)

2. Dragon Strike should be lower than Multi-Strike and "Bite" because it hits a maximum of two units, while the others hit at least three. Multi-Strike can be compared to archers who attack with area-of-effect spells (liches and magogs).

Yes, yes. I was thinking about this, but I didn't voice it. There are many nuances. Moreover, "shooting" + "area of effect" is already a double coefficient.

3. Using expected value is the most accurate way to calculate coefficients, but the essence of the question is not so fundamental as to use heavy artillery. The reason is that we still won't take into account all the influencing factors, and the error will be much higher than the accuracy of the tool.

- Colleague!
- Yes, colleague!
- Allow me to shake your hand!
- Please, colleague! :)
Mathematical expectation can easily replace the personal observation of a small group (even if it's 3-5 people). But "thoughtful" people. :)