Counting the outcome probabilities of single actions is elementary.
To be truly strong, One has to learn to count the probabilities for the different scenarios for his entire Turn so One can choose a likely triumphant scenario, or if need be, knowing the odds choose a more daring one.
Also, if One is able to position Ones players so that he can lead Ones Enemy to execute his Turn in a predictable fashion, One could calculate the odds of next Turns also...
And if One has a mind like Gari Kasparov's and One can predict his opponent exactly, One could count the probability of a drives outcome when the teams have been set up.
A little more seriously:
A word of advise: In a game, probability counting is easy when you do it like this:
The probabilities are rational numbers, id est: X/Y. The denominator (Y in the previous example) in Blood Bowl probabilities is usually some power of 6. When counting probabilities of several consecutive dice rolls, you multiply all the numerators together, and all the denominators together. This is easier to count than if you change them instantly to percentages and use them.
Example:
Roll 1: X1/Y1 Roll2: X2/Y2 Roll 3:X3/Y3
Result: Probability is (X1*X2*X3) / (Y1*Y2*Y3)
In BB this would usually mean (X1*X2*X3) / 6^3
What I'm trying to say is: learn the powers of six by heart! 6, 36, 216, 1296, 7776 etc.
Then you'll always know the denominator. Then you'll only have to count the numerator, and the probability is counted a lot faster.
What about rerolls?
If a certain roll can be rerolled, count the probability as above, but invert the probability of the particular roll (i.e. use the probability that it would fail) and multiply the result again with the non-inverted probability to get the possibility of success in the event that you have to use a reroll. Then
add this probability to the probability of success without using a reroll.
BB example: catch (3+), dodge(3+, Reroll), two sprints(2+).
Success without rerolls: 4*4*5*5 / 6^4 = 400/1296
Success with dodge reroll: 4*3*4*5*5 / 6^5 = 1200/7776
Add these together (remember that to add two rational numbers together they must have the same denominator, so multiply 400 by 6) to get
(2400 +1200)/7776 = 3600/7776
Even without doing the division, we see that the probability is slightly less than one half, division gives us the exact number of about 46.3% probability of success.
If you're not good at multiplying in your head, then you might want to divide such scenarios into parts, calculate the odds for the parts, round to a comfortable number and then calculate the total result. For example above we get
3+ 3+(RR) = 4*4/6^2 + 4*3*4/6^3 =144/216 = 2/3
2+ 2+ = 5*5/6^2 = 25/36 = about 7/10
So the total is 2*7 / 3*10 = 14/30 = 46.7%
Due to rounding, that's not an exact number, but it's close and perhaps easier to count. But if you use this, be aware of the fact that the more you round the more it effects the outcome.
What about if the counting gets complex: for example the reroll possibility of a later roll depends on whether you used a reroll earlier, you have multiple rerolls etc.?
Then (if you want to find out the probabilities) you can use what is called a probability tree:
You start drawing from the left. Every die roll is a node that spans two lines: success and failure. On the lines you mark the possibility of that happening (Line1prob+Line2prob = 1/1). Every line ends into an another die roll node or into an end node (that outcome leads to a TO for example). When you've drawn all the scenarios, you should have a tree starting from the left, dividing up into several branches with several die rolls in each. Note that the same roll usually appears in several branches.
Now you look at the endings of the branches and decide which of them are favourable. Then for each such branch, you take the route that leads up to it, and multiply all the probabilities on that route together.
When you've counted all the favourable routes, add them together to get the total probability for a favourable outcome.
Sounds complicated?
It isn't, but I'd need to draw it to make it appear simpler. The way we calculated reroll probabilities above is a simple example of using a probability tree.
And, as a student of mathematics at the university, it might be that even if I don't consider something complicated, it still might be
