Say NO to nerfism against goblinoids!

[KGB]

Chazar,

I hope I know how the algorithm works. I am the one who suggested and designed it Later when Piranha was implementing it, I implemented it as well to help test his code and I sent a bunch of test cases including the 8vs8 example to make sure it all worked correctly.

What he states there is exactly what I said above. You can see results of <10% because an individual outcome can be less than 10% as long as it's in the middle 80% of the curve (if it's not, the battle is re-run as Piranha says). The original idea was not to create a guaranteed victory but rather to smooth out the standard deviation so that you didn't get extreme results like an 8v8 battle resulting in one side winning while losing 0 or 1 man. The side effect of implementing it created what I termed the 90% rule where once the sum of your chances to win exceeded 90% you couldn't lose.

So when you are observing battles what you want to look for is when you have an 8v8 type battle with identical strengths that the winner will always suffer at least 4 casualties. Otherwise the battle will be re-run until that happen.

KGB

[smursh]

The orc i n its current form is really just a glorified scout, specialized for swamp the same way the lt. calv is an open terrain scout. To look at him as anything else doesn't work at his current strength level.

People are using them with assasins since the new hero came out, but really what did you gain? In the end I find that even scouts work due to group assasination once you hit 30-40% group ambush.

Compare the orc with lt. calv instead of lt. inf and the high cost/upkeep of the lt. calv seems excessive. The main difference is the terrain bonus on open is usually more useful than swamp bonus-but it really depends on the map.

In the end most of the 10 pnt units are really just beefed up scouts. The value here is movement in swamp, not combat ability-not even swamp combat bonus.