Attack bug

Your personal Support ID is 715CCE7D

Did something change in attacking, as in is there now a random factor with SOM intel and attacks?

This is the second time this age this happened to me, the first, even though I could not find what I did wrong I wrote it off as user error, but this time I cannot find any error.

I got the target's intel via SOM. Intel said target has Net Defensive Points at Home 121,481. Target is not online, we are not at war or anything.
I sent a little over 122k, according to ingame calc, and bounced with a army hastily retreats message. I cannot find what I did wrong at all.

Is anyone else seeing any issues with attacking? Most attacks this age so far without any issues but the two attacks I have bounce are not making any sense to me.

Yes there is a random factor on attacks;

http://wiki.utopia-game.com/index.ph...#Random_Factor

sorry, there is a slightly more detailed explanation here;

http://wiki.utopia-game.com/index.ph...Offense_to_Win

Thank you for the link. I remember a few years ago a debate on whether the random factor was still there and the consensus was that it wasn't. I will concede that in the intervening years I may have just been lucky with sending a few points above the actual defense and not getting punished for it. It seems like luck may have just caught up with me twice this age.

I will adjust my attacks accordingly. Thanks again for the link.

Just a further note; Ambush doesn't have random factor attributable to it, so you can win those battles with +1 off point.

A few further notes:

I at least don't know the rounding algorithm on the %def calculated against the box, so it might be possible to bounce while showing 104.04%. Presumably super unlikely, but in theory if 104.035% is rounded up there is room to fail. Someone could go test this though to isolate the rounding behavior.

There are situations where intel can still have error in it. Into or out of war (or inter-kingdom while in war?) is the one I know of off hand.

The precise distribution curve is unclear, so I don't know how likely sending exactly the target's def is to work. We can make a reasonable guess that you'll have a 50/50 shot if you *undersend* by 3% (aka, showing 97% in the box, equal to the .97 in the attack equation), but even that isn't confirmed. (It'd just require weird asymmetric component random factors, which seems like way more work that anyone would have spent on that part of the code.) While it is rarely correct to go under the 1.00404 mark... there are some situations where it might be worth the risk - as you've shown, sending only 1-3% over usually still works.

It is possible to succeed an attack with only 90.44%. What good that info does anyone, I don't rightly know - but I find it funny to imagine.

sending exactly 100% of defense, will result in 50/50 chance of failure. Chance is 0 at 104.04 (not accounting for outside factors, like soldier bombs and last-second defensive modifyers)

That much is pretty straight forward.

Any attack under 95.96% is 100% fail.(not accounting for outside factors, like defense being killed or released)

To get accurate intel into or out of a war, it requires 35% of thieves be sent, as opposed to 10%, if you are grabbing intel through TWO war walls (IE, you are in war, grabbing intel on another kingdom in a separate war), the percentage is MUCH higher.