battle bug found

[SnotlinG]

It has come to our attention that we have a bug (or bad logic) in our battle function.
Whenever a battle is calculated, we run 1000 simulations, to estimate the result.
This is then used to build a range of accepted results from the battle, where we remove the most extreme results, normally the top and bottom 10%.
I.e. if you lose a battle that you had more than 90% chance to win, the battle is re-rolled, since you are not supposed to be able to lose when having those odds.

This re-roll is actually affecting the true probability of the battle outcome.
For example if you are in a battle with 80% chance to win, you should lose 20% of the time.
But due to the re-roll of extreme results, you in reality have 7/8 in chance = 87,5%.
More info can be found in this thread: http://www.warbarons.com/forum/viewtopic.php?f=4&t=2514

Our solution will be to change the code, so that instead of re-rolling the entire battle, the battle will be altered to the closest accepted outcome. I.e. if during the battle, you have reached the minimum number of units left that gives an acceptable outcome, that unit will win its remaining battles.
Same thing goes for if you are about to win a battle, but with too many survivors left (i.e. your battle outcome is too good), the last enemy unit will make sure to kill at least as many units from you to make sure you go down to the acceptable limit)

These changes will take some days to implement, and we will announce it here when it is implemented since it affects all battles in version 1.0 and better.
As part of these changes, we will add a max limit on ambush, to be 95% (instead of the currently no-limit)

[gil]

You should rollback that last battle update
I have more battle reports with extreme results then 50%

[SnotlinG]

You should rollback that last battle update
I have more battle reports with extreme results then 50%

Update has not yet been applied

[SnotlinG]

After some consideration we have decided that we will not include the battle-fix in ongoing games, but it will come in the next release (1.01), which is currently being worked on (no ETA at the moment).

[KGB]

I would think it would be better to introduce it sooner rather than later. Especially since a Ladder season just ended.

The fix from this bug is going to result in more outright wins for the weaker side. That may take some getting used to by players which is why I think it's better to have it sooner than later.

KGB

[Chazar]

Hotfixes should really be avoided!

We have lived with the bug for a long time just fine, and silently updating ongoing games is pretty irritating. That's what versions are for, with all changes being properly announced.

Actually, I think that each game should display its version number somewhere, maybe in the game objective info. Yes, I do know that I can just read a games version number from my browser's URL-bar, but I am sure not every player knows how to spot it.

[KGB]

The problem is that now we are only getting updates at a rate of about once a year instead of 3-4 times a year as we were during early development.

Hotfixes are a way of life. I get updates from Microsoft all the time for my OS for critical bugs/issues. The same for virtually every other type of software as well (I get game patches all the time). Most websites just deploy fixes without even telling anyone since they auto-magcially update the next time you visit the site.

Warbarons already has the capability of having games using different versions (I have a few older games going that still use older rule sets). So I see no reason why we shouldn't get regular updates especially when there is no ETA on any future release.

KGB