Level Testing discoveries and possible bug

[RaustBlackDragon]

I was using some hacking to work out how much of a difference leveling makes on its own, and I think I've discovered a bug, albeit one I don't fully understand:

 

I "Modded" the game to turn cloak of curses into a self-XP boosting spell so I could examine the effects of level on performance when stats were artificially made equal for characters of different levels. My initial tests against the tutorial cave's scrawny nephilim indicated that each level adds a base of 2.5% (rounded up) to your chance to hit and chance to dodge. I decided to test this out with a branched save of my current file to see if there was a difference at higher levels:

 

I used the editor to buff up my ninja's dexterity from 26 + 7 to 33 + 7, since the royal guards in the royal spire had an obscenely high hit rate on them, which I found peculiar. However, to my utter shock, even with that much, the first royal guard I found in the royal spire still had an 86% hit chance on me. I couldn't find anywhere in the code that indicated stat distribution, so either these guards had extremely high strength and I couldn't see it, or their level was ridiculously inflated which would also explain why my characters had difficulty hitting them. I decided to test this on the next guard, and I was in for a bigger shock: The next royal guard to attack me only had a 6% chance of hitting my ninja.

 

That's a difference of 80%, which is equivalent to 16 strength, or 32 levels.

 

This kind of crazily inconsistent variance also showed up before during the battle with the demons in the black spire, which leads me to believe that there's something in the code that allows for much too great a range of possible levels enemies can be. The chance to hit is just all over the place, and I suspect it's a normally reasonable variable that torment has accidentally caused to be exaggerated. I'm going to see if the difference is as wide on lower difficulty levels, and I'll post back with my findings!

[RaustBlackDragon]

Nope, this 80% hit rate gap is there on normal as well. I'm at a loss for what's going on, this sort of thing seems to start happening at the later parts of the game...

[kensuguro]

I wouldn't be surprised if certain npcs were just hard coded that way.. is this just a particular guard in the royal spire? (you're talking about hawthorne's place right?)

 

Also, could it be that the particular guard has modifier items that starts his hit rate way above 100%? Or that torment level starts enemy's hit probability above 100%? But I guess it's still strange that there's such a huge variance.

 

How are you measuring hit rate btw? Are you tracing out results from a hit rate function through a debugger or something?

[RaustBlackDragon]

What do you mean? The message box tells you what the hit rate is whenever somebody attacks. Or is it only that way in the mac version?

 

Anyway, yeah, the main thing that confused me was the fact that they were identical regardless, they both had the same name and appearance, which suggests to me that they were spawned with a level buff or set_attack_bonus.

 

And yeah, I think the hit rate difference is rather exaggerated between difficulty levels, though that's not the cause of this. I'm highly tempted to make a Noble Hack intended to be played on normal, but providing the same challenge as the original torment (just with more reasonable and standardized hit rates).