[Fixed for V0.38] No mana characters

Des · Post by **Des** » Sun Apr 30, 2006 11:13 pm

Whilst testing the RTC version of Angel's Egg, Arnie the Terminator gained a priest and a wizard level despite having no mana as I got him to do the odd spell from time to time by holding a mana-boosting item. In CSBWin and the orignal game, this would normally gain the character 3-4 mana points per level, but under RTC, only a single mana point was awarded even if I cheated and re-ran from a save just before levelling up.

I was wondering if this is "by design"...?

George Gilbert · Post by **George Gilbert** » Sun Apr 30, 2006 11:47 pm

No, I think you were just very unlucky. If you try a few more times I would expect you'll get a larger boost.

I think RTC has a larger random factor than DM did (although the mean value is the same).

Also bear in mind that RTC stores all its stats as sub-integer values, so it's quite possible you've actually gained 1.9 points of mana, so the next wizard level you gain will give you more than you expect due to rounding...

Des · Post by **Des** » Mon May 01, 2006 1:26 pm

Hmmm, well I tried 10 times in total and got 1 mana every time. Could it be because he has very low wisdom (26)?

beowuuf · Post by **beowuuf** » Mon May 01, 2006 1:28 pm

That's low wisdom...Dm characters started with about 30 odd as the low end. At the very least if the RTC model follows the DM one the regen rate would be terrible

copperman · Post by **copperman** » Mon May 01, 2006 2:10 pm

In RTC stats (S,D,W,V) appear to have little, if any effect on stat (H,S,M) progession at lvl up. Some clarification of min and max expected values would be nice, help me finish my character generator spreadsheet. I'm trying to simulate accurate characters for new games.

After using "special" weapons in a modded DM-O dungeon I noted that HSM progress was rather flat and after rasing all 24 original characters the highest mana wasa mere 119 from Chani
Final stats

Code: Select all

H 354
S 388
M 119
Str 46
Dex 60
Wis 69
Vit 51
AM 54
AF 50
Leveled to Artisan F, Expert N, Artisan W {2}Master P

Maybe it's just me, but this mana seems low to me, SDWV fit and these numbers I can accurately simulate. Any help would be groovey.

Yeay for reverse enginneering.

George Gilbert · Post by **George Gilbert** » Mon May 01, 2006 2:15 pm

Des wrote:Hmmm, well I tried 10 times in total and got 1 mana every time. Could it be because he has very low wisdom (26)?

No - wisdom just affects recovery rates not mana gain from leveling up.

I've now had a look at my code though and what's happening is that for the first wizard level it gives you between 1.0 and 2.0 mana. This means that the displayed integer value will nearly always be 1 (because it's quite unlikely that you'll get 2.0 exactly).

Of course, the next level you gain, you'll probably get more than you expect - for example, if this time you gained 1.9 mana, next time you'd only need a gain of 1.1 to appear to get a gain of 2 (i.e. going from 1 to 3).

I don't know what DM did, but let's say it either gave you 1 or 2, then on average throughout your game you'd get the same gain as DM, but for statistically small samples (e.g. just one level) then it will be different.

Des · Post by **Des** » Mon May 01, 2006 4:10 pm

Thanks. It looks like RTC is just stingier than DM on the early levels. I just trained up Arnie in CSBWin and he got 3 mana points for going up his first priest level, then a generous 8 for his first wizard level.

Copperman's stats calculator sounds like it will be a useful tool for designers. For my characters I used PaulH's method here http://www.dungeon-master.com/forum/vie ... sc&start=6 as a starting point, then (mostly) upped the stats a bit.

copperman · Post by **copperman** » Tue May 02, 2006 4:10 pm

Tx Des I'll check that out.

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 6:42 pm

Des wrote:I just trained up Arnie in CSBWin and he got 3 mana points for going up his first priest level, then a generous 8 for his first wizard level.

Blimey - 8 is alot.

I've just tried doing it several times over and the most I've been able to get for the first priest or wizard level is 4 (which clearly shows that there's a bug in RTC, which I'll fix for V0.38).

Are you sure 8 was for the first wizard level - or perhaps the mana raise was greater because you've already got some from the priest level?

Des · Post by **Des** » Tue May 02, 2006 7:14 pm

The 8 was for getting a wizard level with a priest level already in the bag, so as you say the engine must be taking into account the existing level and awarding a bigger range.

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 9:05 pm

OK, I've done a load more experimenting and I've managed to get a good statistical distribution for level gains and put these into RTC. Overall it looks like the low level gains didn't give you enough and high levels were about right!

Interestingly it seems that the problem hit all the stats, not just mana, so there's tweaks across the board.

Just for the record, I managed to get a gain of 8 for the first wizard level with no other levels gained and the additional stats (str, dex etc) made no difference whatsoever.

Des · Post by **Des** » Tue May 02, 2006 9:59 pm

Good work. I guess as it only affects the early levels it shouldn't have too drastic an effect on the relative power of resurrected and reincarnated characters. Reading this thread http://www.dungeon-master.com/forum/vie ... hp?t=25374 it may well redress TheMoreMegil's complaint that his reincarnated Surgical Strike characters ended up worse off.

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 10:14 pm

Yes, to grossly over-simplify it I was using a model of:

Increment = between "new level number times 2" and "new level number times 3" points

whereas

Increment = between "(new level number plus one) times 2" and "(new level number plus one) times 3" points

fits much better and therefore what RTC now uses. For high level numbers (say 10-15) the extra "plus one" makes ~10% difference to the level gain and so isn't noticable. For the first level though it doubles the increment!

Post by **Lunever** » Tue May 02, 2006 10:31 pm

Hmm, does that mean that characters taken from mirrors in V0.38 will have a better stat increase on levelling up than V0.37 or earlier characters?

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 10:39 pm

No - the stat increase is part of the engine, not part of the character (or the dungeon)

Post by **Lunever** » Tue May 02, 2006 10:44 pm

So it might happen that if I import a veteran character from V0.37 to V0.38 that some stats might be changed?

How can a stat increase that has happened to a character not be part of the character?

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 11:09 pm

Erm - you've completely confused me now.

The stat increases are applied by the engine at the time the character gains a level. The act of increasing (specifically how much the increase was) is never recorded, just the new value.

If you import a character its stats will be unchanged because of this fix. If you gain a level with V0.38 then the increase you get may be different from if you had used the same character in V0.37 though.

Post by **Lunever** » Tue May 02, 2006 11:16 pm

Right, what I thought it would be. So if I play 100 reincarnated Daroous in V 0.37 from nothing to Lo-master fighter and then do the same in V0.38, the 100 Daroous from V0.38 would beat the 100 Daroous V 0.37, right?

George Gilbert · Post by **George Gilbert** » Tue May 02, 2006 11:31 pm

Post by **Lunever** » Wed May 03, 2006 12:06 am

Can you then please include a simple functionality in V0.38's import feature that will roughly apply this increase in average stat increment to characters imported from older versions?

George Gilbert · Post by **George Gilbert** » Wed May 03, 2006 12:09 am

Not a chance!

Basically, it's impossible to work out how many levels a character has gained (and therefore what increment to apply). For example, if all you've got is that they're an = master wizard, how do you know if they were taken from a mirror at that level, gained 5 levels, or even gained all of them...

Post by **Lunever** » Wed May 03, 2006 12:14 am

That's why I said roughly. I'm well aware that it is not possible to accurately recalculate this, since the information necessary for that is not available. Just make a very crude approximation, a small even-out to annihilate character developement differences between versions.

George Gilbert · Post by **George Gilbert** » Wed May 03, 2006 10:38 am

It's not even possible to make a crude approximation.

For example, it's possible that the character has come straight out of a mirror with the stats they currently have (i.e. no adjustment is necessary). Equally, it's possible that the character was re-incarnated and so gained up to 15 levels through levelling up (i.e. an adjustment of up to 6 points per level, so a total of up to 100).

So, there's a range of 0-100 points to pick from...how do you approximate that!

On a completely orthogonal note, your characters (or more importantly, you!) are already far too good at playing DM - now that the difficulty has been reduced you of all people don't need any more help!

Post by **Lunever** » Wed May 03, 2006 12:17 pm

Ok, now I understand why it can't be approximated. That is, an approximation would only manually be possible by taking into acount the original mirror characters (which are known), and the player-provided information whether characters are ressurrected or reincarnated (reincarnated in this example). Is there a way I can do this myself when V0.38 is out without violating possible algorithm you might have included to prevent cheat-characters from entering th gHoF?

Edit PS: You are right about too good a party for DM: But as you can guess I'm already thinking about archmastering CSB...

beowuuf · Post by **beowuuf** » Wed May 03, 2006 12:33 pm

Actually, all the importing research would have been far more useful if applied to playing CSB from the start rather than beating DM on settings no one else would have tried

Hopefully the algorithmn holds and are good for various imports onto the CSB win dungeon, rather than being broken and needing tweaked again : (

Post by **Lunever** » Wed May 03, 2006 1:11 pm

I will test that too, but it might be a couple of days until I'll have time to play through.

Post by **Lunever** » Wed May 03, 2006 2:57 pm

Well, I did do a lot of testing of importing into CSB too before V0.37. Yet since of course you are right about DM->CSB port being more important thatn DM->DM port, I will concentrate the test runs for V0.38 on CSB.

Besides, playing and fully completing the same dungeon over and over again is a good way of really finding all bugs that are still in the dungeon file as well as the bugs that appear due to format changes in a newer release, and thus be able to remove them.