View Single Post
  #5  
Old 11-16-2011, 09:33 AM
MattCaspermeyer MattCaspermeyer is offline
Approved Member
 
Join Date: Aug 2010
Posts: 553
Smile Rationale to changes mentioned

With respect to a lot of the changes I made, I used a formulaic approach to the value adjustments.

For example, if any spell was mass at level 3 and the only difference between its level 2 and level 3 values were mass, then I simply multiplied the mana cost by 5 (assuming it could be cast on 5 units simultaneously). This way, all the spells that I made mass (list coming shortly) are essentially automatically balanced if I get the mana cost for level 2 right.

I also used a formulaic approach for computing the mana cost for each spell level based on a formula that compared the increase in spell values from level 1 to level 2 and then applied the average increase to the mana cost. This makes the spell mana cost parameter driven which aids in balancing out the costs.

The only issue is then whether the level 1 values are consistent. This is the subjective part here and may be a part that needs some tweaking based on feedback, but I think I got the "feel" of most spell's level 1 mana cost about right and therefore the level 2 and 3 values "feel" right to me since they are all driven. I still have a favored spell's list so I'm biased by that and so could use feedback on spells that I don't use as often. Time will tell.

I essentially used this same formulaic approach for the skill tree where I compared the level 1 values to level 2 and 3. So for example if I had a skill that gave +5%, +15%, and +30% to a value, then I could simply decide what I wanted the runes to be for level 1, and then level 2 would use twice as many runes (15-5 = +10 - twice as much as +5) and then level 3 would be the sum of the rune cost for level 1 and level 2 (levels 1 and 2 brought you to +15 so spend the same amount for another +15). Once again if a skill had more than one value, then I averaged the increases to get an estimate. I used a master skill list rune cost and compared it to the number of runes you could get via level up to determine if the level 1 skill runes needed to change. Once again, somewhat subjective, but I had the total rune count to aid me with reducing / increasing the level 1 skill value and then the level 2 and 3 runes were automatically computed.

With respect to the additional values that some skils have, I used (I think it was) Razorflame's original post plus people's comments (along with my own opinion) to give weak skills more bonuses to make them better. Now I feel most skills are pretty useful especially with respect to those skills in your hero's class skill tree.

By the way, for those who played my Emperor of the Fading Suns Hyperion mod back in the day, I used this same formulaic approach to design units that were based on their resource inputs. This worked incredibly well for designing well-balanced units and I received a tremendous amount of positive feedback to how good the mod was because of this methodology.

So I approached this mod with the same mindset and I think it has worked incredibly well so far. Time will tell, and certainly I only have a sample of 1 right now and so will rely on feedback to aid with setting up the initial conditions, but I can tell you that I'm having more fun playing this instead of Armored Princess and Crossworlds, which I will hopefully get to modding after I'm done with this mod.

For spirit abilities, I used another set of formulas to aid in setting the power, rage, etc. values. Single-target abilities grow by a factor of 100%, while multi-target, and all-target attacks grow by a formula based on level (which are more percentage based on the level 1 damage than the current level damage). The idea here was that Single-target abilities would do the most damage since they targeted a single unit, then multi-target abilities would do the next most damage, and all-target abilities the least so that there was some strategy associated with choosing an ability.

I did a similar thing for spells so that single target spells would not be so useless and the burn and poison damage system made choosing those spells more interesting because Magic Pole Axe might do more damage this turn, but Flaming Arrow or Poison Skull might do more damage over 2-4 turns. Also Ghost Sword does less damage than Magic Pole Axe, but may be able to do more damage than Magic Pole Axe to units with good physical resistance.

You also probably get more bang per buck with Flaming Arrow than Fire Ball or Fire Rain, but it takes longer to do more damage to multiple targets with Flaming Arrow and so there's the trade-off there. Then with respect to Fire Ball and Fire Rain, Fire Ball does more damage to the central target but less damage to the peripheral targets, whereas Fire Rain does more average damage to all targets.

These and other tradeoffs went through the spell redesign process and how to make spells more usefull and add strategy to their use. The original spell list balance I thought was especially weak and I only thought a handful of spells were useful. Now I feel that there are fewer useless spells.

Anyway, I've actually placed a lot of thought into the game and I've played the Mage 3 times, Warrior once, and am just starting my 2nd Paladin game all with my mod over the past year since I started (I've found that it seems to take me about 2 months or so to finish a game, when I'm playing, analyzing, modding, adding new features, and fixing issues).

I basically learned how to do virtually everything myself with the help of some of the posts in this forum. Unfortunately, since I got into this game about 2 years after it was released, I wasn't able to get really any help from any of the modders and developers of the game. This was especially true when it came to editing the *.CHAT files with their hexadecimal codes that I couldn't figure out how to generate. I PM'd quite a few people and essentially either no one responded or they responded with information that I already knew. So I basically had to just try it out and figure it out as I went.

The changes to River Fairy Diana and the new addition of Orcelyn Ordy seem to work okay, but I have no idea if there are any side effects from what I did since I couldn't get any help. Also, unfortunately, the Crossworlds Editor cannot import The Legend *.CHAT files - it crashes to desktop when I do it. So I had to look at the other wives as examples and have the new wife *.CHAT files use a lot of the same code.

More play testing will tell whether I got this right or not, but I was able to finish my last game with the River Fairy Diana changes. Orcely Ordy is completely new for this go around and I need to work on her dialog to make it more interesting, but you can at least talk to her and have babies!

By the way I'm slowly working on the Change List and have completed the *.ATOM change list and will post it shortly after this post.

Feel free to ask any more questions!

Quote:
Originally Posted by Helios View Post
Wow, fantastic work, Matt! These are some great ideas to spice things up, and hopefully increase the challenge. I haven't played HOMM3, so I'm unfamiliar with that portion, but I can't say I disagree with any of the changes I could see in the YouTube video. The wives, childrens, and Phoenix summons were in desperate need of revision.

I'm definitely going give your mod a try--I'm currently in the middle of my self-modded all Dwarf game--and will let you know when I get to it. This sounds like it will add a lot of replay to KBtL.

(Unfortunately, I can't read the in-game screens on the YouTube video due to my monitor's limitations, so I can't comment on them yet, but will check it out again on a different monitor.)

EDIT: OK, I've seen the trailer on a different monitor now. It sounds very interesting, and will definitely make the game a serious challenge with enemy heroes regaining mana, for instance. How did you come up with the numbers for the changes you made, for example: the huge increase in mana cost for the Helplessness spell? And have you done much checking for game balance, yet?
Reply With Quote