Tilting at Windmills

Debugging the combat engine continues… it is really good to do so, though, since now it’s actually working and the bugs aren’t completely derailing it. Well, most of the time.

A simple register shift being one off caused all the problems with other characters not being able to hit without crashing the game. I was also adding health instead of wounds when striking monsters, meaning they actually got STRONGER with every hit… now monsters are dying correctly.

Currently, I have a few more bugs I want to fix up before I try and add any new FX, like ranged weapons:

• The wrong monster graphics are periodically loading, and I can’t figure out why. It seems consistent for certain monster types, so I will statically fix it to one and debug into it to find out what’s going on
• You can move about the entire map now easily enough, but moving OFF the edge doesn’t end the character’s turn, and actually do some funky stuff with the interface
• You can’t actually end combat, there’s nothing tracking that all monsters are dead at present. If I get this wired up, then I need to test and debug the combat ending routines, including treasure and updating travel mode so the monster you fought is gone
• Syncing up FX and their timing is needed. The regular hit graphic disappears long before control returns to the character, making it seem a little lagged. Part of this is also having a set amount of time always pass at minimum (Say, 1/2 second) so that you can see the combat update of a hit, miss, or critical hit, and also take longer if a FX lingers longer than that

The code and spreadsheets are a’changing… so far, I’m really looking forward to how the changes work. With simpler math, the combat engine should be much easier to balance out.

As usual, changing some spells was in order. I had several that affected defensive bonuses that no longer exist, so I removed them. A big challenge is with damage; while hit and defense numbers are fairly static and a +1 is ALWAYS good, damage will scale up a bit more. So I decided to have all damage effects be things like 1/2 or double. This means changing the bonus calculations slightly to account for changes of this nature instead of adding/subtracting.

One major addition is that I’ve added party rations as a game factor. I don’t intend for food to be a major pain like it is in Ultima, where running out of it effectively kills your party in short order. Instead, rations serve the same purpose as they do in Tunnels of Doom; you periodically consume rations on the world map and heal fatigue and possibly wounds. Rations also serve as a money sink.

I’ve removed some of the more lame spells and added new ones to replace them. A corrosion spell now exists that halves normal damage resistance, while protection doubles it. Like many spells that have counter-effects, one will remove the other if cast upon a target. A spell to create rations also seemed like a good idea.

One bit of caution I’m taking is to not make ranged weapons too overwhelmingly powerful. The first step to this is to make sure hit bonuses for ranged weapons are lower than melee weapons. Another is that spells that increase damage potential will only work on melee. That way, melee attacks still have good value at the expense of range.

My final step, before I try to start compiling and running the changes and looking for regressions, is to update the combat movement system. Before, you had to press the M key and then indicate a direction. I want to scrap this and have you be able to use the arrow keys immediately to move and attack. This will speed up combat considerably.

The refactoring continues… part of which is the redesign of how weapons and armor work.

Playing World of Warcraft in the last few years has really made me recognize the serious problem of high-scaling. Basically, if you create a system in which bonuses continue to rise at a steady rate, eventually you make a system where more experienced characters just wipe the floor with lower-level encounters. In WoW, since the Cataclysm expansions, the attribute bonuses on items got so ludicrously high that a maxed-out character could solo heroic dungeons from the prior expansion with ease.

Dungeons & Dragons suffered a similar problem, although not quite as extreme, with the 3rd edition. Although it was better than most at higher levels, the 15-20 level range had characters overloaded with magical items that boosted their skill and attacks to extremely high bonuses.. so much so that the 1-20 random range became a non-factor. The epic level rules (which clearly had never been play-tested) made things even worse.

I’ve addressed some of this already in my designs by dropping levels as an abstract measure of power, and experience gained from monster slaying. However, I still had items featuring steadily increasing power levels, to the point that an item could make a significant difference in power level, plus a monster with high statistics could easily overpower a low-level party.

On my re-design, I took it one step further by redesigning weapons and armor to have static values for attack/defense bonus based on type. So a sword always has +4 to hit, and a cost of 2 action points to use. A suit of mail armor has a defensive bonus of +3, and a move cost of 3 action points. And so on. I also removed all the other defensive options such as dodge, parry, and block. (Too bad, since I already had some sound effects for them.)

So how are weapons different? Well, their potential damage (measured as a base amount and a range) is different for each weapon, as is the regular and magical damage reduction of armor. In addition, the last two bytes of the weapon/armor array serve as a generic index modifier. Basically, an item can adjust another statistic in the array. A rapier, for example, may give you a +1 defensive bonus.

The nice thing with this change is that it stabilizes your ability to hit to a static amount. So a high defensive score remains high, no matter how powerful the characters are. Damage will vary, but I plan on keeping that curve very slow. This means a higher strength encounter won’t automatically wipe the party out, although they may get reduced to doing only 1 point of damage per hit.

I didn’t remove QUITE as much code as I thought I would, removing the pre-processed action system. The refactoring of the attack routine did reveal quite a few small little bugs, though. Some of them were straight values used as addresses, which was probably what caused the increasing instability in my combat engine. So hopefully this will help debug and get the FX engine rolling.

Okay, I got my new system figured out in my mind… actually re-configuring the code base to use it is a long arduous process I probably won’t start for awhile, but at least I know what I’m going to do.

All the pre-processing of actions is going away. When you move your characters in combat, you take all your actions then and there. Just like Tunnels of Doom. But how to track things like haste and speed? How does fatigue figure in? I was wracking my brains on a solution when it came to me unexpectedly.

Each character has “action points” which renew every turn. These points are used to take actions, such as move, attack, fire a bow, cast a spell, use an item, etc. Once all your action points are gone, your turn automatically ends.

So, here’s how it works: If you take an action that takes you to negative action points, you take that difference in fatigue damage. Fatigue serves as the first kind of “damage” you take, which recovers over time outside of combat, or through use of potions. If you are at maximum fatigue, any additional action damage becomes wounds instead. Maxed out wounds renders your character unable to do anything. Guarding takes any remaining action points as a defensive bonus and ends your turn.

Example: Eric Seablade has 5 action points. He’s wearing mail armor, which makes movement cost 3 points. He has a sword, which he can swing for 2 action points. So he can move once and attack for 5 points, ending his turn. Alternatively, he could move twice for 6 action points but he would also take 1 fatigue. If he moved once and then just guarded himself, he would end his turn with a +2 bonus to defense.

When I had this epiphany, it just worked across the board for everything. Weapon attacks, heavy armor slowing you down, expensive spell casting? All covered! There’s a slight bit of confusion with spells that can be cast in both combat and travel mode… the high cost of them should technically be less outside of combat. I may have to subtract standard action points from spell costs to reflect this difference.

The nice thing is this will greatly simplify the code as well. It may reduce the combat code base by as much of a kilobyte, which will be much more efficient. Processing actions immediately instead of storing them up for later solves other issues as well, such as tracking item types used.

Unfortunately, the retcon of the code will be murderously difficult, as many things in many places have to change. On the plus side, it would serve as a quasi review of the code for me to refresh my memory about what’s all in there.