Vapourware Ultimate Clusterfuck RPG

[hrose]

This is vaporware.

But also a concrete project I tinkered with in spare time for the past few years. I'm the kind of guy who is unable to think and plan realistically, so instead of making a game that is within reach I instead plan for the most epic, insane RPG that can be conceived. Also, I cannot really code, even making a simple menu takes me an insane amount of hours, so it would take me many, many lives to even get close to the plan I have.

Why doing this? Because it's fun. I enjoy experimenting with rulesets and tinkering with game systems more than I enjoy playing a good game. And maybe 50 years in the future there would be someone like minded, but with far better technical skills, who could pick up my plans and actually deliver a game that is fully old-school yet progressive, the same as I plan mine.

This project is built on concrete ideas and goals and that's what I'm going to write about. It's vaporware because it cannot realistically be completed, but the groundwork is solid.

The very best, but obviously unlikely, scenario is that someone actually is interested and helps me with code actively. After all this project exists because I had people helping me with suggestions as I moved along, but everything would change if there was someone who can actually write code and implement stuff without spending 20 hours on a 4-options menu like I do.

General summary and plan:

A roguelike-like game, mostly ASCII, coded in C++ and libtcod (but because I don't really know C++ it's mostly standard C spaghetti code written at a very, very low skill level). Everything I make is open source.

The plan is to have character creation for one character and eventually a party up to 4. So three more companions that can be recruited in-game, and then managed directly after they join. Rogue-like ASCII top-down perspective. When out of combat it works like a "blobber", so you just move around your party. When combat starts you "deploy" the party on the same top-down map, and combat becomes turn-based and tactical, with a sort of action points system. When combat is over you're back to blobber mode.

The rulest would be built as very old-school pen and paper RPG with explicit rules and dice rolls. But the aim is to make a mechanically very, very complex and intricate system. Not only I implement and adapt well known complex systems, but I blend them to take every noteworthy mechanic. The systems I already worked on use a blend of these rulesets: Harnmaster, Harnmaster Gold, Dangerous Journey, The Riddle of Steel, Rolemaster, Elric!, AD&D 2nd, Pathfinder. I take a mechanic, analyze how it's used in each of these different P&P systems, and then blend all into one trying to retain the interesting bits of each.

Because the foundation is Harnmaster, the system has no levels or hardcoded classes, and it would be entirely skill based.

The features that are not roguelike are: it's planned to have a fully handcrafted world and no permdeath. Instead of trial & error & experimentation of classic roguelike gameplay, it encourages to play and experiment on the same character and dig deep in the story/lore. It's open-world.

Theme: it's an almost-classic RPG with a "gritty", Dark Soul-ian theme. Because of this, the world is "artificial". It's not a realistic reconstruction of a bucolic medieval world. There's not a castle with farmers working nearby. It's a deliberate game space. This is part of the concept and the story (I want to use Shin Megami Tensei as inspiration here). The plan would be to have very dense philosophical and metaphysical themes, and I plan to outdo Torment at its own game (I know what I'm doing)

As I have time I'll explain more details about how the combat system works (including the plan to use abilities layered on top of a standard skill based system), and what I plan as the structure for the first dungeon, the general feel of how dungeon exploration is supposed to work, style of dialogue, and also "research" to be done about blending in systems from other outlandish games, including Vagrant Story (breaking down and rebuilding equipment) and Umineko (?) (specifically: color-coded truths).

I think reading about these outlandish ideas can be interesting, even if they aren't going to be implemented any time soon, or ever. I only wish I could code at an acceptable speed and get more things actually done...

 

[hrose]

Combat System

As I said above when out of combat the game works like a blobber, so you move an individual unit that represents the whole party across the dungeon. Whenever a monsters appears in line of sight combat is initiated, and you would deploy your party of 4 on a small area. But because I want flexibility, I also plan to have a button so that the player can trigger the combat mode even without an enemy on sight. This means you can manually deploy your party and move around the single units. It can be used to solve puzzle-like situation that rely on accurate positions, or, for example, to create an ambush. You know the enemy is patrolling, so you might use a party member as a bait while another stays close to a switch that would trap the enemy in a room and then fill it with lava. It's just a simple and flexible system, and it can also be used to plan to surprise an enemy.

Initiative

When actual combat starts, initiative is rolled. I've looked at the various systems. As a basis I use Harnmaster, and if I remember correctly it uses an "initiative" skill with no dice roll. So the order is simply set by that skill number. In Stormbringer you use the Dexterity value, and it's still fixed. This mechanic is not satisfying, because it's static and because it doesn't include important elements like weapon type or encumbrance. Other systems, for example optional rules for AD&D use a die roll and then add weapon speed. Rolemaster uses a stat + dice roll + bunch of situational mods.

I liked the idea of Harnmaster of a skill that can be improved, I also want an amount of randomness, every turn (because while P&P it's a pain in the ass to roll millions of dice, in a computer game you can retain this kind of complexity), but I also want to consider weapon type and encumbrance. I think the AD&D Combat & Tactics expansion offered an interesting and complex system. Instead of using a weapon speed value to add to a roll, it uses a more complex "phase" system. This phase system adds complexity because it breaks movement and attack so that you could move in an early phase, and then attack in another. It creates a very complex dynamic that factors all the elements I wanted. (as a standard you can perform an half move and one attack, or two attacks, or a full move)

So I built (and coded) an hybrid system. Unit type and weapon class would determine what phase you play in. Then, WITHIN the same phase, units are ordered through an initiative roll. All skills are percent based, but I wanted some randomness. I'm still undecided about the final formula, but right now I use: skill value + 1d20 -10. That means that if you have the initiative skill at 80, then the outcome ranges from 70 to 90. It includes randomness every turn, but it still overall hovers around the skill level.

I'm not sure if I coded also this next part, but I'll also add the option for critical rolls. So that if you roll a 1 or a 20 then you can JUMP to an earlier or late phase (by doing an actual skill check roll).

Phase system

So we have a phase system. If your character has fast stats or a fast weapon then you might act at an earlier phase than the monsters, regardless the initiative roll. But because it's a granular system you cannot use all your action points at once, only 4 for each phase (of a standard total 8, that are modded by a speed stat). So, you can use now, before the enemy, your 4 APs. If you are right next to an enemy you can attack (usually uses the 4 APs), instead if you use a melee weapon you'll have to move first. Let's say you move first, so you use the current phase to move 4 steps (the 4 APs) closer to the enemy.

By having spent your 4 APs you conclude the current phase (because no other units acts in that phase). So the next phase starts. But because the next phases also includes 4 monsters, this time the order is established by initiative rolls. But let's say the player had a better initiative skill, so he acts first again. Since he's now next to an enemy he can use the 4 APs he has left to make an attack. And if he does he'll be done for the rest of the turn (because he had 8 APs total, spent the first 4 to move in the previous phase and the other 4 to attack in the current).

This also means that when your turn comes up you could still decide to "pass" in order to see what happens and act later (or to have better chance at defending since you didn't make any attack, since making an attack decreases the parry skill check). The system should normally have just 5 phases, but the hardcap I coded is 10, to factor the possibility of characters passing their turn. In general a normal character uses its 8 APs in two phases, and characters with extraordinary stats might grow it to almost double that, but it's not something that would normally happen.

This is enough for now, it's already a good example to show you not just how this systems works, but the logic I followed when building it. I take a system, analyze how it's implemented in each P&P ruleset, then hybridize to retain all the interesting aspects of each. So in this case we have a complex turn based system that is structured in phases. Every phase is 4 APs and standard characters have 8. So they act in two distinct phases. Attack attack, or move attack, or attack move, or move move. Within the same phase, the units are ordered by the initiative roll (it's fixed in the same combat turn, so it applies to every phase, but is rerolled with the start of each turn). With the caveat that a critical success or failure roll can also push the unit up and down the phase system. Optionally the player can also decide to "delay" the action and act in a following phase (but you usually want to act first and kill the monster before the monster has its chance of attacking you).

Next I'll explain actual attacks, but the really CRAZY stuff is when I'll explain the details of ranged combat...

(a simple example, the colors on the right let you visualize the phase. So you can see all the known units are acting in the SAME phase, yellow, in this example. Because they are all the same phase, the order is set by initiative. The initiative is the number you see on the left, from higher to lower. The Troll is red because it already made its action, now it's the player's turn. When this phase is over the units that have APs left will be correctly moved on the next phase. If the next phase is empty, a new turn automatically starts)

(all of this, statistics, phase system, initiative, movement, attack, use of APs and so on is already coded. I think I haven't yet coded the critical successes and failures on the initiative roll, but I'm pretty sure it's the only feature described that is not already in code and working smoothly)

 

[hrose]

I'm going to write about a few aspects of melee combat.

Attack Profiles

First, an aspect of design that can only be coded when everything else is in place: since I plan for a complex ruleset and a lot of specific options about a single attack, this can obviously become very cumbersome to play, especially if you have to fight against very easy goblins and yet you have to go through the laborious process of selecting a number of things. So I thought a system that would offer flexibility. You should be able to set, for each character in the party, a number of attack presets, call them Attack Profiles. They are pages that you can bring up and set the various attack options in detail.

This means that when you want to attack you can just select the profile you want to use. Or, you can bring up the attack profile and set every option individually if you need something tactically specific this time. The system is as flexible as you need: when you open an Attack Profile you have an "Ok" button and a "Save" button, so you can change the profile and save it, or you can also use that profile as a basis for that specific attack, tweak the options you need tweaked, and use those only for that one attack.

--

When it comes to actual combat there are simple systems like AD&D where a lot is abstracted. So armor for example doesn't do the armor job: absorb damage. Instead it is abstracted as a cumulative value. Armor makes you harder to hit, and the dexterity bonus is added to it. It's of course not very realistic: a guy with a small dagger would still consistently hit and damage a guy wearing full plate. Using a long sword would do more damage, but would still hit the same as the dagger.

That's why I always liked a system that was more accurate and less abstracted. A system where armors works as passive absorption, making you easier to hit if it's heavy, but still soaking up that damage. And I also like a system where defense is also modeled. So the possibility to actively parry, dodge and counterattack. If you don't parry or block, then you'd only have your armor to protect you.

A simple system that does this is Stormbringer (and Elric!). You have a percent skill to attack and to parry. If your attack is successful, your target can decide if attempting a parry or a dodge, or just soak that damage with the armor.

But a lot more detailed and complex, and still on the same line, we have Harnmaster and Harnmaster Gold, and that's the foundation I use.

Harnmaster offers the possibility of the class-less, level-less, and skill based system.

It seems complex but it's actually fairly easy and not very different from Stormbringer. It's still an active system where both attacker and defender perform an action. the result is mapped on that chart. Each as four entries: Marginal Failure, Marginal Success, Critical Failure and Critical Success.

What I don't like of this system is that it ends up very random, and because combat is very deadly the result is that the dice decide too much. You really don't have a lot of control on the outcome of a fight. That's why I still use that system but by heavily reducing chance.

In Harnmaster, to make the system simple to use, a "critical" is if you roll a 5 or a 0 for the unit of the % dice. So you roll a 45 or a 40, or a 35 or a 30 and it's "critical". If your skill value is above what you rolled, then it's critical success, if it's above it's a critical failure. That means, say you have 70% skill value, that you have a 14% critical success rate, and 6% critical failure rate. Or, regardless of the skill value, that every dice roll will always have a 20% critical rate. Or, that 2 rolls every 10 are always criticals.

It's a bit too much, so I decided for something milder, and eschewed toward successes more than failures. I use the index of the skill value for critical successes. So it goes from a minimum of 1 to a maximum of 10. If you have a 70% skill, then a critical success is simply if you roll 1-7. Or 7%. For critical failures I use the same system, but halved. So it would be just 1% or 2% (I don't remember how I rounded it).

Even with criticals downplayed, we still have a system where the dice decide a lot. You need to roll less than your percent skill to make a successful attack or parry. Other systems like Rolemaster and The Riddle of Steel use something like a flexible pool where you decide how much focus you put on attack and how much on defense. It's realistic, the more you push for an attack the more you expose your defense. And what I liked of this system is that the outcome isn't simply: successful or unsuccessful hit, but also the AMOUNT of success that determines the outcome. So not just a binary hit/not hit, but a granular outcome that is determined by the difference of the attack and defense.

That gave me an idea to change Harnmaster rules, and use the hit-roll to affect the damage.

See this example (it's read from bottom up):

Player skill is 48/65. 48% would be the sword skill you see in the paperdoll. It grows to a final 65 because of added bonuses I'm not explaining here (like weapon class). But 65% is the percent roll you need to pass in order to make a successful hit. It's a success because an 11 is rolled. The enemy has a 50% change to parry, and it fails rolling a 81.

I use then a system borrowed from The Riddle of Steel to determine the direction of the swing. Damage type is distinct in Blunt/Edge/Point, so a sword would have all three options, doing different damage on each type, and armor also having different protection values for each attack aspect. The armor system also uses a layered system, so you can wear various layers of armor, all doing their job. It would look something like this (the bottom right part):

The first line with colors lists the possible locations, like SKull, FAce, NecK, SHoulder... and so on. You then see the individual armor pieces, their material, weight and at the bottom the actual protection distinct for each weapon aspect. So if you hit the shoulder with an "edge" type of attack, the armor absorption will be 7. If the attack was Blunt, then armor would only absorb 4.

As I said my idea was to use the to-hit roll to play a role in the severity of the damage done, so retaining the mechanics offered in The Riddle of Steel and Rolemaster. So that's why I use a "Damage Index" added to the damage value. In the example the index is 6. Or 5+1. 5 is obtained from the difference between the target number (65) and the dice roll (11). Removing the unit it's 6-1, so 5. This means, mechanically, that the more "distance" between the dice roll and your target number, the more successful is the attack. For the monster parry failure we use the same calculation, but halved. The troll rolled a 81, the target was 50. So 81-50, or 8-5. It's 3, but it's cut in half, and truncated it becomes just one. 5+1 become 6, and that's the value ADDED to the damage.

Notice how for damage we don't roll any dice. Damage is almost entirely due to the character's strength. A sword is a deadly weapon, but it's the strength of the swing that determines actual damage. That's why you see that the weapon type only contributes to a 5 to the total damage, while we use 16 taken directly from Strength attribute. Then we add the Damage Index, that is the only random part, and that is deduced as explained above, from the success rate of the to-hit roll.

And then you subtract from that damage the armor value to see what actually happened. Then the system stops for the time being because I still haven't coded the actual health system. Harnmaster uses a very complex system of wounds. I want to retain a lot of that, but I'll break down in a different way: there will be an individual wound for successful hit, that depends on the severity of the hit and the location (as both Harnmaster and Riddle of the Steel want) but after the effect is applied (statues like shock or bleeding) the rest is going to use a general health pool that has various severity stages (like Cyberpunk 2020, for example). The difference is that Harnmaster instead continues to keep track of individual wounds, and even when you heal them you treat each wound received individually. Since this is just too cumbersome, my system is going to use the individual wound to determine the effect of an attack, but after it's applied that wound is "forgotten" and the effect goes in a generalized health pool with tiers/severity.

This should give an idea of how the melee system is working. There's more to it in design, for example the possibility to focus defense on a particular spot that is exposed. For example you could direct a shield to focus blocking blows to your head, but at the expense of exposing yourself from attacks aimed at different locations. The shield is also both active and passive. Active because you can roll to block an attack, same as you can use a sword to parry it. But you can also use it passively, and so it would normally work as simple armor absorption as long your are hit in the location covered by the shield.

Overall, you have a % skill you have to roll under to make a successful attack, and your target will do the same to try to parry it. The original rule I added is that I use the amount of "success" in that roll to also determine the amount of damage, instead of rolling dice for damage separately as one usually does in most systems. So it's not anymore just a binary hit or not, but it also says how effective was the attack, depending on your actual skill value and the hit roll.

- Active parry, block, dodge, a successful attack isn't enough if your opponent keeps successfully parrying
- to-hit roll used to adjust damage, so skill values still matter even in the randomness of the dice
- Use of the arc of the swing, weapon aspect (blunt/edge/point), and precise hit location
- Armor is purely damage absorption
- Armor gives penalties through encumbrance system, so you don't automatically put on the best protection
- Armor can be used in layers (but encumbrance goes up exponentially)
- Melee damage is mostly driven by strength

 

[laclongquan]

pity post~

Also bad tittle~ We want something more concrete.

 

[hrose]

I'm using the writing of this post as a personal refresher of ideas

For me this is meaningful because as far as I know I'm the only human being out there attempting this "style" of game. You can see other RPG projects here and they use all similar rules and similar combat systems. I'm doing something completely different, and I'm doing it not by inventing, but by looking back and re-learning the past.

 

[hrose]

I said that ranged combat was where things really were kicked to next level absurdity, so I'll try to explain that now. Since it's pretty much the most recent system I built, it's also the most technically complex and that went through many stages. I cannot write again all the various steps that brought me there.

In general I relied on The Riddle of The Steel ruleset, that has the most complex combat system I know, and more specifically I used also its expansion called The Flower of Battle to make things even more messy. Hybridizing as usual with Harnmaster, since most other systems don't have much to add of their own when it comes to ranged combat.

- Draw Weight. When you try to equip a bow a check will be performed to establish if you can actually use it. The attribute used is "Strength", because you need to be strong enough depending what type of bow you use (every bow has a different strength requirement, a short bow could have a requirement of 60 and a long bow of 90). But it's also important to notice that in melee Strength is added to the damage, instead a bow has its Strength fixed and unchanging. Your own strength doesn't make the arrow hit stronger or fly faster, but your Strength is required to know if you're strong enough to draw that bow. Formula is STR + Skill Index * 5. So you have 12 Strength and your bow skill at 60%, it becomes: 12 + 6 * 5 = you can equip a bow up to 90 Strength.

- Nock and Draw delay. Nocking and Drawing are separate actions, along with taking Aim and Shooting. Because I have a complex turn based system all these four actions are modeled individually. It's useful because if you are engaged in melee combat, you won't be able to Nock and Draw. But if you instead prepared the arrow you will indeed be able to shoot at point-blank. You can also nock the arrow, move, take aim and shoot. Instead, whenever you move your aim is reset to zero, but with the arrow still nocked. The standard formula right now says that on average it takes 2 AP to nock, 2 AP to draw, 2 AP to fire (and do notice I haven't included the time it takes to aim). In general the idea is that nocking/shooting process takes quite a bit longer than swinging a sword (4 AP).

Those values are all modified by character statistics:

Nock can be reduced to 1 AP if Agility is at least 16.
Draw can be reduced by 1 AP if Dexterity is at least 15, and by 2 AP if Dex is 18. So a character with 18 Dex has Draw action performed automatically at zero delay.
Fire can be reduced by 1 AP if Dexterity is at least 10, and to zero if Dex is at least 20 (above normal humans).

A character with 16 Agility and 18 Dexterity would be able to go though this "dance" at 2 AP total, 1 for nocking and 1 for firing. But we still need to consider aim.

- Aiming mechanic. The idea I had was to use Dexterity to "fill up" an aim gauge. Of course we still have percent base "Bow skill". So you'd normally roll on that skill, after adding penalties for distance and whether the target is moving or not. But instead of rolling against the skill target number, I added the aim mechanic. Basically you can fill up your aim, up to the value of your skill. The "time" it takes for aim to fill up to the skill level depends on the Dexterity character attribute. That means that an archer with very good overall skill but poor Dexterity will still take a long time to aim. He hits well, but he's slow. Whereas a character with low skill, but high Dex would aim quickly but still miss a lot because the skill value "caps" the value of the aim. He's fast, but untrained.

At least 1 aim is required to shoot, so in the example above the guy can shoot an arrow in 3 AP, but at what %? The aim follows a progressive formula. It means if grows slower the more time you take, again to simulate what would happen in reality. Time improves the aim, but with diminishing returns the more time passes. I think right now the formula is: Dex *2, Dex, Dex/2 (not sure if I tweaked this). So, the first aim you spend gives you twice your Dex value. In the example above it means that with just 3 AP you'd be able to shoot at 36% (as long your bow skill is above that). It still sucks, so you want to aim more. The second AP you spend on aim (so the whole action grows to 4) brings your aim to 54%. This is better. But from now on, every new AP spent on aiming will only improve the aim by a 9% (Dex/2). Say you have 70% bow skill, to use that 70% fully the whole process would take you 6 AP total: 1 AP for nocking, 1 AP to fire, 4 AP to aim fully. So 6 AP to fire an arrow at max potential, considering that a character that has 16 Agility and 18 Dexterity is already close to the very top of the line.

- Range tiers. Depending on rules you can deal with the problem of ranges differently. I use a mix of Riddle of Steel and Harnmaster. Instead of having fixed range tiers, and then the weapon detailing the penalties you have on each, I instead have six range tiers mechanically identical (close, very short, short, medium, long, extreme). Each ranged weapon will tell you the distance value corresponding to each tier. So what counts as "very short" range for a type of bow would differ from another bow type. The penalties would work something like this, for each tier:
+10 +5 0 -20 -40 -80

Meaning that at close and very short range you have a bonus to the skill instead of a penalty (this bonus is added to the aim for the first three tiers, and to the cap for the last three).

- Movement penalty. depends on the movement of the target. Whereas range penalty is fixed, the movement penalty is proportional to the character skill. So the target might move in a way that gives a 25% penalty. This penalty is applied to the skill. If the character has an 80% skill then he shoots at 60%, if he has 40% he shoots at 30%. So it applies proportionally and in a different way to the range penalty.

- Aim location. The idea is to accept a severe penalty to the aim, to aim precisely at a body location. In this case deviation will be calculated depending on how successful was the roll. You aim for the head, but you don't quite nail it, you're still more likely to hit high on the body than to hit the legs. Depending on how much you screw up the actual roll.

- There were infinite attempts at modeling various "misses". For example what happens if you miss your roll. Do I calculate a deviation to see where the arrow goes and if it hits a different target? What happens if there are five monsters in a straight line and I miss the first, do I roll again to check if the one behind is hit instead? Or, I have a mage standing behind a warrior, and I really want to hit that mage behind. Do I still roll to check if the arrow hits the warrior in the front even if I was aiming at the mage in the back? All these possibilities are tactically interesting to model. I don't quite remember the solution I coded, so take it as an aspect of design that is still subject to experimentation.

- Damage. Another rule is "impact speed" and it is tied to the distance tiers I explained above. Not only distance gives a penalty to the to-hit, but the impact speed of a missile decreases with every tier. If a short bow firing within 15 yards (very short) has 6 for impact value, the same bow firing within 60 yards (medium) has only 3. To the impact speed you add the weapon Strength. It depends on the hardcoded, fixed Strength of the ranged weapon, plus the effect of the arrowhead type. So there might be arrows that make worse wounds, cause bleeding, or pierce armor, or stun.


More stuff to add to those "basic" mechanics.

- Dungeon ceilings. Realistic ranged weapons can shoot a very, very long distance. But this is done by shooting "up", as an arc instead of taking a straight aim. But in dungeons you have ceilings, so the range tiers will have to consider how high is the ceiling to decide how far you can fire. Because of this, everything past the "medium" tier counts as "indirect fire", and all indirect fire is forbidden in the presence of ceilings.

- Dodging. You can try to dodge an arrow, it's not very efficient. I planned a system that uses the dodge index and scales with distance tiers (if you are far away you are more likely to be able to dodge, if you are close it's more like pure instinct, and very unlikely).

- Blocking. There's a split between high velocity and low velocity missile (like a rock). Range is not generally used for block calculations, but very long ranges make the missile turn to low velocity. On low missile velocity you can also use a weapon to block, wheres high missile velocity can only be parried by a shield. Chance to block depends on both the skill level (active block) and shield type. If a Tower shield allows to block an High Velocity arrow at -10%, a Buckler gives instead a -35% penalty, being smaller. Also to remember that a shield is also ALWAYS working as passive armor, meaning that if you aren't trying to block and the arrow hits in a location covered by the shield the armor absorption could effectively work as an active block (the difference is of course based on the material, and the fact that an Active Block is able to block an arrow regardless of where it's going to hit).

In general the idea is that a good swordsman can take a while to get hit in melee, because he'll keep actively parrying the blows. This makes ranged weapons much more deadly, because on a successful shot the dodge chance for that swordsman will be very low. So he's gonna be hit. He's gonna be hit if he doesn't have a shield. A shield gives some decent chance at defending from successful ranged attacks.

Otherwise, the system seems to favor a lot ranged attacks. But it's supposed to be balanced by two aspects:
1- You can nock the arrow as long no one is around you. So when you are engaged in melee you are going to switch weapons.
2- It still takes a sweet long time to properly aim.

That means archers are deadly as long they stay in the back and are left alone, and it kind of makes sense.

 

[laclongquan]

No graphic cue is just sooooo turn off~

 

[Isiloon]

Hope you don't mind if I steal your combat mechanic ideas for my own game that's way better