Time to re-examine the ruleset in a slightly different angle. What is it I’m trying to achieve?
- A fairly lightweight game
- Mechanics to reinforce the sense of being a reptile
- Different species feel different
- More broadly, reinforce the feeling of being a cyborg reptile space wizard secret agent
- Designed to support a variety of secret agent activities, not just for combat
- Quick, fairly pulpy combat
- Competent PCs who excel in particular areas
At the moment, what have we got?
I’m fairly happy with the Heat Points model. Ambient temperature is important, characters have to consider gear and activities in terms of temperature, and it highlights some differences between herps and pretty much all RPG characters ever. It also lets you turn this consideration into an advantage, not just another thing to micro-manage.
I think at the moment species are perhaps too specifically defined. Instead, I might just suggest you pick one major and one minor lineage trait that are particularly important to your character, and these grant you benefits. Other traits aren’t necessarily absent, they just aren’t game-mechanically relevant. I do want to try and ensure that these traits feel significant.
For supporting different activities, I feel like there are three main approaches. One is to have extremely basic mechanics, so that everything is done on a small pool of abilities, or indeed without any stats at all. The second is to have relatively detailed mechanics for all the desired activities, so that all of them can be modelled interestingly without getting repetitive, and none of them are handwaved with a single die roll. The third, my current approach, is to mix-and-match sets of abilities, so that within any activity there will be a range of different abilities coming into play depending what you actually do. I’m hopeful this will prevent the tendency to veer towards either combat, as the most mechanically interesting, or freeform gaming that relies on player ability rather than PC.
Pulpy combat... well, characters can’t die from game mechanics and recover rapidly. Weapons are assumed to be powerful, so it’s just about hits and armour at the moment – and I’m seriously considering paring this back even further, perhaps just having two or three classes of armour (Light, Heavy and maybe Medium), which would make it lighter. Competence plus powerful weapons should equal fast combat. This should help prevent it dominating play.
Magic has its own system that will (hopefully) make magic feel like a Thing. I'm hoping to evoke cyborghood through pervasion rather than lots of mechanics, and aim to model upgrades as reliable slick effects, to contrast with esoteric magic (magic may need to get even more esoteric, but we'll see). There's only so much evoking you can do with mechanics; the setting and the playstyle will need to handle a lot of that.
As ever, I’m drawn back to the issue of core mechanics. This is mostly because I am phenomenally indecisive, but also because any system is going to have benefits and drawbacks.
Single die rolls are simple, but are best for modelling highly random events where the outcome isn’t a bell curve, like where many factors complicate the outcome. They work well with modifiers, and providing degree of success isn’t important, this combination can readily represent competence that provides a baseline skill.
Dicepools are good for modelling things that it’s satisfactory to model as a bell curve. This is good for ensuring characters are consistently competent (or incompetent) at things. It's more difficult to intuit your odds of success with a dicepool, due to the increased complexity, but a basic roll-and-pick system is simpler to actually do than roll-and-add single die. Designing dicepools is (to my eyes) complicated because of that same complexity, which makes running calculations in Excel a pain. While the chance of improbable failure/success is reduced by that tendency to the mean, it's not prevented, so extreme rolls are still possible. Variance also tends to increase with skill, which can be unwelcome, though some models avoid this.
The current system is still d20 blackjack with modifiers. Difficulty, special skills and any other relevant factor are represented by modifiers. These may make success assured (attribute reaches 20) or impossible (attribute reaches 0). As modifiers are 2 or 5, the maths is relatively simple. The penalty die system means any roll that's less than the penalty die score is a failure if that impairment would hamper that task.
A second possibility is a dicepool. There are many dicepool models: variable/fixed target number, any/multiple successes, fixed/variable die size, count/sum, and so on. Far more than I can easily begin to evaluate, to be honest.
My inclination with dicepools would be to have fixed target numbers (keeping maths simple) and track difficulty by number of successes needed. Training could be represented by granting an automatic success, so professionals would succeed at even non-trivial tasks without rolling - only quite tricky ones would call for a roll. Other factors could be represented by granting additional dice. Penalty dice... I dunno. With small penalty dice you could discount successes equal to the number rolled, but that gets brutal fast.
I'm also a little bit cautious of how easily dicepools seem to render targets impossible. A modifier-free system doesn't easily handle auto-success, but if you don't have enough dice, you can't reach a target. This can happen in most systems, it's just something I want to keep an eye on. On a related note, I'd want to adjust the scope of attributes, since rolling 10 dice is both highly variable and a bit faffy - we'd probably want them to top out at about 5.