This old pearl has been running around on this site for a while now.
“D100 role playing games have an undeserved reputation for swingy resolution mechanics”.
I say undeserved mainly because the perception and the reality, especially in comparison to some of the big RPG names out there, are more about misunderstanding than reality.
Ironically, Achtung Cthulhu can be used with Savage Worlds, a game with an even bigger rep for swinginess.
d20 games have a linnear curve to their rolled value chances, but unlike the d100 games, they usually have roll over not roll under mechanics. It’s an illusion that this is less swingy, but that is the perception. The reality is, it is more limiting, creating a “dead” zone at the extreme ends, which leads to player dissatisfaction with early editions “topping out” chance curves (ever played a level 12 3e Fighter against some level 1/2 goblins?).
Even those with a curve like 2d6, 3d6 or 2d10 games, still have the same number of actual chances of rolling any single combination, they are just clumped towards the middle (called a Bell curve). This also has its own issues, mainly with mods that can become quite unbalanced at high or low levels.
Even Savage Worlds, a game with a very small and tight resuts spread has a reputation for swingness due to its “exploding” dice, so there is always a catch.
Take swinginess away completely and you have no risk.
D100 games offer a wide spread of evenly distributed values, and an equally large number of options within those. The two dice used allow for a built in critical system (doubles), a gentle growth path (1% increments) and several ways of implementing the actual mechanics.
To alleviate the swinginess of d100 skill systems, there are a couple of basic tricks available;
Let the players choose the order their dice are read in for unopposed, non time critical tests or when they have the “advantage”. This means they will have a roughly 75% chance of rolling 50% or less. In other words, introduce a Bell curve of sorts (maybe Bell-bottom curve). Combat and pressure tests should stay “taken as they come”, but basic tests are more player controlled. This also allows a player to choose between two passing tests. In my own system, difficulty is measured in “how the dice are read” and for extreme tests how many are rolled, not mods.
Increase the core skill level, either with a fixed base or increased skill starting formula. BRP uses the sum of 2x3d6 Chrs (average 21%), but maybe use three or even two chrs and a base depending on untrained likelihood (+0% for Jumbo Jet pilot, +20% for Bike riding).
Always pay attention to the consequences of failure, especially critical failures. Make the penalty (or reward) fit the risk. I use a table for my “doubles are crit pass/failure” system, with escalating effects towards the centre of the spread and “softer” results at the extremes. Another option is to have levels of success or failure based on the variance of the needed roll and the acual, a bit like difficulty, but backwards. This removes the rigid line of pass/fail and turns it more into a softer indicator of success-failure levels. 20% levels seems to work.
Re-interpret “failure” using the “fail forward” or “failure that succeeds, but“ models, especially when a test comes close (maybe a 20% either way rule). There are plenty of good systems out there to bring role playing into failure mechanics rather than just maths.
Re-define the test parameters. Will the lever break on a fail, or just not turn enough? Are you maybe even turning it the wrong way, so no test is needed next attempt.
Only test when needed. The test of a good GM is knowing when and when not to test.
Break a test into parts, increasing tension and story. This makes for a better game and story anyway, but it allows the skill level to smooth out to a more likely outcome, not a swingier yes/no question and allows the GM to apply different chrs or skills to the different elements of the test.
The reality is, most d100 games, indeed most RPG’s suffer from some form or other of mechanical swinginess, which is where the risk is found. They all come up with methods to get around this (or not), but there is always a way.
Bare Bones uses the higher base level model, using the actual characteristic -/+ based on skill.