Years ago I read three graphs, showing the success pattern for novitiates, professionals, and veterans at any given skill.

The novitiates were all over the place, sometimes successful with beginner's luck, sometimes middling, most times failing.

The professionals were consistently successful, with the occasional minor mistake.

The veterans were almost consistently successful, with the occasional major failure.

This suggests the Veteran Rule to me: anyone with over 30 years of experience in a skill has a flat 5% chance of failure, regardless of whether they are skilled enough to ordinarily have a 100% chance of success.

I would roll this separately from the ordinary success chance, should one be using a  bell curve system like 3d6 versus a target number.

This models unpredictable failures, like the time I saw a video clip of a master tightwire walker fall to his death. He didn't become a master with a 5% chance of dying per attempt, but only had that chance after he had become a veteran in that skill.

There you go.

So...you get worse after you reach mastery? No thanks.

The problem with realistic skill representation is that people tend to get up to 95% and then develop fractions of a percentage from there.  The Olympic games are often won by the narrowest of margins where the tiniest advantage can be measured and quantified.   This is somewhat reflected in exponential skill and level costs but really it's not very exciting.

In my percentile games I treat doubles as exceptional successes and failures, but they're also specifically flukes.  Thus rather than the guy with a skill of 98 always having exceptional failures they're fluke failures.  The only kind they're likely to ever have.

As far as failure ranges I always liked how High Fantasy started combat skill at 100% and then reduced it by defense.  I didn't go the route with my own designs buy you can generally scrounge +40 by taking extra time and other precautions.

I was always a fan of Silhouette's mechanic where you got 1-5 dice (six-siders) based on your skill level, but when you rolled you only kept the highest result (or if there were multiple sixes you added +1 for each additional six rolled).

The result was that novices were all over the place, but as their skill grew it trended to the upper range and eventually got to where a "six" was routine and the true masters got their "win by narrowest of margins" factor against each other in those extra sixes that only show up reliably when you have 4-5 dice (but might show up as a fluke at even 2 dice).

Quote from: Chris24601 on August 06, 2024, 07:55:07 AMI was always a fan of Silhouette's mechanic where you got 1-5 dice (six-siders) based on your skill level, but when you rolled you only kept the highest result (or if there were multiple sixes you added +1 for each additional six rolled).

The result was that novices were all over the place, but as their skill grew it trended to the upper range and eventually got to where a "six" was routine and the true masters got their "win by narrowest of margins" factor against each other in those extra sixes that only show up reliably when you have 4-5 dice (but might show up as a fluke at even 2 dice).

This might be a tangent, but one issue I have with these types of 'throw a hand full of dice rules' is my own personal dice cursed ill luck. I ran a Traveller game modified by an idea I had from the Alien RPG. Throw additional dice that are different color than your normal d6. Keep the highest 2, and any of the other non-normal dice are 1s then something bad happens. When my girls and I were doing a test run & I was rolling for an NPC helping the PCs fighting the monster, I had 7 or 8 dice and rolled nothing over a 3 and at least 3 of the extra dice were 1s.

With that, I'm not saying any rule like this that looks at narrating the success or failure including flukes, then be ready for some poor soul with ill lucked dice to be a multi fluke inducing outlier !

Doesn't it really depend on the difficulty of the skill (or task). Like pretty sure airline pilots succeed 99.9999999999% of the time. Even if you just count emergencies for rolling, 95% seems low

The way I wrote my current system is that if your skill is high enough then a task of given difficulty, say walking a tightrope, is going to be 100% successful.  You can just skip the roll.  However once in a while that task isn't going to be as easy as it looks.  Frex a gust of wind comes up whilst walking the tight rope, and the difficulty you thought was going to be a 13 is now a 17 and that roll is now really important.  This works for me.

Quote from: JeremyR on August 07, 2024, 02:02:33 AMDoesn't it really depend on the difficulty of the skill (or task). Like pretty sure airline pilots succeed 99.9999999999% of the time. Even if you just count emergencies for rolling, 95% seems low

Pilots have an awful lot of people helping them. Take away all the maintenance technicians, safety officers and air traffic control and see how they go.

With stupid DMs, they'd represent this by having everyone involved in the activity roll, making failure almost certain. With more intelligent DMs, this is handled by complimentary bonuses or rolls - some arbitrary number or fraction of others' skills are added to the primary roll, making success certain, except for whatever proportion of attempts you as DM want there to be a problem. You can also handle it as, when one of those involved fails, another gets to make a roll to detect and stop the problem before it's too late.


For example, Technician Drunkard with a skill of 50%, less 25% for being drunk, rolls 74 against his 25%, and fails to bolt the aircraft doors on properly. The stewardess follows the instructions, "close doors and crosscheck", and while doing so, with her own technician skill of 10%, with a +30% for a simple task (doors are fairly simple) rolls 32 against her adjusted skill of 40%, and notices an odd noise as she closes the door. She alerts the pilot who calls the ground crew to come and check the door. Senior Technician Crusty with his skill of 85%, again dealing with a simple problem gets 30%, success is certain - except the DM says that 00 always fails. Whether he throws 01-99 or 00 depends on whether he qualifies to work for Boeing.

Quote from: JeremyR on August 07, 2024, 02:02:33 AMDoesn't it really depend on the difficulty of the skill (or task). Like pretty sure airline pilots succeed 99.9999999999% of the time. Even if you just count emergencies for rolling, 95% seems low

This gets into granularity. A tightrope walker (as mentioned) only needs a single failure to fall to his death. An airline pilot has hundreds of small tasks to perform, few of which will cause his death if he makes a mistake.

However, were he flying a biplane through a hurricane, you'd want him to be a younger master, not an old vet, because one mistake could mean you're through.

Quote from: Kyle Aaron on August 07, 2024, 08:51:43 AM

Quote from: JeremyR on August 07, 2024, 02:02:33 AMDoesn't it really depend on the difficulty of the skill (or task). Like pretty sure airline pilots succeed 99.9999999999% of the time. Even if you just count emergencies for rolling, 95% seems low

Pilots have an awful lot of people helping them. Take away all the maintenance technicians, safety officers and air traffic control and see how they go.

With stupid DMs, they'd represent this by having everyone involved in the activity roll, making failure almost certain. With more intelligent DMs, this is handled by complimentary bonuses or rolls - some arbitrary number or fraction of others' skills are added to the primary roll, making success certain, except for whatever proportion of attempts you as DM want there to be a problem. You can also handle it as, when one of those involved fails, another gets to make a roll to detect and stop the problem before it's too late.


For example, Technician Drunkard with a skill of 50%, less 25% for being drunk, rolls 74 against his 25%, and fails to bolt the aircraft doors on properly. The stewardess follows the instructions, "close doors and crosscheck", and while doing so, with her own technician skill of 10%, with a +30% for a simple task (doors are fairly simple) rolls 32 against her adjusted skill of 40%, and notices an odd noise as she closes the door. She alerts the pilot who calls the ground crew to come and check the door. Senior Technician Crusty with his skill of 85%, again dealing with a simple problem gets 30%, success is certain - except the DM says that 00 always fails. Whether he throws 01-99 or 00 depends on whether he qualifies to work for Boeing.

Haaaaaaaaaaaaa ! Reading that I almost laughed out the gulp of cherry-limeade (extra lime of course) that I took before I started reading it ! NICE

Your example is a good reason I have adopted the 'Traveller task chain rule' in my games now. It works really well in Traveller & Call of Cathluhu (btb CoC and in different home brew / DG mash up games). Haven't tried it in a D&D type class game so I can't say if it works in them too.