A surprisingly good idea came from the Atlas official forums by user: Humboldtscott for Ars Magica stress dice rolls, which is worth promoting. As said in the forum there are some quirks in ArM for dice rolls which might be confusing to new players.

I prefer that the dice be more consistent. I have proposed to my troupe the following variant, and they have agreed, but I’m curious what more experienced players and storyguides think:

With a stress die, 2-9 are just that. 1 is a potential botch. Roll the botch dice, and if any come up a 1, you have botched. If not, the result is calculated as if you have rolled a 1. 0 is a potential great success. Roll again and double the result. If 0 again, roll again and quadruple, etc. If the final result of this re-roll is less than 10, treat the result as if you rolled 10.

This keeps the stress die on a more or less 1-10 range, while still allowing for the occasional dramatic success or failure. I think this change will make dice slightly more important, but not enough to really worry me. Botches will be just as likely, but a potential botch that does not do so will be slightly less detrimental, and great successes will be slightly better, as they will always be at least 10.

I think it is worth considering.This change means that a result of a 10 on a stress roll is always at least a reasonable roll; rather than the current rules where rolling a 1 and the doubling the next result can sometimes feel like the player is being cheated if the next roll is very low.

eg. a roll of a 1 then a 2 would mean that a potential great success results in a total of 4. which is very disappointing.

It also means that a 1 is at best only a 1, and at worst a botch result. The check for botch could also be changed so that a botch is confirmed on a re-roll of 1 instead of 0, with no affect on the chance of this occurring. It grants a consistency to rolls across normal and stress rolls.

Method One: **Multiplicative Stress, minimum 10**.

So Ars Magica Stress rolls become:

- 1 is a potential botch, roll the number of additional appropriate dice where a roll of 1 indicates a botch.
- 10 rolls again, with a doubling result. However a minimum result of 10 regardless of the roll.
- 2-9 as normal.

Method Two: **Additive Stress, minimum 10**

Perhaps an alternate to this to avoid even more confusion is that a result of 10 is considered good, and the next result is added to 10, rather than doubled with a minimum. A ten is added and the player rolls again, etc. It makes the super success increase in a consistent manner, as a 10 always adds to the to next result.

So Ars Magica rolls become:

- 1 is a potential botch, roll the number of additional appropriate dice where a roll of 1 indicates a botch.
- 10 rolls again, with the next roll adding to the first. Every 10 rolled adds 10 to the value and roll another dice.
- 2-9 as normal.

eg. roll 0, then a 4 = 14. Or Roll 0, then 0, then a 4 = 24.

I think it you are using this rule it will change the spread of potential values achieved, as a stress roll which multiplies up will now always be great. It raises the average roll result slightly upward.

These alternates do not (in my opinion) change the concept of botch, pass, and exceptional results in the system. It makes a result of a 1 consistently disappointing, and a 10 consistently valuable.

For Humboldtscott’s method / Method One:

- After the first potential “crit roll” (meaning a 10 is rolled), the final result is never higher than the standard Stress Roll result if the next number rolled is 5 or more for single crit rolls. ie. roll 10, then another number.
- For dual-crit rolls (where the player rolled a 10, then 10, then another number) the difference is far smaller.
- For three-crit rolls (where the player rolled a 10, 10, 10 then another number) the result is only different if the number rolled is a 1.
- Makes any “crit” better than a standard result, but not any different to normal stress rolls at the extremely low probability.

For my suggested method / Method Two:

- The method scales in a linear way, where it is not possible to gain a good result. Suits games where rare exceptional results are not wanted, but crits consistently perform better.
- the degree of variation for very rare circumstances (10,10,10, + normal roll) is very significant; being both drastically higher is the number rolled is low (compare 8 vs 31), and far lower than standard if the number rolled is high (72 vs 39).

My conclusion is that for any method rolling two 10s, and a number greater than 4 will be a very large total (16+ stats). Also that my suggested method is pretty flawed, as it removed the possibility of “off the chart” totals for dice rolls. It would never produce totals in the 70s.

So I like Humboldtscott’s more, and think it is an improvement on standard crit rolls. Nice work sir. The spreadsheet for ref: ArM-roll-variations