Thank you all for the interesting and valuable answers and suggestions. Through your feedback I have learned that one of the main problems is that my playgroup receives too many confidence points because our play sessions are too short compared to the sessions for which the rule was designed. A simple solution would be to divide the confidence points gained in one of our sessions by 2 or 3.
However, your contributions have inspired me to develop a new House Rule on confidence that I would like to submit to you to hear your thoughts (@David_Chart ).
Aureliano Confidence House Rule
Confidence points are earned at the rate of 2 or 3 points for every 5 hours of play approximately depending on how each character meets the game requirements as indicated by the 5th edition of AM. You can have a maximum of Confidence points equal to Confidence Score x6. So with a score of 1 a maximum of 6 points and with a score of 2 a maximum of 12… and so on.
advanced optional rule
It is possible to use a confidence point to reroll a die instead of adding +3. If the character has multiple confidence scores then the player can decide how many Confidence Points to spend in the reroll (up to the maximum of his Confidence Score) and then reroll a die for each confidence point spent and use for the purposes of the test the score of the die he likes the most (the highest probably...). When evaluating which die result to keep to solve the test, the result of the first die roll made on the test is also included in the choice of the best score, therefore with a confident score of 2 and the use of 2 confidence points, 2 dice would be re-rolled and the player could choose which of the 3 die results to keep (the first one made + the two re-rolled using the confidence points). This reroll option can also be used in the case in which a zero is rolled on a stress roll to avoid a botch test, but it must be declared and chosen before checking for any botches (as it was in the 3rd edition of AM). The rerolled dice are of the same type (Simple or Stress) of the initial test. As in the original rules, confidence points cannot be used in laboratory or seasonal activities. With this system the player can defend himself from a potential botch in the stress rolls in the event of a zero, but if in the reroll a zero is rolled again, that score is kept and the normal check for botches is made. I would also add that defeats and failures in important battles or challenges make you lose confidence points (1), representing the effects of a drop in morale and self-esteem.
A player may, if he wishes, instead of rerolling a die, add a cumulative +3 bonus to the chosen die roll by spending a confidence point (just like in the official rule), and may do this with any combination of rerolls and cumulative +3 bonuses as long as his Confidence Score allows him to spend Confidence Points on that action. So for example a player with a Confidence Score of +5 (a legendary hero) could, by spending 5 confidence points, reroll 2 dice keeping the best of the 3 die results made and add +9 to the result (+3 x 3 = 9) and then calculate the total by adding the appropriate ability and characteristic bonuses.
This allows you to defend yourself from botches that potentially in combat can be the cause of a character's death and allows you to better protect yourself in defensive combat rolls by using them to compensate for an extraordinary success of your attacking monster. It also leaves more uncertainty and variability, and in this way more confidence points are spent. This alternative rule, although it protects from botches, does not protect absolutely. In practice, using this option the risk of running into a botch is reduced by 10 times, but is not completely eliminated. Since a botch in a defense or soak roll in combat probably implies the death of the character, this rule allows a player to avoid the death of their character due to a botch, but does not completely eliminate the possibility. Personally, I have always thought that seeing your character die from a botch is not the height of an epic climax. Moreover, according to the official rules of Ars Magica, a character in combat should on average make a botch in defense every 100 tests, which represents the character's life spam in a fairly predetermined way and on which the player is completely helpless, instead with the implementation of this rule the death of the character from a botch in Defense rolls is postponed (therefore more variable) but not completely eliminated. However, this could unbalance the game in the case in which the characters had too many confidence points to spend, for this reason I set a maximum of accumulable points as a rule, which also encourages players to use them. This way the number of points awarded to players has less relevance than the official rule.
What do you think? Could it work? Do you see any drawbacks to this alternative rule?