Level as saving throw

A Close Call (from Wikipedia)

So if the saving throw is really just a shorthand for level, a reward for extensive successful play, why do we need another number at all? Maybe Swords & Wizardry doesn’t go far enough with its single saving throw. Why can’t we just use the level directly? As frequently noted, I really like saving throws (see here, here, and here). Seriously, why am I writing this? The save system is maybe the least perverse part of the traditional game. Anyways, I have these ideas in my head, so here I go.

There are a few problems with using level directly. For one thing, saving throws shouldn’t be too hard at first level or too easy at high levels. In the original game, a fighter has to roll 12 or higher at first level to successfully save versus death. A beginning magic-user has to roll 15 or higher to save versus spells. The endgame saving throw target numbers range from 3 to 8, depending on category and class. They cluster around 5. These probabilities should be our guidelines.

In discussion on G+, Lance T. pointed out that Searchers of the Unknown does something very similar:

Saving throws: when such a roll is needed for any reason, roll 1d20 under the character’s level, +4. So 7th level adventurer must roll under 11 to escape a magical charm from a harpy. This “level+4” rules apply to every other action which aren’t covered by the “stealth & stunts” rule, but fits the common adventurers knowledge like searching for secrets doors or picking locks.

That’s good, but the magic constant of 4 is aesthetically unappealing. Also, as written, at high levels the saving throw becomes either 100% or 95% (if 20 is always a failure). I think more risk should remain at high levels. Like many D&D rules, it breaks down somewhat at high levels. It would be nice to avoid that.

Claytonian JP suggested how bout just a d20, roll under level plus ability mod? Lethal, but dramatic. That’s relatively pleasing. Maybe factoring in class too, so it would be under level + class and/or ability bonus (assuming something like the B/X ability score modifiers). The class modifier could be get a +2 bonus to the if it made sense (magic-users versus spells, thieves versus traps, fighters versus dragon breath, etc); similarly for ability scores. This means that in the most favorable case, with class factoring in and a maximum relevant ability score, the chance of success at first level is 6 or less, which is 30% (and most saves will have the insane difficulty of 5% at first level). Given the guidelines given above, this is probably too hard at first level, and also unbounded at high levels (and thus too easy). So that doesn’t quite satisfy the requirements for a good saving throw system.

Paolo suggested limiting the level bonus to the hit dice total, which is nice and elegant, especially for a system that caps hit dice at or around name level. It also made me think of how I have hit dice and traits set up in Hexagram. Hit dice total, like pretty much everything in Hexagram, is limited to 6, and characters get one every level for the first six levels. So that could be the beginning of the save bonus, and each path (battle, guile, wonder) also has a pseudo-defensive trait that could be applied to appropriate saves. This is starting to stray from the clean idea using the level number directly though.

4E (and 3E, I think) use “half-level” bonuses for some things, but I think that’s ugly. If we’re going to replace the saving throw subsystem with something more streamlined, I want it to be really elegant. This is similar to something I hacked into my previous Fourth Edition game (see the luck throw), but that was designed to explicitly fit with other 4E assumptions (like the level range of 1 to 30).

What about d20 + level, 16+ = success and 5- = failure? That’s simple enough to remember, is symmetrical, and handles low and high levels well. Ranges from 30% success at first level (since there is a +1 from being first level) to 25% failure at level 10 (after which saves wouldn’t functionally improve anymore, though you never need to track anything other than the level). That’s kind of what Paolo suggested, but is independent of hit dice (which is nice, as level and hit dice are not always so correlated; consider the all d6 progression of OD&D that uses +1 bonuses at some levels).

Here are some examples of how this system might work. Roll an 8 at level 3, total is 11, which is less than 16, so failure. Roll a 10 at level 6, total is 16, which is a success. Roll a 3 at any level, that’s a failure because it’s 5 or less. Roll an 18 at any level, that’s a success because it’s more than 15. I think this satisfies all my original contraints, though you have to remember the slightly odd failure and success ranges.

So… what do you think of replacing saving throws with an explicit level check? Too far off the deep end?

14 thoughts on “Level as saving throw

  1. Ramanan Sivaranjan

    Is it just a shorthand for level, though? I feel like there is also a lot of flavour that comes from the each of the Saving Throw types, and how they differ from class to class. Dwarves are super hardy so they have good saves. Haflings are super lucky so they have good saves. Fighters save better against breath weapons. Etc. 5e keeps this to some extent by tying them to your attributes, but you do still lose something.

    That said, the extra set of numbers seems silly. I was glad to see 5e attempt to simplify things.

    1. Brendan

      I agree; I like the flavor too. The save versus “death ray” has some power above and beyond the mechanics. That said, I think a class or race bonus could accomplish at least the kind of distinction that you mention. I do think that traditional saving throws are dominated by level (though the dwarf +4 is a big deal and kind of an exception; I would point out though that within the set of dwarves, level is still dominant in the same way).

      That’s why I worry about making saves based on ability scores too closely. I think that makes the variance in starting saves too high (and tempting to min/max). Even assuming the ability modifiers are what counts and not the score itself, that is a potential difference of 7 points (or 35%) between a starting character’s save versus poison, just based on the ability score roll.

    2. Drance

      NOTE: Castles & Crusades simplified things way before 5E was even a twinkle in Mike Mearl’s eye. The SIEGE Engine mechanic tied saves to the D&D attributes years ago (2004?). God bless the Troll Lords! Trollzah!

    3. Peter K.

      I’d actually like to see how it played out with ability score being the dominant factor and level not so much. It would reinforce the idea that for last ditch attempts at survivability, your training isn’t everything, and some characters just have what it takes starting out.

      The variant I’ve been considering is that saves are just:

      Ability score + 1d20 vs. difficulty of 21
      (occasionally modified)

      Level would come into play more gradually in that each even level a character gets to increase an ability score by +1 (up to racial maximums). Possibly every odd level they get to shift a rank from one ability score to another.

    4. Brendan

      In your example, is “ability score” the full score or the modifier?

      Trent B. from G+ uses this, which seems interesting:

      Roll under half of the relevant ability+level to save. Assuming more or less average ability scores, it numerically correlates quite strongly with most existing old-school save progressions. Once you have dudes with multiple 14+s it starts to blow out, but whatever.

      (That’s a paraphrase-quote, not a literal quote.)

    5. Peter K.

      Ability modifiers are small (even smaller pre-3E). Small modifiers highlight the importance of training and experience (skill increase or character level).

      But the ability scores (not modifiers) offer more dramatic variation, showcasing innate character differences which don’t normally change much. I thought it would be interesting to use the scores themselves as saves, to suggest that you are dealing with situations that normal training doesn’t account for. When afflicted with poison, the hearty peasant may possess better stamina than the experienced knight, even though the knight could survive longer in combat.

      But on the other hand I also like the idea that a character could, very gradually, improve their innate ability scores without resorting to a magic book of attribute increase.

  2. heylonghair

    I’m sorry but I’m a little confused by your example.
    I read it as 16 or greater is a success and 5 or below is a failure. In your example though you keep referencing 15. Was that supposed to be 15 or greater is a success? Then you say rolling a 3 at any level is a failure because it’s under 5 but if you rolled a 3 at level 5 wouldn’t it be 8? Are you adding the level to some rolls and not others?

    1. Brendan

      Rolling 1, 2, 3, 4, or 5 is like rolling a 1 on a traditional D&D attack roll: it always fails, no matter the bonuses. So even if a character is level 20, that character still has a 25% chance of failure. Rolling 16, 17, 18, 19, or 20 is like rolling a 20 on a traditional D&D attack roll: it always succeeds, no matter the bonuses. If rolling any other number, then the level bonus comes into play, and you are shooting for a target number of 15.

      Maybe it is too complicated, but structurally it is identical to an attack roll against AC 15 with different auto-hit and auto-miss ranges.

      You’re right the first “less than 15” should read “less than 16” though, so I fixed that.

    1. Brendan


      The original tables don’t seem to be that difficult. A first level fighter’s save versus death, for instance, is 12. The first level saves in Men & Magic range from 11 to 16. There are none as hard as 18. The average target number is 13.93. AD&D makes them slightly harder (a first level fighter has two saves at 17), but not that much if you look at the average. Also, doesn’t AD&D give you more ways to get bonuses on saves?

      On G+, Roger left the following comment:

      Yeah, I have my easy save (body) start at +7, hard one (mind) at +3 and middle (speed) at +5, and you have to roll 20+ on d20 plus those modifiers and any ability and class modifiers that are relevant.

      Those kinds of modifiers along with proto-3E-style fortitude, reflex, and will saves would probably approximate the original spread pretty well, I think. Perhaps that would be the most comprehensive way to do it: level bonus along with category penalties.

    2. Matthew James Stanham

      If you are adding level to saving throws, though, the spread needs to be longer. For example, comparing 16 + Level to original saving throws you get this spread:

      01: 12 vs. 15
      02: 12 vs. 14
      03: 12 vs. 13
      04: 10 vs. 12
      05: 10 vs. 11
      06: 10 vs. 10
      07: 08 vs. 09
      08: 08 vs. 08
      09: 08 vs. 07
      10: 06 vs. 06
      11: 06 vs. 05
      12: 06 vs. 04
      13: 04 vs. 03

      On the face of it, that looks okay, but then we note that many of the other saving throws simply do not scale that rapidly. For example, the magician at level 13 has saves of 8/9/8/11/8, and fighter 4/5/5/5/8. In the former case it is a lost cause to add level in order to get anything like a similar progression, in the latter case it is clear that to get a similar spread of saving throws you need to start a bit higher.

    3. Brendan

      Remember that by hypothesis 5 or less is always a failure in this proposed system.

      It’s true that as stated it homogenizes the classes. However, if fighters got +2 to physical saves, magic-users got +2 to spell saves, and thieves got +2 to agility saves, that would make them unique again.

      I’m not claiming the modelling is perfect. One is probably going to either be closer to the original numbers in early levels or later levels, but not both. As stated above, I think the exact probability is not that important at high levels as long as there is still a nontrivial chance of failure, so I would probably privilege the functioning of the low level saves.

    4. Matthew James Stanham

      I reckon if you are using +2 to saves for specific classes it is even more reason to start higher. One alternative fix, though, is to use half level instead of level. That lets you start with a better probability without thundering down to “5” before hitting ninth level or whatever. My feeling is that probability distribution for (A)D&D does contribute significantly to the feeling of the game, and requires more careful consideration than game masters sometimes deign to give.


Leave a Reply