There are two ways that spells are rolled for in D&D 5e, depending on the spell: attack rolls rolled by the players and saving throws rolled by the DM.

But as far as I can see, saving throws could be made as attack rolls instead. I know D&D 4e did it this way, with players making attacks against AC/Fortitude/Reflex/Will Defenses instead of the DM making Fort/Ref/Will saving throws and players making attack rolls against AC.

I don't really see why the designers of D&D 5e would choose to have two different ways to roll something that, as far as I can tell, could be simplified into one mechanic (and I have asked about their stated reasons in a separate question), so I wondered if I'm missing something important that distinguishes the two methods.

I guess it boils down to: Is there any difference in probabilities or practical concerns between one or the other method, or are they interchangeable?

This is related to Why do attacks function differently to spells? and Why did saving throws return in D&D 5E? This question is focused on if there is any practical difference, regardless of why they actually made that design decision.

  • 2
    \$\begingroup\$ The last question would be a near duplicate of Is "Unearthed Arcana: Players Make All Rolls" Correct? \$\endgroup\$
    – Dale M
    Nov 25, 2016 at 21:32
  • 1
    \$\begingroup\$ This is not a duplicate. The other question asks 'why was the game made how it is?'. This question asks 'what does that decision for the players?' \$\endgroup\$
    – Daron
    Nov 26, 2016 at 16:29
  • 3
    \$\begingroup\$ I don't think it's a dupe. The linked question focuses on "this is a math error, isn't it?" without trying to determine what would change between the two methods should the success probability be the same, which is what this question asks. \$\endgroup\$
    – Zachiel
    Nov 28, 2016 at 18:03
  • 1
    \$\begingroup\$ This must not be a duplicate because I can understand this question and not the other \$\endgroup\$
    – Daron
    Nov 28, 2016 at 20:19
  • 1
    \$\begingroup\$ Daron, I just did a major rewrite of the question. I think part of the problem is the quantity of text that was copy-pasted from your other two questions, which makes the distinction much harder to see. I've removed or rewritten most of the background, to try to make your main question more obvious. However, I had to use some judgement, and I may have misinterpreted the point of this question while attempting to clarify it. Could you please review the question's current wording and purpose, and confirm whether it is still asking what you meant to ask? \$\endgroup\$ Nov 28, 2016 at 21:17

2 Answers 2



There are a number of game mechanics in D&D 5e that interact with Attack Rolls/Saving Throws to create significant differences between the two. The most notable are Advantage/Disadvantage and Critical Hits.


Having advantage or disadvantage on any roll significantly changes the probability of success. While this mechanic can influence many situations in the game, there are two situations in particular where advantage/disadvantage creates practical differences between spells that call for an attack roll and those that allow a saving throw: when making ranged attacks in close combat, and when using inspiration

Ranged Attacks in Close Combat

From PHB p. 195

When you make a ranged attack with a weapon, a spell, or some other means, you have disadvantage on the attack roll if you are within 5 feet of a hostile creature.

This means that spells which require an attack roll will sometimes suffer from disadvantage in situations where those that require a saving throw do not.


From PHB p. 125

If you have inspiration, you can expend it when you make an attack roll, saving throw, or ability check. Spending your inspiration gives you advantage on that roll.

This means that spells which require an attack roll can benefit from the use of inspiration, while spells that require a saving throw cannot (you cannot use inspiration to give your target disadvantage on a saving throw).

Critical Hits

From PHB p. 194

If the d20 roll for an attack is a 20, the attack hits regardless of any modifiers or the target's AC. In addition, the attack is a critical hit ... If the d20 roll for an attack is a 1, the attack misses regardless of any modifiers or the target's AC.

This means that spells that require an attack roll have a chance to hit (or miss) regardless of modifiers, and a chance to deal additional damage. Spells that require a saving throw cannot score a critical hit.

The previous points are intended to show that attack rolls and saving throws are not simply interchangeable, due to the ways that they interact with the larger system. In particular, the rule regarding ranged attacks in close combat makes spells that require a saving throw a bit stronger at short range; replacing the saving throw with an attack roll would weaken them. Recall that D&D 4e, which uses only attack rolls, has to have additional rules for 'close' attacks in order to allow short-range spells. With saving throws in 5e, such rules are no longer necessary.

  • \$\begingroup\$ +1 It's much easier to modify your own rolls than to modify someone else's. Other examples are Bardic Inspiration and the Lucky feat. Contrast that with the clumsiness of the Shield spell. \$\endgroup\$ Nov 29, 2016 at 7:41
  • \$\begingroup\$ Another point is information - NPCs might have roll modificators, that players aren't aware of. \$\endgroup\$
    – enkryptor
    Nov 29, 2016 at 13:44
  • \$\begingroup\$ Another minor point of difference might be the 5% change in the chance of success due to ties. The person who rolled the die wins in the case of a tie vs AC or a saving throw DC. \$\endgroup\$ Nov 29, 2016 at 17:35
  • \$\begingroup\$ A good answer that addresses the specifics of D&D 5E rather than just the generic difference between who gets to roll. \$\endgroup\$
    – Daron
    Dec 1, 2016 at 15:03

The difference between these two systems is who is given apparent control over the outcome of the random roll, and nothing more. This has more to do with psychology and what 'feels right' to your players than it has to do with any real mechanical difference.

If you were to, say, target a player with a spell that causes instant death and you make a random roll to determine the success, and succeed, a player would feel like they 'couldn't do anything' to avoid it. They died, and no event under their control could prevent it.

In the case of a defense roll, a player gets to throw the die, and that comes with a greater feeling of control, even though it isn't really different from the other system in terms of probability.

Where there may be a practical difference, though, would be in GM fudge factor where some GMs will alter the results of a random roll to achieve some end. Perhaps a player has cast their death spell on your big bad and they've surely killed him with a nat 20 and your story comes to an abrupt, unsatisfying end. Had you rolled to save behind the GM screen and gotten a 1, you could happily declare your big bad made the save and carry on as normal. Defense rolls give you the option, as a game master, to spare your NPCs from story destroying results, if you wish.

In short, any roll the GM makes is under their complete control. The GM may choose to use the random result or not. Any roll a player makes is under their apparent control, even if that is not actually so because the die roll they are making is random.

Mechanically, the difference is who gets a flat added to their stats vs. who rolls the die. In 3rd Edition the DC of a save is:

10 + Spell Level + Ability Modifier

The Saving throw itself is calculated as:

d20 + Class Modifiers + Ability Modifier

The d20 is opposed by the +10. To convert between attacker and defender making the roll, you swap these two values in the calculations.

  • 3
    \$\begingroup\$ GM fudging is more often used to save player characters than NPCs, I think. \$\endgroup\$
    – Erik
    Nov 26, 2016 at 8:41
  • 1
    \$\begingroup\$ The purpose of the example is to illustrate why it matters who rolls the die. The frequency distribution of the types of rolls a GM may fudge and in favor of whom is a subject for a different question. \$\endgroup\$
    – Dale
    Nov 26, 2016 at 9:04
  • 1
    \$\begingroup\$ This completely ignores the very important fact that some effects/abilities/*proficiencies* work differently on attack rolls than on saving throws. Saying the only difference is who rolls the die is just incorrect. \$\endgroup\$
    – nitsua60
    Nov 26, 2016 at 13:24
  • 1
    \$\begingroup\$ @Dale: Your answer goes a good way to address the psychological meaning of attack vs/ save rolls. But it does not refer specifically to D&D 5E or take into account, as nitsua60 points out, the precise mechanics of how those rolls operate in that game. \$\endgroup\$
    – Daron
    Nov 26, 2016 at 13:34
  • 2
    \$\begingroup\$ I am asking about the difference between saving 5E saving throws and attack rolls. \$\endgroup\$
    – Daron
    Nov 27, 2016 at 8:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .