For example:

A player (speed 6) moves two squares towards a human bandit (pg. 162 MM) becoming adjacent to him, then proceeds to move around him to the right, provoking an opportunity attack. The bandit hits with his basic melee attack (+4 vs. AC; 1d8+1 damage, and the human bandit shifts 1 square), shifting one square to the left immediately.

This free shift makes it so that the PC would no longer flank the bandit at the END of his move. Can the player adjust his remaining three squares of movement so that he can once again flank the now-shifted bandit, or does he have to stick with his earlier movement designation?

This question boils down to: Is movement counted as you go, or do you say, "I am moving to here, it is within my range, and on the way to my destination 'blah' happens?"


2 Answers 2


The rules (PHB, page 291) say this about opportunity attacks:

An opportunity action takes place before the target finishes its action. After the opportunity attack, the creature resumes its action.

So, based on this, we know that an opportunity action can happen in the middle of another action.

That leaves us with the two possibilities you bring up:

  1. The movement action consists of the player plotting a course, which then proceeds as programmed regardless of what happens between the start and end.

  2. The movement occurs one square at a time, and can react to changes in the battlefield.

Unfortunately, I can't find a page citation to answer which of these two interpretations is correct. My groups have always used option two, as it's a lot easier (no need to remember the exact path), and has a lot fewer ugly edge cases (like "what happens if my path becomes blocked?")

  • 2
    \$\begingroup\$ +1 - This is something that falls under what we call "DM's Discretion". \$\endgroup\$
    – Iszi
    Dec 15, 2010 at 22:09
  • 16
    \$\begingroup\$ Option 2 is the only thing that makes sense. This is D&D, not RoboRally. \$\endgroup\$
    – okeefe
    Dec 15, 2010 at 22:31
  • \$\begingroup\$ This came up in my first D&D 4e game my D&D group tried. My DM Decided that this falls under the benefit that the Human Bandit gets from his basic attack. Since the bandit's slide doesn't interfere or stop the movement of the player then the player's movement remains unchanged (aka resumes the Exact action he was doing before the OA) Now, we went on with the game but I asked him about it because I felt the player could change his move any time during his move as long as he doesn't "Take back" a movement. So he prompted me to find more info and so I ask:) Ty all and I await more answers. \$\endgroup\$
    – Gav
    Dec 16, 2010 at 1:27
  • 2
    \$\begingroup\$ @Gav Another question for your DM is "What would happen if the bandit's attack had a push or pull, instead of a shift?" would the player's turn end as soon as they were forced to move, because they were no longer on their pre-defined path? \$\endgroup\$
    – AceCalhoon
    Dec 17, 2010 at 15:02
  • 2
    \$\begingroup\$ Must concur with okeefe; regardless of rules, option 2 is the only one that doesn't make my brain scream from a roleplay or game design perspective. Otherwise we have to rule that characters don't react to events during a combat round... which is not consistent with the whole concept of opportunity attacks, or interrupt triggers. (For a real-world example, take it from a not-particularly-good fencer... the time it takes for a trained warrior to react to an opponent's steps and change their own movement is well under a second. Or you die.) \$\endgroup\$
    – Tynam
    Jan 9, 2011 at 16:35


The best answer is by inference.

We start with Walk:

Movement: Move a number of squares up to your speed.

Provoke Opportunity Attacks: If you leave a square adjacent to an enemy, that enemy can make an opportunity attack against you.

And continue with an Opportunity Attack:

Interrupts Target’s Action: An opportunity action takes place before the target finishes its action. After the opportunity attack, the creature resumes its action. If the target is reduced to 0 hit points or fewer by the opportunity attack, it can’t finish its action because it’s dead or dying.

The persuasive inference, however, comes from the Fighter's class feature, Combat Superiority:

You gain a bonus to opportunity attacks equal to your Wisdom modifier. An enemy struck by your opportunity attack stops moving, if a move provoked the attack. If it still has actions remaining, it can use them to resume moving.

From the "interrupt's target action" of the opportunity attack and the explicit note that Combat superiority, as a class feature, stops movement, the specific before general rule suggests that anything that does not so have that text does not inhibit movement.

Therefore, as forced movement can happen in the middle of voluntary movement, DMs should encourage players to note the exact path they take across the battlefield. While most of the time this won't matter much (as a safe path exists), the presence of OAs and hidden obstacles suggests this social rule.

-- Edit --

Clarification: looking at the text of move suggests that all movement proceeds square by square, rather than as any kind of path. This suggestion is vague, however:

Whenever a creature, an object, or an effect leaves a square to enter another, it is moving, whether that move is done willingly or is forced. This means shifting, teleporting, and being pushed are all moves, for example.

Thus movement is the act of leaving one square to enter another. As walking may only choose adjacent destination squares (unlike teleportation) all moves must proceed square by square, rather than as some sort of (as @ace puts it in one of his examples) preprogramed route.

Happily, this conception of square-to-square movement also makes sense. Because, in the real world, people can choose to not move somewhere in the middle of the movement to that place.

  • 1
    \$\begingroup\$ The fighter's text doesn't really help here, because an AoO without associated movement won't stop movement in either model without the specific reference. As to the last paragraph, the difference between model one and two isn't "I move to this square, six squares away" vs. "I move along this path," it's "I move along this exact path" vs. "I move up, then left, then up again (i.e. one square at a time, making decisions based on the results of that movement)." \$\endgroup\$
    – AceCalhoon
    Dec 16, 2010 at 14:19
  • \$\begingroup\$ Very interesting but not quite answering the question Brian. It wasn't whether we could have forced movement during voluntary movement but instead whe... what Ace said. :) (Ty though, it was an interesting read.) \$\endgroup\$
    – Gav
    Dec 17, 2010 at 5:54
  • 2
    \$\begingroup\$ Re: edit. Don't get me wrong, I totally agree with you on how it should be played. I'm just perplexed at the lack of rules clarity here. I'd hoped to use rules stating that you perform powers one clause at a time as a guide, but I can't even find that in the rules. \$\endgroup\$
    – AceCalhoon
    Dec 17, 2010 at 15:04

You must log in to answer this question.

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