Monday 21 May 2018

Surprised and Confused

In Conclusion

This turned into a long look at how surprise has worked through the early editions in order to try to fix the mess that later AD&D surprise became, so I decided to put the conclusion first so it’s easy to refer back to.

So what’s the fix? Quite simply, to go back to the original idea of the surprise die indicating both surprise and range, with “complete surprise” being the case where melee can start without any closing of range. Literally “in your face” in the case of the catoblepas and some other monsters.

I don’t want to simply go back to OD&D, however. Many players feel that multiple surprise segments is a core feature of AD&D and so I want to retain that. But I do want to get away from the idea that anything can cause an adventuring PC to be unable to react for half a minute or more.

The proposal is:

  1. Each party rolls for surprise (usually on 1d6).
    1. If surprise exists, individuals modify this rolled “surprise number”:
      1. Unencumbered figures may apply any reaction bonus.
      2. All figures apply any reaction penalties.
      3. Figures with heavy gear get +1 to their surprise number.
      4. Figures who are encumbered get +2 to their surprise number.
    2. A modified score of 0 or less is un-surprised.
    3. A modified score of 1 is surprise.
    4. A modified score of 2 or more is complete surprise.
    5. If no one is surprised after these modifications, then it is a standard encounter at normal ranges.
  2. Surprised creatures are “frozen” for one segment.
  3. Completely surprised creatures are “frozen” for two segments. If one is holding a crossbow of speed, it fires in the second segment.
  4. The less surprised (un- Vs surprised or completely; surprised Vs completely) figures have automatic initiative on the first immediately following combat round. If both sides are equally surprised then there is no initiative advantage.
  5. If all the members of a party have the same score then that modified score is used to determine the party level of surprise and encounter range. Otherwise, range is determined based on the party die roll (and maximum available space, of course):

    Party A Party B Effect Range
    Unsurprised Unsurprised None, proceed to 1st round. Normal
    Unsurprised Surprised B immobile for 1seg; A gains initiative 3”
    Unsurprised Completely B immobile for 2seg A gains initiative 1-3”, at A’s choice
    Surprised Surprised 1seg wasted 2”
    Surprised Completely A: 1seg wasted; B: 2segs 1”
    Completely Completely 2seg wasted 1”

If any opposing figures are not surprised then any “frozen” segments are part of any combat round which is engaged in by those unsurprised figures.

Notice that this already allows for the possibility that a surprised party will receive three sets of attacks before being able to answer back, so I don’t think there’s any need for a “3 segment cap”.

Note roundup

These notes are in priority order where there is a clash - so #3 can not override #1, for example.

  1. Monks above second level can never be completely surprised by anything.
  2. Piercers only ever surprise, they never completely surprise no matter what penalties the target has. This is more to do with their need for a ranged attack than the level of shock they create.
  3. Aerial Servants which are not detected always gain complete surprise without a roll; otherwise normal rolling is done based on a chance of 1-4 on d6.
  4. Quiet things grant +1 to the chance of surprising others. Quiet things include undead, bugbears, hobbits or elves in non-metalic armour, rangers, golems, cats of any kind. Some of these are already noted in the books, some are not.
    • A party only gains these bonuses if all the members qualify. There’s no use a ranger and a hobbit sneaking about if they have a donkey tapdancing on the stone floor.
  5. Modifiers to the chance to surprise others are cumulative only insofar that they reflect different advantages. Being silent is one advantage; being hidden is another. Hiding in shadows while invisible is a single advantage. As a specific example, an elven thief hiding in the shadows of a forest surprises on 1-4, not 1-5. In effect the hiding in shadows is assumed as part of the elven ability to blend into woods. Monsters appear to have slightly different rules, but PCs generally can only get +2 to their chance to surprise.
  6. A party with any members with a reduction in chance to be surprised gains the best such reduction.
  7. Surprise reduced from 1 in 6 goes to 1 in 8, then 1 in 10, 1 in 12, 1 in 20. In the opposite direction, 5 in 6 goes to 7 in 8, to 9 in 10 etc.
  8. When encountering settlements, and fortresses use the standard rules on DMG p182-183.
  9. Wilderness encounters with surprise are as per general encounters (1-3”).
  10. Wilderness encounters without surprise are as per DMG p49.
  11. Ruins are spotted at 1d10 half miles if no surprise; and 1-3” if the party is surprised. This score is shared by any monsters within.

The Problem

On the surface of it, surprise is a very straight-forward mechanism in AD&D: when any encounter happens each side rolls a d6; on a 1 or 2 they (not the opponent) are surprised and spend that many segments in a state of basic helplessness as they fumble or drop weapons or what have you.

Characters with low dexterity are not surprised more often, but instead take longer to recover if they are; high dexterity is the same, not reducing the incidence but improving recovery time. There is a slight asymmetry in that high dexterity can eliminate recovery time, so in a practical sense the high dex character is actually surprised less often.

This basic system is then absolutely screwed up even at this early stage by the sort of over-explanation that melee weapons Vs spells suffers from. Because, the above is all you really need to know and it is a simple and obvious consequence of these rules that if a PC walks around a corner and meets an orc and both roll 2 for surprise then the net effect is that both are helpless for 12 seconds but unable to exploit their opponent’s state, while being surprised for 1 segment against an opponent spending 2 segments fumbling gives you one segment to act.

Sadly, it was decided to spell this out in gruesome detail which (aside from a misprint in early printings, which didn’t help) gave the impression that the DM was supposed to subtract actual time spans from each other, leaving a muddle about what happens when two parties are surprised but one or more extreme-dexterity individuals do not share the result of their companions.

So: the first rule of surprise club is that surprise is measured in segments and you never, ever subtract one side or individual’s surprise from another. If I wait for ten minutes and you wait for 15, that doesn’t somehow mean that you only have to wait for five minutes.

But, sadly, things did not rest even there. That 1-2 on d6 has some exceptions even in the days of MM/PHB/DMG AD&D. For one, elves surprise on 1-4 on d6 when in natural surroundings. For another, rangers are only surprised on a 1 on d6.

Now, the elves are a problem. Because each segment of surprise allows an unsurprised opponent a full rounds’ worth of attacks. And four rounds of attacks is deadly. Those elves could be 7th level fighters, so they get 6 attacks in that time. Given that this is all down to basically pure luck, that’s a very big advantage, and if you have a dexterity penalty then leaving the house is basically suicide.

You're surprised for...4 days!

It gets worse. Later monsters introduced different dice, and eventually we have the example of the greenhag from MMII who can attack from invisibility, gaining surprise on 1-19 on d20. Strictly speaking, that’s a potential 19 rounds’ worth of attacks from a 9HD monster with 2 attacks per round - against AC -5 she’ll do an average of 71¼ damage!

Additional complications include what to do about a ranger’s implied -1 to surprise when faced by such a monster. Do we say that the greenhag surprises a ranger on 1-18? Big deal! Note that the DMG specifically rules out the alternative suggestion made by some that we treat the ranger’s bonus as a 50% modifier, which would reduce the greenhag to surprising on 1-10 (and only doing an average of 136hp damage to an unarmoured opponent).

There are various monsters which surprise on x on d8 and d10 floating around too.

What a mess.

How it was in the old days

So, how did we get here? That’s a multi-stage process.


The very first published version of D&D included surprise. In book three (The Underworld & Wilderness Adventures) surprise is introduced as a roll of 1 or 2 on a d6 and it states that the existence of surprise indicates that encounter distance is between 10 and 30 feet.

The primary effect of surprise is to grant “a free move segment, whether to flee, cast a spell, or engage in combat” (my bold).

In addition, the surprise roll is referred to as the “surprise distance determination die” with a roll of 2 indicating a distance of 10 feet.

At this point I’m going to get controversial. The normal belief is that melee range in D&D is 3”, because that’s what it was in Chainmail. However, Book III of D&D contradicts this once in plain language and once more by implication, in both cases when dealing with surprise. The first is when describing surprise distance in the wilderness: the statement is made that a range of 10 yards (1”) will enable a monster to attack. Under the 3” rule, this would be a redundant statement since the surprise range is between 10 and 30 feet - all of which would be in Chainmail’s melee range. The second case is in the primary discussion of surprise in the underworld where surprise is described as being an opportunity to close distance and also the encounter example which continues from determining a range of 10’ (1”) to the wyvern attacking as if there were a connection. I believe that these passages were written with a 1” melee range, the same as AD&D, in mind. This becomes more important later.

In addition to granting one “free move” having surprise also grants initiative on the first round of combat (a rule that is referenced in DMG without any real explanation).

There are no rules given for both sides being surprised but it’s not hard to imagine that the expectation was that the DM would count this as a cancelling out. Since there was only one segment of action possible, there was no real need to go into any detail.


Greyhawk introduced nothing new except for the bugbear, who surprises parties 16⅔% more often (in one sense, in another 50% more often). In other words, they surprise on a 1-3.

Notice that this has no real problematic aspects in the rules as stated so far. There’s still only two questions being answered by the surprise roll:

  1. Does either side get a surprise bonus move?
  2. If so, are they in melee range to attack?

As before, if both sides are surprised then there’s just a general fumbling about and then combat proceeds as usual. Some surprise is still all surprise and we can probably assume that a roll of 3 from the bugbear means that it’s in melee range, just as a 2 did for the wyvern. The binary nature of both surprise and melee range means there are no questions about being very surprised compared to only slightly surprised.

Although thieves are introduced in Greyhawk, the only mechanical application of their ability to move silently is to allow them to make a backstab attack.

The guy at the bottom is still not
completely surprised


Blackmoor does very little about surprise except for one paragraph in the introductory material dealing with monks, which are a new class based on the Destroyer books of Warren Murphy and Richard Sapir.

  • Third-level monks, it says, are surprised only on a 1 on d6.
  • Fifth-level monks are surprised on a 1 in 8.
  • Seventh and higher level monks are surprised on a 1 in 10.

It seems from this that a 3rd or higher level monk can not find themselves in a surprise situation where their opponent is in melee range unless they have the advantage (i.e., that is indicated by the opponent’s die). This situation is still not a complication, however because both parties are still surprised equally, in terms of time.

The paragraph (still dealing specifically with monks) goes on to note that “extremely silent creatures will double surprise possibilities, i.e., halflings, thieves, bugbears, and undead”.

In the context of a monk who has a 1 in x chance of being surprised, doubling the chance is the same as saying that the monster increases the chance by +1 - so 1-2 on d6 or 1-2 on d8 etc. Which fits what we were told in Greyhawk about bugbears but is innovative here, as far as I can see, in extending this to thieves and so forth.

Nothing in the text implies that this +1 to surprise is extended to situations not involving monks, but I think that’s logical in the same way that I assume that the thieves in question must have rolled a successful move silently check.

Eldritch Wizardry

EW took a look at the whole question of initiative which had been studiously ignored in D&D up to this point, with DMs assumed to have cobbled together something based on Chainmail’s Man-to-Man Combat section. As well as being incredibly opaque it introduced many key concepts such as segments and (implied) spell casting time and the surprise die as a number of rounds of free attacks. It may also hold the seed of a mistake.

EW tried to create a individual clockwork initiative system whereby the round was divided into segments and each combatant would be assigned an attack rhythm within that round which would essentially repeat from one round the next, at least as long as they fought the same opponent with the same means.

On the initial round of combat, surprise was used to modify this system and the notes on the tables involved included the statement that “complete surprise is basically a die of 2 when checking” (EW p6).

So you're saying...what exactly?

To throw another spanner into the works there is also a note that says that “surprise is basically a die 1 when checking, or a die 2 in those cases where the creature in question is difficult to surprise and has only a 1 in 6 chance of being surprised” (my ephasis).

The implication of this last statement is that in the EW system the chance of being surprised is 2 in 6 for everyone and the effect of being hard to surprise is just that you don’t suffer the penalties for complete surprise. It’s pretty weird if it’s not a typo.

Anyway, we are also told that “If surprised lose the 1st segment on a die roll of 1 and the 1st and 2nd on a die roll of 2”. This seems to mean that you lose your first one or two attacks, in the context. And another way of looking at that is that the opponent gets one or two rounds of attacks free of reply. Sound familiar?

And note the phrasing: the text does not say that a character that rolled 2 missed 2 attacks because he was completely surprised. This creates a dangerous link between the number on the die and the length of time a character is unable to fight back. EW never considers the possibility that a surprise die might be 3 or more - much more.

My suspicion is that Tim Kask (who wrote the EW initiative system) was simply carrying on the previous tradition of surprise starting either outside of melee or within melee, and that was the distinction between plain surprise and complete surprise with no further consideration given to higher values.

But by not saying “If surprised lose the 1st segment and the 1st and 2nd if completely surprised” he unwittingly left the door open to 19 attacks from greenhags etc.


  • The Monster Manual

    No further tweaks were made in official supplements to normal combat before the Monster Manual came out. So, what initiative system was assumed to be in play for MM?

    It’s very hard to tell. I believe that Gygax never even considered using the EW system - he simply wasn’t that mad - and that it died a quick death just like the hit location system of Blackmoor (which, to be fair to Kask, was actually worse).

    Orcus is listed as having a tail which strikes as if 18 Dex, but this could either look back to the “adjusted dexterity” system in EW, or it could look ahead to the AD&D system where high dexterity grants bonuses to both surprise and ranged attacks (which the long tail might be seen as).

    There are several references to “complete surprise” in the MM and also in PHB and DMG. That’s an interesting clue as it suggests that even when Gygax had definitively abandoned EW’s system (which he had by PHB where he had gone back to a foundation in Chainmail with its weapon factors and first strike rules, even though he had not yet formulated the final system), he had decided to retain that division.

    In the Monster Manual and PHB we encounter monsters and characters gaining surprise on 1-4 on d6.

    Piercers have a 95% chance of surprise.

  • DMG

    With DMG we finally got the official word that surprise was the number on the die, even if it was a 4, and no explanation of what “complete surprise” actually was, despite the fact that whatever it was it gave you a bonus in various situations (or a penalty, if on the receiving end).

    As far as I can see, Gygax’s next step was to forget all about this and go back to pre-EW days, at least as far as initiative and surprise was concerned.

    There was never a single official explanation or expansion of the initiative and surprise rules in Gygax-period Dragon magazine. Nothing. The world of AD&D DMs and players was baffled by the presentation of the rules yet this was never addressed in over 100 issues of the in-house organ.

    The simple reason was that Gary was as ignorant of the DMG initiative rules as he was of the falling damage rules. He continued to play to the rules in his head, leaving the formal rules for people who wanted to play in tournaments. It’s entirely possible that Gary didn’t even write the DMG rules, just as “Gygax & Blume” had not written the system in EW. No one else has admitted responsibility, however.

    The DMG combat rules were certainly not well proof-read in any case, with dead-end cross-references and a table explaining surprise which had a hideous error in it (to say nothing of the screeds of combat examples in PHB and DMG which simply didn’t match up to the rules given).

    As Gary went on creating monsters, he retained the distinction between surprise and complete surprise in his head but nothing much else. So, even in the MMII we find several new monsters still referring to complete surprise alongside monsters that cause crazy numbers.

    Like Kask before him, Gary simply didn’t think about the consequences of making that link between the number on the die and the number of segments of free attacks and largely acted as if it didn’t exist. Because, I suggest, that link did not exist for him.

    So, what to do about all this? Go back to the top of this document to see.


  1. Early on in the DMG, Gary explained (poorly!) the way scale inches work. In his example, he chose to use 3 one-inch squares on the table equaling 10 feet in the game world. I think this probably came from the exact same part of his memory as the 3" melee range in the Man-To-Man section of Chainmail.

    Or to put it another way, you're right - the melee range is about 10 feet. At some point Gary was playing Chainmail Man-To-Man rules with a one inch on table = 1 yard in game world scale.

    Begging the question, how long was his yardstick? :)

  2. Very interesting write up worth a re-read or two. Small point of order. For OD&D, you've conflated melee range with melee distance. It's easily done, so no fault of yours. Melee range, per CHAINMAIL, is the distance within which opponents fight - kinda like a boxing ring around the combatants. A figure can join the fight by entering the ring, so to speak, and you must leave the ring to leave a fight. Melee distance, aka Zone of Control or Area of Control is the actual fighting space of each figure, and that is 1". A ghoul can touch you when you are within 1", for example, but not when you are 2" away. Note also that Melee range doesn't exist at the moment of surprise - because melee hasn't started yet.

  3. Gygax realised that keeping his probability resolutions linear helped with the abstraction he understood was an important aspect of gaming.

    But with surprise he makes a mistake in extreme monster cases. I wonder if a useful step would be to use a d12 for surprise and force fit the outlying monsters to that to retain linearity.

    There is a crucial decision to be made about decoupling probability of surprise from length of surprise. My instinct is the same as yours to try to keep the concept as close to 1 in 6 standard , and 2 in 6 very rare.

  4. Around these parts, any DM who required a character to attacked 19 times due to "surprise" would be drawn and quartered. There is a thing called Common Sense.