Wednesday, September 9, 2009

An EPT Oddity

In re-reading Empire of the Petal Throne recently, I noticed a peculiarity in its use of percentile dice. Take for example this passage on seeking permission to change one's alignment at a temple:
It is possible to change one's alignment by going to a temple of one of the Gods (not a Cohort) of the opposite alignment and seeking permission to change. This is granted upon a roll of 60 or more on two percentile dice.
This is opposite the usual practice where a referee might judge a 40% change of gaining permission from the temple, with the player succeeding in the roll if he success 40 or less on D100. This peculiar practice appears throughout the rulebook, such as on the table for acquiring bonus spells as a character gains new levels. The table only makes sense if one must roll a certain number or higher on percentile dice. Otherwise, the chance to gain a Group I spell (the least powerful type of spell) decrease from 80% to 70% once a 2nd-level character gains 3rd level.

Granted, this "higher is always better" interpretation is consistent with Professor Barker's freeform "Thursday Night Rules," but it's very different than the way percentiles have been used in nearly every other game I have ever played. Or is it? Perhaps I've missed something somewhere. Can any think of using percentiles in a fashion like this elsewhere, particularly in the early days of the hobby?

23 comments:

  1. It seems on the rarer side to me; I can't think of any games I've played that use a d100 to hit a target number instead of using it for handling "naked" X percentage probabilities in the roll-under method.

    ReplyDelete
  2. Original Rolemaster has a lot of rolls where you roll D100, add the modifiers (there are always modifiers), and if your total score is over 100 you succeed. Not quite the same situation as you found in EPT, but a similar "high is always better" concept.

    ReplyDelete
  3. You know what I'm finding weird about about EPT right now? A lack of charsheets online. I've found exactly one and it looks like it was laid out by a blind gorilla. I was honestly expecting at least one super sweet EPT sheet festooned with elaborate borders and flowery script.

    ReplyDelete
  4. Theoretically, higher-is-strictly-better systems make a lot of sense:

    (a) They're naturally open-ended upwards, and thus tend to be more scalable, especially if the initial prob curve is flat (as with any "single-die" roll).

    (b) They don't involve subtraction, and thus seem to play "faster and smoother": I seem to remember that addition and multiplication are cognitively much easier to do than subtraction and division, and so any game where you don't have to subtract or divide will seem to play faster and smoother. Smart system designers seem to extend this philosophy to the point where there are no negative modifiers in the game at all: if it benefits you you add it to your generated total, if it hinders you, the moderator adds it to the difficulty total you're trying to match/beat.

    Other higher-is-strictly better systems abound, but typically don't involve "percentile totals" because this seems to fly in the face of the old "you have a 40% chance of success", and to present it otherwise kind of front-loads subtraction (my rating is 40, so I start with a 60% chance of success?).

    For example: D20 is a "higher-is-better" system, and could pretty easily embrace the no-negative-modifier philosophy, too by just adding bad stuff to the DC, and good stuff to the acting creature's total.

    Here are the big systems I know about that use d100 generation, but in the "higher is better" way:

    - RoleMaster and all the "xLaw" derivations: HARP, MERP, CyberSpace, RM, SM, etc.

    - Aces & Eights from Kenzer does exactly this: having a lower skill number is always better, because you want to roll over the number to succeed. I'm not sure if HackMaster (Basic?) does this as well.

    There may be others that do this with d100, but I can't think of them off the top of my head.

    ReplyDelete
  5. It is odd because it uses the term "percentile" but then goes and uses the dice as if they were a d100 rather than a d%. It is interesting that we've grown to equate a hundred-sided die with a percentage-based resolution system, when it clearly doesn't have to be so.

    ReplyDelete
  6. While creating my character for your play by post game I noticed that in the Greyhawk supplement there is something similar in the little blurb about a thiefs ability to climb walls. Rather than being expressed as a percentage chance of success, it is expressed as a percentage chance of failure - 13% at first level. So you would roll d100 and if you got higher than 13 you would succeed.

    I am so used to the other way of looking at percentages that I automatically converted that to an 87% chance of success on my character sheet.

    ReplyDelete
  7. Yes, the Rolemaster is a prime example, and later, I believe the Pacesetter games (StarAce?) may have done the same.

    However, in EPT, tests to grapple an opponent in combat (section 725) are roll-under on d%, while in section 1130, a Dexterity save is made on d20::high to avoid drowning as a ship capsizes. 2000 has a fairly standard Saving Throw on d20 by level, however 2110 Divine Intervention is d% roll high as modified by score and level.

    A mixed bag.

    ReplyDelete
  8. Roll-under makes sense because you immediately know your percentage chance of success. You don't have to perform a funky calculation in your head. Furthermore, if you have a bonus you apply it directly as an addition to your chance of success. It's counter-intuitive to subtract a beneficial modifier.

    As I recall, in 2E it was generally lower = better. I recall the Psionic Handbook for Wild Talents required a very low roll. A high roll resulted in Consequences. Thief skills required a roll-under. High-percentage Strength was a good thing.

    1E was sort of a mixed bag. The gem and jewelry tables definitely rewarded a high roll. Psionic talent determination required a very high roll. You got a special familiar if you rolled high enough.
    But the Deities and Demigods book for calling for aid from your god required a low percentage roll. Same with Thief skills.

    CAPTCHA: ponfacc. n. The Vulcan period of life of being really horny but with no desire to do anything about it.

    ReplyDelete
  9. One thing I have noticed when using systems that mix both systems (aim high, aim low) with the same dice is that it drives players crazy. You can almost count on getting the twenty when you are rolling an attribute check and a one when you are rolling to hit. It is still just random probability either way, but it feels like you are getting cheated when that natural 20 comes up and it actually means you fail horribly!

    ReplyDelete
  10. This comment has been removed by the author.

    ReplyDelete
  11. The wording seems to refer to d10's as percentile dice. So my take is that it's not a percentile roll, but a d100 roll with two 'percentile dice'. Huh.

    Then again, why not just roll 1d10?

    ReplyDelete
  12. Indeed, that's why it is so funny from the GM's point of view. It also prevents players from bringing loaded dice to the table (as yes, this has happened more than once in games I ran).

    ReplyDelete
  13. Using percentile dice, higher was better on the Marvel Super Heroes' Universal Table.

    ReplyDelete
  14. >> Then again, why not just roll 1d10?

    At the time, d10s weren't available. d20s were numbered 0 - 9 in two different colors so they could double as percentile dice. Even though a d20 could be used as a d10, I don't think it was common to use them that way.

    -- James

    ReplyDelete
  15. Re: Greyhawk thieves climbing being given as a chance to fail. Here's a nifty snippet from the AD&D DMG, p. 9:

    "SLIGHTLY SLIPPERY surfaces DOUBLE chances of slipping and falling. SLIPPERY surfaces make chances of slipping and falling TEN TIMES more likely."

    Of course, I was always confused by how to compute this on the fly, because AD&D PHB tables thief climb as a chance to succeed (same as Carl does above). However, this particular rule only makes sense in the context of a chance-to-fail system. (Who knows, maybe it first appeared in a Dragon article in an OD&D context, then copy-and-pasted to the DMG).

    There's a LOT of stuff like that, confusing in the DMG, that only made sense to me after I (much, much later) got access to the OD&D LBB, to see they were really additions to the OD&D context.

    ReplyDelete
  16. Psionics in 1st edition AD&D.

    ReplyDelete
  17. the system in Fantasy Wargaming, which you mentioned in a retrospective a little while back, also uses "high is better" d100 rolls for action resolution.

    i don't think that it was an uncommon mechanic, but it is somewhat more unwieldy than the RQ/BRP-style "roll % chance or under".

    CAPTCHA: "fambion" - a biological unit which makes up a celebrity.

    ReplyDelete
  18. Timeship and Man, Myth & Magic by Yaquinto Publications both have combat rules whereby one must roll higher than one's combat skill on percentile dice in order to score a hit on an opponent. If I recall, the damage was calculated by subtracting the dice roll (if one succeeded) from the attacker's skill, and then adding the weapon's modifier. Very simple, but the low skill being better always irritated me, especially since the attribute checks were normal percentages with high attributes being better. I tended to convert everything to straightforward percentages when I GMed.

    ReplyDelete
  19. Sherpa by Steffan O'Sullivan also has a split system in which one rolls high for combat rolls and low for skill rolls, but it seems more user-friendly since a high attribute is best in both cases. In combat, you add your roll to your attribute and the highest wins. In skill checks, you simply try to roll under your attribute. In either case, it is beneficial to have a higher attribute. (The randomizer, incidently, is a digital stopwatch or a d10.)

    ReplyDelete
  20. In Runequest, you had to roll on percentile skill higher than your current skill rating in order for you to improve. Every adventure that you used your skill successfully to make a positive difference, you had a chance for you skill to improve by a few percentage points. That icluded the combat skill with a particular wepon. Say you used your tracking skill to lead the party to the Ogre lair, where the party subsequently slew the beast. The DM checks off the box near player's tracking kill. After the adventure the player rolls peercentile dice to beat his own skill rating. If he does, he rolls 1d6-2 and adds any result above zero to his skill. So, if your skill is 27% you have a faily good chance to improve the skill by 1 or 2 percentage points every time you get together to play. But say that you are a a master tracker wiht a skill of 92%. You have to roll above 92 to improve. This only applies to experience based learning. If you find a teacher with a skill better than yours, you can improve through a course of instruction.

    ReplyDelete
  21. Jeff, I don't want to make any promises, but I have long thought about an EPT char sheet, so once I get a couple of other projects out of the way, I may look into it. If you have ideas, let me know...

    Cheers,
    Patrick / The Mad Irishman

    ReplyDelete
  22. As a gamemaster I would deliberately flip-flop which end of the number line the players had to hit to succeed. I'd announce, "roll d100." They'd cringe and roll, never knowing what to "aim" for. Sometimes I'd even mutter, "20 to 29 is a success" or something, as the dice hit the table, just on the spur of the moment. Mad days.

    ReplyDelete
  23. A few years back when I was re-reading my AD&D books a lot, I got the definite impression that Gygax tended to think in terms of probability rather than dice. And that he expected the same of the reader. When specific die rolls are mentioned, they often feel more like an example than a rule.

    I imagine him saying: “High or low? 4 in 12 or 2 in 6? Who cares? Just pick one and go.”

    I get a similar impression from classic Traveller.

    ReplyDelete