HeroQuest rules revisited - A simplified mechanism

[Gelasma]

I propose to simplify the HeroQuest rules, to reduce the handle time of contests. The original rules require multiple comparisons of actual rolls, trait value and masterships to get first a level of success for each roll and finally comparing these to get a level of successs for the whole contest. One can acchieve this with a far more simple mechanism.

The proposal is to use the mechanism known from the D20 games, that is just compare the the sum of actual roll and trait value. The higher values wins the contest, and depending on the difference of the results we can define levels of successs. For example for each ten points of difference a level of successs. /footnote{To get a probability distribution that is more similar to the original rules, we count a rolled 1 as -10 and a rolled 20 as 30.}

Example - Character A has 'Foo Fighting 17', Character B has 'Bar Fighting 25' (5M in HeroQuest notation). Both players roll and add their trait values to the result. In this example player A rolls an 8 and player B an 14, the sums are 17 + 8 = 25 and 25 + 14 = 39 and thus player B wins the contest - with a minor success, as the difference is 14.

Note - how to handle transfers in extended contest is still kind of an open issue. But as we use a system based on d6 instead of d20 we just definied these levels of success for extend contests:

[*]difference 1 - loser loses 1/2 x AP bid.[*]difference 2 - loser loses 1/2 x AP bid, winner gains 1/2 x AP bid.[*]difference 3 - loser loses 1 x AP bid, winner gains 1/2 x AP bid.[*]difference 5 - loser loses 1 x AP bid, winner gains 1 x AP bid.[*]difference 5 - loser loses 2 x AP bid, winner gains 1 x AP bid.[*]difference 6 or more - loser loses 2 x AP bid, winner gains 2 x AP bid.[/list:u]Using a d6 instead of a d20 ist a further simplification of the rules, as there is no need for a granularity of twenty values if in the end there are just four different levels of success (or in the case of your d6-rules even just three levels.)

Why d6 instead of d20? - Comparing two d20s gives 400 different possible results, breaking them down to four levels of success (acctually nine, as there is also a draw and four levels of failure) maps 44.4 results per level of success. This is a waste of precision. On the other hand comparing two d6s gives 36 possible results, breaking them down to seven levels maps 5.1 results per level of success, which is more than accurate enougth.

[lev_lafayette]

Of course, if you're just using d6, you're making a fairly radical change to the distinction betwene Karma and Fortune. At least with a d20 system they are (assuming a nominal ability scale of 1-20) roughly equal. With a d6 that equality disappears. Which isn't necessarily a bad thing in a lot of situations.

Luck should only play a role when luck is appropriate (to the sim or to the nar, or for fun)

[Mike Holmes]

Yeah, I think that the HQ system produces a lot of effects in it's curve, the benefits of which aren't always easy to see. But they're there. And this system doesn't come close to replicating them.

Actually, I have come up with a simplification that doesn't quite cut down on things as much, but still helps, especially in terms of not needing the chart. Here's what you do. Instead of the crit fail, fail, Success, crit success results, just output 0,1,2,3. That is:

roll = outcome
1 = 3
2 to skill = 2
skill to 19 = 1
20 = 0

Add one for each mastery and one if you bump (can be done later). The higher result is the winner. Check the difference to see the result.

0 = low roller gets marginal victory (loser loses half bid)
1 = minor victory (loser loses bid)
2 = major victory (loser tranfers bid)
3+ = complete victory (loser transfers twice bid)

This actually replicates the current system almost exactly. But you never need a chart (you can memorize the above four results easily), and there's never the worry about can I bump etc. Note that this is slightly different than normal system in that you can essentially bump a critical success. But it seems to me that the option to do so was kept out as a vestigal part of the original system that had a need for unopposed rolls. With all opposed rolls you're going to go off the chart anyhow, so there's no need to worry.

The other variation is that there are less possible results in extended contests. But I don't think that's a big deal anyhow.

What's really good about this is that you don't have to wrap your head around, "I got a success, but my opponent got a critical success, so I get a minor defeat." Given that the initial roll doesn't determine success, calling it success or failure is just confusing.

Mike

[Gelasma]

Yeah, I think that the HQ system produces a lot of effects in it's curve, the benefits of which aren't always easy to see. But they're there. And this system doesn't come close to replicating them.

I'm not sure if all of the strange effects of the HQ system are a desired benefit - for example if your opponent is two masteries above you, the chance for a failure is nearly the same, regardless wether the remaining value used to roll against are 5 versus 5 or 5 versus 15. But in one case mostly over 80% of your failures will be major failures, while in the other over 80% of your failures will be minor failures. So if you are in a contest against a much stronger opponent it is wiser not to bump your value, since this would significantly  ncrease the degree of your failuers while the chance of not failing at all does only marginaly change. If the remainders are near the mid-value of 10 the system works well, but if they are near the end of the d20-range the mechanism often yields strange results...

Of course, if you're just using d6, you're making a fairly radical change to the distinction betwene Karma and Fortune. At least with a d20 system they are (assuming a nominal ability scale of 1-20) roughly equal. With a d6 that equality disappears. Which isn't necessarily a bad thing in a lot of situations.

Of course we also adjust the ability scale to reflect the change of dice-size, thus both effects neutralize each other and the only thing that changes is the granularity. Which is in my opinion a good thing, as I see no reason to roll d20 x d20 = 400 if in end the we only get a result with a resolution of 9 levels.

Furthermore humans are not able to meaningfuly distinguish twenty possible ability values, while we can distinguish six values in a meaningful way. Just try to find twenty adjectives describing the level of your ability - you wont! - while you can easly find five to seven such adjectives. Which is also the reason why HQ, in the end, breaks down the rolls into only nine levels of success, since we can name them with "minimal, minor, major, total". But then: Why this detour with those fine-grained ability value? Why not just use through the whole system a granularity that humans can grasp?

[Mike Holmes]

But in one case mostly over 80% of your failures will be major failures, while in the other over 80% of your failures will be minor failures. So if you are in a contest against a much stronger opponent it is wiser not to bump your value, since this would significantly  ncrease the degree of your failuers while the chance of not failing at all does only marginaly change.

Huh? You bump after the die roll. So you always know the effect, and that's part of the decision making process. It's a feature that sometimes a bump means more in one case than in another as that means that not all situations will be the same in terms of incentives. In fact, I love this feature.

I'm not saying that the curve is perfect. But I'm not seeing any flaws here. Another real advantage of the curve is in it's minimal marginal increases in odds. That is, if two characters are anywhere near close in ability, this means that contests are a very, very unsure thing. How real life this is. It explains why people don't fight often - you just aren't often garunteed of being that much better at it than your opponent to win. Some people see this as a flaw, but it also means that the underdog gets to win a lot. Which is also pretty dramatic. It doesn't support the simmy "X does Y" sort of POV, but it's great for randomizing events for dramatic purposes. Dramatic swings are constant.

But the system still allows for someone to be so superior that their victory is assured - so you don't get people taking swings at gods on the off chance that they roll well. They're just outclassed. No other system does both of these things.

I can go on and on about the benefits of the curve and have in other threads.

Furthermore humans are not able to meaningfuly distinguish twenty possible ability values, while we can distinguish six values in a meaningful way.

Well, I don't think this is true. That is, I know that a 13 is not as good as a 14, or a 17. Check out this page at the bottom: http://random.average-bear.com/Heroquest/WhatSkillRatingsMean

Yes, there are several ranks in between each level, but these represent incremental development. Which I think is intersting. The question would be for your proposed system, how do you alter ability purchasing costs in terms of HP?

Again, I very much appreciate the granularity. One thing that bugs me about FUDGE and similar is that I love percentile granularity or thereabouts in terms of what it means for incremental advancement. In fact, the range on abilities for humans in HQ is not just 20, but more like 80. And the overall scale goes to about 240 practically speaking. The fact that such a fine description exists, and is functional across it's entire range so that you can describe everything from gnats to giants the size of planets is - well just too cool for words.

In any case, there's no more difficult math with the d20 method. You're only doing comparisons anyhow, so it's not like the larger range means anything in terms of math difficulty. Unless you're saying that adding +2 to 22 for augmenting is tougher than +2 to 5.

Just how do you handle augments?

Mike

[Gelasma]

@Augments:
Each character starts with 9 traits and the player can distribute 17 points within a range from 1 to 3. A trait with a value of 3 or more can be used to augment other traits with +1. The same with 6 and +2, and so on.

Per game session the players get 1 heropoint, they can either use it to a) increase one traits by +1, b) add one new trait at +1 a, or c) save it for a later use as "karma point", ie boosting one roll with +6.

But this is a variable not a constant: I once did a 12 hour one-shot with a very high density of narration, and thus gave each hour of play one point to express that dense narration.

@Topic:
There are different preferred styles of play, and we should avoid a discussion about which one is the better. This is a matter of personal preference. My simplifications aim at a style of play with fast and dense narration, where rule handling is reduced and improvisation made easier. I try to achieve this with two modifications:

[*]Reducing the complexity of resolution mechanism.[*]Reducing the granularity of game parameters.[/list:u]From a computation theory point of view we can measure handle time with the kolmogorov complexity. That is the size of the smallest computer programm that can execute the conflict resultion. The lower the kolmogorov complexity, the shorter the handle time. But the shorter is not necessarily the better. There is a trade-off beetween shortness of resultion and the number of game relevant factors taken into account. A good mechanism maximizes this ratio.

1) My first change, the reduction of the resultion mechanism, aims at a shorter handle time while preserving the input and output and keeping the probability distribution as close to the original one as possible. The original HQ-rules first use division-modulo to preprocess the input values and finaly use a two-step resolution of the output values. This can be simplified by switching to a simple roll-with-addition, without loosing the granularity of neither input nor output and keeping the possibilities very close to the original. If you look at the mechanism as a black box, you have the same input and output values and a similar probability distribution. I did a "least square fitting" of both curves and they are closer to each other than you might guess.

A shorter handle time is good, as it reduces the time wasted to mere mechanism. You dont lose any feature of the system when switching from HQ's division-modulo-two-step mechanism to roll-with-addition; its just the mechanism black box that changes. Thus, independent of your style of play, this change is recommendable.

Simply said: its a shorter piece of software that computes the same values (or at least, approximately the same).

2) The second change, the reduction of granularity of game parameters, affects the style of play. But this is what I want, I want a faster but more meaningful development and denser narration. This a mere matter of taste: None style of play is better than the other, just different. From your post I can see that your prefer another style, but nevertheless let me explain why I want this sytle and how this change facilitates this style.

The less the granularity of a game parameter, the more distinct the meaning of each possible value. And hence the more meaningful the difference between two consecutive values, ie the meaning of increasing a trait by +1 has much more meaning. Its not a small increment, but a change that really matters and has real impact on the narration --> the narration becomes denser.

Furthermore a less granular scale facilitates improvisation. The smaller the set of possible values, that easier for the game master to choose among them. I always judge a game systems from this game master point of view, as I dont have the time to do big-design-up-front and thus need a system that facilitates to come up with values on the fly. The descion of choosing the concret value of a trait has always the same impact on the narration, independet wheter one chooses among few or very many values. From an information theory point of view, we can measure a descision in bits - and as the game impact is an invariant, the fewer values, the higher the meaning per bit --> the game master's descisions become more meaningful, and thus the narration becomes denser.

Simply said: I bundle multiple values into one (i.e. 1..10 -> +1 11..20 -> +2 and so on) and thus the meaning per value increases.


.

[Gelasma]

PS: thanks for the link to the roleplay/heroquest wiki, its a great website.

[Christopher Weeks]

Furthermore humans are not able to meaningfuly distinguish twenty possible ability values, while we can distinguish six values in a meaningful way.

Well, I don't think this is true.

I spent about half my time while earning my masters degree in education on test theory and psychometry.  There are well-conducted studies that clearly suggest that teachers, even using mechanical systems for derivation of grades, that use a ranking with 100 stops (i.e. they assign a score of 1 to 100) do so with a remarkably low degree of reliability and validity.  Rankings with 11 stops (i.e. A,A-,B+,B,B-,C+,C,C-,D+,D,F) are much more valid and reliable instruments.  The human being appears unable to differentiate between 100 levels of achievement, but is quite capable of 11.  This obviously doesn't say anything about 20, but there is a known phenomenon being brushed by this discussion.

What I really wonder about using a d6 is the role of the crit and fumble results.  When coupled with the simpler comparison method, you can gloss over that change (maybe), but if you were d6ing HQ without that, something else would have to give.

Gelasma, if you're using both of these mods, at what steps in the scale of possible differences do you distinguish between the success/fail types?

Also, the 400 roll combinations produce a level of complexity that could be valued merely for its difficulty to predict precise comparitive outcomes.  Wouldn't that be lessened by the reduction do a d6 scale?  HQ wants you to know that higher is better and that one mastery greater = a 75% success rate, etc.  But maybe being able to more accurately assess the odds is a disadvantage?

[Peter Nordstrand]

Just for your information, here is a little something I once created for Hero Wars. I think it needs to be reworked to appropriately mimic extended contests in HeroQuest.

http://www.geocities.com/doctorpeace/bookoflies/extended.html

All the best,

[Gelasma]

What I really wonder about using a d6 is the role of the crit and fumble results. [...] Gelasma, if you're using both of these mods, at what steps in the scale of possible differences do you distinguish between the success/fail types?

There are no crits or fumbles, and the steps are each two points: difference 1 to 2 is a minor success, 3 to 4 a major succes and 5 or more is a total succes. The nice thing is, that 2.415 is the standard deviation of 2d6. Thus minors, majors and totals match the 66%-95%-99.7% rule of normal distribution.

[simon_hibbs]

The main problem with a roll-and-add mechanic is that you very easily get situations where one side or the other cannot win. If you are using skill + d6, if the difference between two character's skill levels is more than 6 points then the lesser skilled character will 'allways' lose.

This is not the case in HQ. If Character A has an ability of 5 and character B has an Ability of 45 (2W5), it is still possible for Character A to win. A rolls 1 (critical), B rolls 20 (fumble). B's roll is bumped to a success, but it's still beaten by A's critical. In HQ you need an advantage of 3 masteries (60 points) to guarantee at last a tie.

This is what makes the game so scalable to powerful characters - they can still meaninglfuly interact with ordinary characters. Most starting or near-starting PCs will have a few abilities at one level of mastery, which means they can still theoreticaly beat a character with an ability at 3 masteries, and if the PC has spare Hero Points and a few good augments, even world-class heroes with 4 masteries are potentialy beatable given aggresive and 'heroic' efforts by the players. This is an essential feature of the rules.

Simon Hibbs

[Mike Holmes]

This is an essential feature of the rules.

Quite agreed. I love the granularity that HQ provides without it being a burden in play.

Mike

[Gelasma]

Heropoints okey, they are an essential feature of HQ.
But the the example with critical and fumble is just that, the usual critical-fumble case of "1 out of 400 pure luck" you'll find in almost any game. Nothing special nor essential about this.