review and analysis of current questworlds srd

[radmonger]

part 1: basic mechanics

The fundamental mechanism driving questworlds is the opposed roll, called a 'contest'. The pc rolls d20 and tries to get under their skill, the gm rolls for a resistance which can be the skill of an npc.

A roll equal to the skill/resistance (so a 5% chance) is what other most systems call a critical, and is worth 2 regular successes.

Masteries (i.e. skill over 20 add extra successes.

In the simplest form of contest, victory is determined by whoever has the most successes; ties are broken by the higher dice roll. 

This produces a table of outcome probabilities:

 

 

The above table shows the probability of a skill of n (row number) immediately winning in a simple contest against a resistance given by the column number. Equal skill and resistance is highlighted in green. It is not .50 as there is a still a 5% chance of a tie for equal skill; when the same number is rolled on both dice.

Of note:

- an equal difference in skill always gives the same chance of winning, whether or not the difference crosses the master boundary.

- the band where there is a 20 to 80 % chance of a win, highlighted in yellow, is perhaps surprisingly narrow.; a modifier of +/-10 will usually move you outside it.

- you are guaranteed a win once you exceed the target number by two full masteries, as they need a critical to match your worst roll.

 

 

 

 

 

 

[radmonger]

part 2. tie breaks

The tie break mechanic is built into the description of the basic contest, although it is not actually used by some of the contest varieties. I find this perhaps a bit confusing; I think it really logically belongs in section 4.3, _Multiple Contestants, One Prize. 

Certainly in both reality and fiction, let alone sport, it is pretty normal for things to end inconclusively for now. In act 1, the villain shows up, fights with the heroes for a bit, and then departs unscathed. This works much more naturally if the dice say 'it was a draw' rather than the gm having to say 'you win, but nothing happens'.

Without any tie break rule, the chances of getting more success than the resistance are;

 

This is a lot less regular, with the effects of mastery boundaries evident.

 

 

Edited March 30, 2023 by radmonger

[radmonger]

part 3:contests

section 4 covers contests. i think the interpretation of this section is that contests come in 3 flavors;

4.1. [simple] contest: one player makes a single opposed roll, narrate results and consequences, using the tiebreak rule

4.2 group contest. Each player gets to make one roll against a single resistance. The number of successes are added together, outcome is based on the net success/failures as for a simple contest. The tie break rule is not used, presumably as it is not clear which player's  die roll would be used.

4.3 single-prize contest. Each contestant gets to make one roll against a single resistance. The number of successes are added together, outcome is based on the net successes/failures as for a simple contest. The tie break rule is used to pick a single winner.

Note that for any kind of contest, there are never multiple (mechanically represented) opponents, just players making opposed rolls against a single resistance. So if multiple players search for a team of assassins, some visually, some by listening, some using a magic 'sense assassin' ability, they all face the same base resistance. However, different bonuses and penalties may apply, as per s2.5. 

Key to this approach, I suspect, is to follow the advice to work backwards from the desired resistance to the situation that justifies that resistance. if 1 master assassins and 5 apprentices sounds right for a resistance of say 3m1, then that is how many there are. Defeating that resistance means finding all the assassins; getting more than a minimal success means capturing an appropriate number. before you roll, it is never mechanically stated whether you are facing the master or an apprentice. That is only determined afterwards. Facing the master is a good narrative justification for bad PC rolls. Capturing or killing him a good justification for an exceptional outcome.

Not stated in this section, but i think implicit in the section 2.6 rules, is that in a group contest a player may choose to not roll themselves, and instead augment someone else's roll. Of course, this hits the issue that even 3 pcs giving the minimum +5 each virtually assures victory against any resistance a player would normally have a chance against. section 2.6 acknowledges this, but says the solution is to use a group contest, which is where we are. So maybe there should just be a limit of 1 augment per player from any source? Or the base for augmenting should be lower than 5, so several people can do it without leaving  the narrow yellow zone of the table above. Or maybe it is can just be handled by imposing situational modifiers on anyone doing things that are mechanically legal, but implausible in fiction for more than one person to do.

For a single-prize contest, which typically represents something like a race, advice is given on what to do if the 'race wins'. I'm pretty sure that situation can be avoided by setting the resistance sufficiently low that it is deeply implausible that it wins. If the unlikely does happen, then it's ok to narrate an unlikely outcome. So if a resistance of 5 beats 8 runners with skills of 17 or more, then going by the tables above, that is something like a 0.000001% chance. Which is a number that corresponds to at least freak weather conditions, and maybe a broo attack or volcano erupting.

Lacking the tie break rule, a group contest with an even number of players will produce a high number of ties: nearly 50% for two PCs.

 

 

 

 

 

 

 

[BruceMri]

Very much appreciating all this. Neurological side effects of systemic health problems have done mean, unkind things to my number crunching abilities. But I can follow your trail. 

[Natter]

Curious to see how this compares to HQ2e probabilities, I don't think much has changed, but I know there is at least a tiny bit of one with the lack of a fumble.

[Ian Cooper]

On 3/30/2023 at 1:08 AM, radmonger said:

part 2. tie breaks

The tie break mechanic is built into the description of the basic contest, although it is not actually used by some of the contest varieties. I find this perhaps a bit confusing; I think it really logically belongs in section 4.3, _Multiple Contestants, One Prize. 

Certainly in both reality and fiction, let alone sport, it is pretty normal for things to end inconclusively for now. In act 1, the villain shows up, fights with the heroes for a bit, and then departs unscathed. This works much more naturally if the dice say 'it was a draw' rather than the gm having to say 'you win, but nothing happens'.

Without any tie break rule, the chances of getting more success than the resistance are;

 

This is a lot less regular, with the effects of mastery boundaries evident.

 

 

So a tie usually means - the contest is inconclusive; it's only a PC victory if the GM can't decide on an interesting story reason why the contest ends in a stalemate.

[radmonger]

part 4: sequences

And now we hit what I would consider the problematic part of the rules.

As i read this section, a sequence is a set of opposed rolls that, instead of stopping immediately, continues until some condition is met.

 

There are 4 different variants of this, representing things that have been tried by different editions and spin-offs.

5.1.2 group sequence; pair up until all of one side defeated. The heading level of this version is a bit ewierd, perhaps it is supposed to be a default with the others variants?

5.2 scored; first side to 5 RP wins (RP = net sucesses + 1, or 1 if tiebreak rule applied)
5.3 wagered: set AP based on skill, win lose lose or transfer an multiple of a stake set based on how significant the roll is to the contest.
5.4 chained: classify opponents as mooks/named/PC, they lose individually when their resolve hits zero, where resolve is the negative of RP.

 

 

 

Edited April 18, 2023 by radmonger

[radmonger]

For sequences, then, using the tiebreak rule assuming the same skill is used each time, you get the following table:

 

 

the outcome is a little irregular because of the varying relative chance of a critical. bu the main takeway is that, due to the law of large numbers, the output is all but deterministic. A 3 point advantage in skill can give a 85% or more chance of winning, a 7 point disadvantage is less than 5% chance of a win.

this seems problematic to me, as it will taker a _lot_ of dice rolls (minimum 10) to get to a result that is almost always pre-determined. what's more, the difference between this and a simple contest result means players are incentivised to request this method in boring cases where they have a skill advantage, and avoid it in climactic scenes where they are outnumbered.

You can avoid this by using different skills, but it seems hard to make that an interesting process when, mathematically, the best solution is clear.

 

 

Edited April 18, 2023 by radmonger