WFRP 4 analysis

The enemy lurks in shadows
User avatar
Orin J.
Posts: 403
Joined: Wed Feb 06, 2019 10:39 pm

totsuzenheni wrote:
Sun Feb 17, 2019 7:38 am
Orin J. wrote:
Sat Feb 16, 2019 11:22 pm
I don't think opposed rolls are bad. but opposed rolls with a percentile di(c)e is a recipe for trouble [ ... ]
I'm curious to know what it is that you ( all ) think is a recipe for trouble about putting percentile di(c)e together with opposed rolls. I'm developing a system that does just that and, though the way in which the rolls are opposed in the system i'm developing is different to the way in which they are opposed in WFRP 4th edition, this is one of the few things about WFRP 4th edition that i think i like, at least at 'first glance'. Is it the particular combination of percentile di(c)e and opposed rolls that appears in WFRP 4th edition, or is it any such combination that you ( all ) find troublesome?
opposed rolls more or less have to be judged on a bell curve to balance them in my experiance, and roll-offs against what is effectively a single die means that you either end up discarding the advantage in using the opposed rolls (the odds are too unpredictable to judge) or relegating the purpose of the roll to an outlier (modifiers in place overriding the any meaningful chance of getting different results). this is made a lot worse by the fact that single die roll has huge variance, so there is no real predictibility in comparing odds against relitively equal opponents. 4th's system mean you either have a large margin of un/favorable modifiers, or you have a crapshoot.

they obviously went with piling on modifiers to convey the appearance of balanced combat, but it doesn't take much at all to game the system into uselessness from what i've seen.
Graak
Posts: 51
Joined: Mon Jan 14, 2019 11:50 pm

Pretty much what Orin wrote.
It's not as much a matter if personal taste as a matter of statistical facts.

You chose a dice mechanic? Fine, there are consequences for that. I'm not sure anyone in C7 was really aware of what opposed d100 rolls with DoS really meant to the game.

@Capnzapp: I almost fell into that trap, I had very high expectations about C7 ability to convey very nice mechanics with strong ties to the setting (The One Ring would be my first example too), luckily an alarm bell rang when I've heard the first rumors about the opposed DoS mechanic proposed as a fix for whiff factor
CapnZapp
Posts: 263
Joined: Tue Jan 08, 2019 2:15 am
Location: Norsca

Let's try to understand the problems inherent in opposed d100 rolls.

The main issue is that the target number is no longer static.

By that I mean that if you have 56% in some skill, and the difficulty is Average (+20), your target number is 76%. A number that doesn't change. It is static.

This lets you trust your skill level. You basically know that you have a seventy-six percent shot at success. You will pull off the task more than three out of four times. (Then you have Fortune Points, which can make this even more favourable, but I digress)

Now let's use an Opposed Test instead. Instead of saying the test is Average (+20), your skill is pitched against that of some NPC. For this purpose we'll assume the NPC has a skill of 50%. The NPC skill score that corresponds to an Average static modifier is 30%.

This means that it's equally probable your Test is going to be harder and that it is going to be easier. We simply don't know.

If the GM rolls 26 for the NPC, that means your Test effectively gets a -30% modifier (the NPC gains 3 SL). But if the GM rolls 82, that instead means your Test gets a +30% modifier (since the NPC gets -3 SL).

This means huge shifts in the modifier in ways you as a player can't control, and maybe more importantly: in ways the GM can't control either!

The difficulty becomes essentially random. This robs the player of agency (since his or her inherent skill now matters much less). A skilled character can fail, not just by rolling badly, but also because random bad luck.

Now, the important part: with the NPC having a 50% skill score, on average the SL is zero. But this doesn't matter when you only make one or two rolls.

Don't look at the statistical averages. Instead consider: how likely is it that the average result comes up. With Opposed Tests this is much less likely. The probability of an outlier is greatly magnified by rolling two d100 rolls instead of just one.

Do we want a skill score of 70% to mean "you succeed seven times out of ten"? (Yes, we do) If so, don't use Opposed Tests for isolated die rolls since that is only true when we roll ~50 for our NPC.

In at least eight out of ten times, the NPC won't roll 0 SLs. In eight out of ten times, the Opposed Test adds a random modifier that diminishes the importance of your own skill score (as well as any assigned difficulty modifiers).

You could have a 77% chance with a +20% modifier, and look forward to not just making the Test, but collecting a healthy number of SLs too, but if the NPC rolls well, all bets are off.

This is where the designers went astray, I believe. They look at average outcomes without considering the large spread.

When you make many rolls (as you do in combat) Opposed Tests work fine, and this is also why WFRP4 has solved the whiffiness of WFRP2.

But when it comes to singular rolls (common in social and exploration), Opposed Tests work really really badly, since they more or less displace player and GM agency to tell a story.
mormegil
Posts: 56
Joined: Wed Jan 09, 2019 10:48 am

How many out of combat opposed rolls are needed as per your experience Cap? I mean in which situations you would use opposed rolls?
User avatar
totsuzenheni
Posts: 286
Joined: Mon Jan 07, 2019 12:38 pm

I appreciate all your responses on the topic of putting percentile di(c)e together with opposed rolls and i see what you all mean about the opposing roll becoming a random modifier for the protagonist's roll, and a random modifier that appears to have lost some of it's meaning in gaming and simulationist terms at that. I'm not sure i would be entirely put off using the system because of that alone, though i'm not enamoured of the 4th edition system as a whole, nor many of it's other components, so it's harder to judge this one part in that context. I'm yet to do fuller statistical analysis with the system i'm designing but i've avoided this particular problem by using the opposer's ( equivalent of WFRP 4th edition's ) static skill score as the modifier. I would suggest this as possibility but i anticipate it will work in the system i'm designing because of other differences it has with WFRP systems ( such as what a minus number means ) so i don't think that would be a quick fix. Anyway, i don't want to hijack the thread so i'll leave that there.
User avatar
Orin J.
Posts: 403
Joined: Wed Feb 06, 2019 10:39 pm

CapnZapp wrote:
Mon Feb 18, 2019 1:28 am
When you make many rolls (as you do in combat) Opposed Tests work fine, and this is also why WFRP4 has solved the whiffiness of WFRP2-
that's not how 4th solved whiffing with opposed rolls. they solved whiffing by removing the ability to apply multiple layers of failure to an attack and forcing it into a single opposed roll for each attack. instead of an attack being successfull if it hits AND if it isn't parried AND if it isn't dodged, it's now either successful OR it's missed, defended or dodged somehow based on the one roll. it doesn't make the odds any better over several rolls (you could argue it makes them worse once advantage is applied) it just removes the ability to do nothing by forcing an action out of the roll.
mormegil
Posts: 56
Joined: Wed Jan 09, 2019 10:48 am

Orin J. wrote:
Mon Feb 18, 2019 11:51 am
CapnZapp wrote:
Mon Feb 18, 2019 1:28 am
When you make many rolls (as you do in combat) Opposed Tests work fine, and this is also why WFRP4 has solved the whiffiness of WFRP2-
that's not how 4th solved whiffing with opposed rolls. they solved whiffing by removing the ability to apply multiple layers of failure to an attack and forcing it into a single opposed roll for each attack. instead of an attack being successfull if it hits AND if it isn't parried AND if it isn't dodged, it's now either successful OR it's missed, defended or dodged somehow based on the one roll. it doesn't make the odds any better over several rolls (you could argue it makes them worse once advantage is applied) it just removes the ability to do nothing by forcing an action out of the roll.
I am with Cap on this, with the opposed tests there is no need to have parry and dodge. The odds are not better but who cares? When you roll you know whether you succeeded or failed based on two rolls and this is all that you need, fast and efficient. Applying advantage, or how to apply advantage is the discussion that Cap is trying to resolve for his game. From my point of view, I started with RAW, but I am considering to deduct advantage as you use it.
Bitsa
Posts: 18
Joined: Thu Jan 17, 2019 5:53 am

mormegil wrote:
Tue Feb 19, 2019 3:06 am
Orin J. wrote:
Mon Feb 18, 2019 11:51 am
CapnZapp wrote:
Mon Feb 18, 2019 1:28 am
When you make many rolls (as you do in combat) Opposed Tests work fine, and this is also why WFRP4 has solved the whiffiness of WFRP2-
that's not how 4th solved whiffing with opposed rolls. they solved whiffing by removing the ability to apply multiple layers of failure to an attack and forcing it into a single opposed roll for each attack. instead of an attack being successfull if it hits AND if it isn't parried AND if it isn't dodged, it's now either successful OR it's missed, defended or dodged somehow based on the one roll. it doesn't make the odds any better over several rolls (you could argue it makes them worse once advantage is applied) it just removes the ability to do nothing by forcing an action out of the roll.
I am with Cap on this, with the opposed tests there is no need to have parry and dodge. The odds are not better but who cares? When you roll you know whether you succeeded or failed based on two rolls and this is all that you need, fast and efficient. Applying advantage, or how to apply advantage is the discussion that Cap is trying to resolve for his game. From my point of view, I started with RAW, but I am considering to deduct advantage as you use it.
You find opposed rolls faster than an attack and a parry roll?
mormegil
Posts: 56
Joined: Wed Jan 09, 2019 10:48 am

Bitsa wrote:
Tue Feb 19, 2019 4:29 am
mormegil wrote:
Tue Feb 19, 2019 3:06 am
Orin J. wrote:
Mon Feb 18, 2019 11:51 am


that's not how 4th solved whiffing with opposed rolls. they solved whiffing by removing the ability to apply multiple layers of failure to an attack and forcing it into a single opposed roll for each attack. instead of an attack being successfull if it hits AND if it isn't parried AND if it isn't dodged, it's now either successful OR it's missed, defended or dodged somehow based on the one roll. it doesn't make the odds any better over several rolls (you could argue it makes them worse once advantage is applied) it just removes the ability to do nothing by forcing an action out of the roll.
I am with Cap on this, with the opposed tests there is no need to have parry and dodge. The odds are not better but who cares? When you roll you know whether you succeeded or failed based on two rolls and this is all that you need, fast and efficient. Applying advantage, or how to apply advantage is the discussion that Cap is trying to resolve for his game. From my point of view, I started with RAW, but I am considering to deduct advantage as you use it.
You find opposed rolls faster than an attack and a parry roll?
Yes, because opposed rolls are rolled simultaneously and you compare level of success. The other way was: roll, check if you succeed, then make a second roll for parry or dodge and check again if you succeed.

In addition, now there is no need to roll for damage, as you get it from the difference of the opposed roll plus other static modifiers. So eventually, it is at least one roll less.
CapnZapp
Posts: 263
Joined: Tue Jan 08, 2019 2:15 am
Location: Norsca

Making rolls opposed reduces whiffiness.

This can trivially be shown. Consider two poor combatants (generally the case where whiffy complaints are the strongest).

If they both have 25% the probability of an attack actually "landing" (dealing damage) is ~18%. First the character needs to hit: 25%. Then the opponent gets to parry, which succeeds in one out of four tries.

If the attack is opposed you should easily see that the probability is 50%* for all cases where attacker and defender has the same skill (poor or good).

50* > 18. Q.E.D.

*) Not necessarily exactly 50%. The exact figure depends on how you count Success Levels when the result is close to your skill score.

---

The opposed roll also speeds up combat, since both rolls are made at the same time, instead of one being contingent on the other. That is, there is no wait until the defending roll is made (when it is made).

(Not saying 4E is faster than 2E. Only that out of two otherwise identical systems, the one using opposed Tests become faster)

---

When combatants are actually good, stakes are usually so high few 2E players minded the occasional miss. Yes, you hit more often, but parries/dodges are also much more often successful.

Generally the main reason whiffiness is reduced is because 2E offense increases faster than 2E defense: you go from 1 Attack vs parry/dodge to 3 Attacks vs parry+dodge.
Graak
Posts: 51
Joined: Mon Jan 14, 2019 11:50 pm

mormegil wrote:
Tue Feb 19, 2019 5:59 am

Yes, because opposed rolls are rolled simultaneously and you compare level of success. The other way was: roll, check if you succeed, then make a second roll for parry or dodge and check again if you succeed.

In addition, now there is no need to roll for damage, as you get it from the difference of the opposed roll plus other static modifiers. So eventually, it is at least one roll less.
Regarding speed, not a major point, but worth mentioning:

Ehm, I'm quite sure that checking 2 rolls for binary success/file should be faster than checking 2 (simultaneous) rolls both for success/fail AND for relative degrees of success.

At our table dices outnumber players more than 2:1 and the two rolls are almost simultaneous. We don't pass each other dice BTW (2e)

I'd stress out that in 2e you have to choose if you are going to parry OR dodge a succesfull attack. Hence: maximum 2 rolls (swift attack aside) and then roll for damage. At this point is applying damage in 2e slower than 4e? (I don't know, I have to dig deeper into 4e)

Speed chapter closed :)
CapnZapp
Posts: 263
Joined: Tue Jan 08, 2019 2:15 am
Location: Norsca

Graak wrote:
Tue Feb 19, 2019 10:58 am

Speed chapter closed :)
Or at least, maybe you guys can take it to a new thread :)
Hell Hound
Posts: 4
Joined: Fri Mar 29, 2019 10:40 am

Hey there CapnZapp,

I just picked up the core book and I'm reading through it and in my head I'm agreeing with most, if not all, of what your initial post states. But I'm wondering if your feelings have changed at all having had a bit of time to play or house rule it?
I'll reserve my final judgement for when we actually play the game however, me being to GM, I have to admit, it looks like a lot of work to actually run a game especially if you're not using something prewritten. 2e had it's challenges in that regard but I've got "Red Alert," flashing as I read through the 4e book.

Cheers.

Edit: I did not read this entire thread when I posted so I realize now you have been working on house rules. All the best on making something your'e happy with. At this point I think I'd be happy playing in a game of 4e to see how it goes but not running it.
Capitaneus Fractus
Posts: 104
Joined: Mon Mar 11, 2019 7:41 am

I don't have WFRP4 (I wait the French translation) so I didn't had the opportunity to read its rules, but I presume that the point raised by CapnZapp on pure randomness of opposed tests could be solved with dividing opposed tests between:

Opposed tests
Used when two opponents directly oppose with the same aims, the same means in actions where the success of an opponent necessary imply the failure of its opponents. E.g. combat with melee weapons and in the intent, by both combatants, to strike to hit; chess game; arm wrestling; tug of war, card game without cheating...
--How it work: rolls are made simultaneously, there is not an aggressor and a defender, both rolls are considered as attacks and defences.
  • if both succeed, success levels are compared:
    • if both success levels are equals, there is an interchanged thrust. Both opponents does their damages (it have to be adapted to the context, obviously, in case of tug-of-war, none player move...). All advantage is lost;
    • if success levels aren't equals, the opponent with the highest success level hit, get the advantage and does its damage. His opponent lose all his advantage;
  • if only one succeed, he hit and get the advantage and does its damage. His opponent lose all his advantage;
  • if both fail, failure levels are compared:
    • if failure level, the opponent with the lowest failure level hit, get the advantage and does its damage. His opponent lose all his advantage;
    • if both failure levels are equals, no one have hit this round, all advantage is lost.
In this case, the randomness isn't much problematic as there have to be someone who succeed!

Contested tests
Used when two contesters indirectly oppose in an action, either because they doesn't share the same aim, or they doesn't employ same means, or because the success of a contester doesn't necessary imply the failure of his antagonist. E.g. combat with melee weapons where the intent of one combatant is to protect himself without trying to hit, card game where one player cheat, trying to outwit someone's vigilance...
--How it work: rolls are made simultaneously, there is an agent and a contester:
  • if both succeed, dice results are compared:
    • if both dice results are equals, both actions succeed (it have to be adapted to the context: the guard didn't seen the thief, but he quite immediately remark that the jewel have disappeared);
    • if dice results aren't equals, the highest dice result succeed;
  • if only one succeed, then, he is the one who succeed...;
  • if both fail, both fail... (the thief never found an opportunity to approach the jewel, but the guard didn't remarked anything);
(comparing dice results instead of success level seem faster and have more or less the same function: if you succeed with an high roll, that is that your skill allow you tu succeed with an high roll...).
Veniam, Duelli Malleum, phantasticum ludum personae uidebo, in fera terra periculosorum aduenturorum ludebam.
User avatar
Orin J.
Posts: 403
Joined: Wed Feb 06, 2019 10:39 pm

So...ignoring our little derailment over the details of the issues with 4th's rulesystem, how's zapp's overhaul coming along?
Frore
Posts: 1
Joined: Sat Apr 20, 2019 5:26 am

I've been following this thread with interest, along with similar discussions on rpg.net and reddit. Very much looking forward to seeing your collected house rules CapnZapp.
User avatar
Yepesnopes
Posts: 15
Joined: Tue Mar 19, 2019 4:29 am

I am afraid CapnZapp went RAW
User avatar
totsuzenheni
Posts: 286
Joined: Mon Jan 07, 2019 12:38 pm

ROAR more like.
CapnZapp
Posts: 263
Joined: Tue Jan 08, 2019 2:15 am
Location: Norsca

Frore wrote:
Sat Apr 20, 2019 5:30 am
I've been following this thread with interest, along with similar discussions on rpg.net and reddit. Very much looking forward to seeing your collected house rules CapnZapp.
Thank you - and welcome to the forums!

We're still playing, though we've had a lull in the fighting lately (and combat is what needs playtesting the most).

I still believe I will get back to this eventually :)
User avatar
Rat Catcher
Posts: 46
Joined: Fri May 03, 2019 1:11 am

magnus the flyest wrote:
Thu Jan 31, 2019 12:12 pm
Sounds like this isn't the system for you.
Was thinking the same. Not that I have much of an idea, seeing as I'm still going through the rulebook.
Post Reply