hsi379
Topic Author
Posts: 33
Joined: Fri 10 Sep 2021, 19:24

Re: Council Skill Endeavor math odds success rate

Wed 20 Oct 2021, 18:58

hsi379, please, could you explain your math?
Here's the python code I used for 4S TN14, at least 6 successes in 6 tries. The probability of getting 0 success, 1 success, etc. in the 4S dice pool is from Anydice:

https://anydice.com/program/24582

I manually changed the parameters to fill out the table because I am not a coder...
import random
count=0

for i in range (1000000):    	
    numberList = [0, 1, 2, 3, 4, 5]
    mylist= (random.choices(numberList, weights=(11.433, 38.593, 36.78 , 11.574, 1.543, 0.077), k=6))
    if sum(mylist) > 5:
        count = count +1
        
print (count/1000000)
Thanks for also calculating. If I'm reading them right, your TN14 table shows around 10 percentage less than my numbers, with the same trend.

I haven't had a chance to carefully look at your code yet, but I'd like to figure out which is correct! Does your program take into consideration the auto success of a Gandalf (seems like it does)?
 
fahrgast
Posts: 37
Joined: Sat 03 Jul 2021, 16:20

Re: Council Skill Endeavor math odds success rate

Wed 20 Oct 2021, 19:22

I haven't had a chance to carefully look at your code yet, but I'd like to figure out which is correct! Does your program take into consideration the auto success of a Gandalf (seems like it does)?
Yes, it does!

I think that the problem with this rule comes from two facts (and skill endeavours may have the same problem):
  • Success icons count as additional successes.
  • Number of rolls is tied to number of successes needed.
So I think there are two possible solutions:
  • Ruling that success icons don't count as additional successes. I think that this solution goes against the spirit of the rules.
  • Decoupling the number of rolls and the number of successes needed.
I have imagined a rule that sets the number of successes needed the same way as written in the rulebook, but makes the number of rolls depend on the success of the introduction roll:
  • With a simple success in the introduction roll, the fellowship gets 3 rolls.
  • For each success icon in the introduction roll, they get 3 additional rolls.
I have run a simulation with a TN of 14 and these are the results:

Skill Rolls  Resistance    
2S           3        6        9
      3      22,84 %  0,82 %   0,00 %
      6      57,87 %  11,99 %  0,82 %
      9      80,30 %  34,00 %  6,62 %
3S
      3      59,76 %  7,56 %   0,07 %
      6      93,89 %  57,40 %  15,05 %
      9      99,33 %  89,15 %  56,29 %
4S
      3      84,16 %  21,49 %  0,57 %
      6      99,59 %  88,90 %  47,18 %
      9      99,99 %  99,45 %  92,05 %
These numbers might be appropriate or not, but they don't have the problem observed with the official system.
 
gyrovague
Posts: 591
Joined: Tue 28 Apr 2020, 16:52

Re: Council Skill Endeavor math odds success rate

Wed 20 Oct 2021, 19:34

Sorry, I had misunderstood you. The rules for the phase of interactions say "The players choose their own course of action as they see fit." However, I'm not sure that this means that players can plan ahead a series of rolls. I think that choosing the skill for each subsequent roll should depend on the dialog between the heroes and the npcs. Otherwise, a council would be a silly and boring mechanic if a company could choose to roll always the same skill.
I agree that should be the case, but all-in-all I would rather have an RPG mechanically encourage interesting, unexpected narrative, rather than ask players to have to make poor tactical choices in order to vary play, or to require the Loremaster* to improvise and make constant judgment calls in order to keep things interesting.

*I recognize that this sort of improv/judgment is one hallmark of an excellent LM. However:
1) There is a global shortage of GMs (in all RPGs) and to the extent that rules can scaffold the LM's job I think they should
2) When a changing tactical environment is unsupported by rules and is up to the LM's whim, it can be hard (read: less fun, YMMV) for players to figure out what's going on. They have to guess at what the LM is thinking, rather than use rules to predict outcomes.

In general I prefer rules that guide NPC/LMC behavior deterministically, so that the LM can remain neutral. (This is a preference, not a claim of "correct" RPGing.). For example, although the old 1e rule about adversaries making called shots when the player rolls an eye had counter-intuitive results, I did like the fact that the LM didn't have to decide when to make a called shot (although they could for narrative reasons), and instead the use of adversary called shots was deterministic.

So the notes in the section on the Interaction phase, e.g. "The objective of the Skill attempt must be obvious; otherwise, even if the roll is a success, its effect will be weak" are not helpful. Is that supposed to mean that if the LM doesn't think the player's intent is obvious, the successes shouldn't be counted?

Imagine a combat system that included the note, "If the attack isn't sufficiently vigorous, even if the roll is a success, its effect will be weak." WTF?

For the same reason, I'm wary of the sidebar on page 107 about "Rewarding effective roleplaying." I get it, and I do play this way (across RPGs) but at the same time there's a risk to giving mechanical benefit to players who are good at persuading the LM.
 
hsi379
Topic Author
Posts: 33
Joined: Fri 10 Sep 2021, 19:24

Re: Council Skill Endeavor math odds success rate

Thu 21 Oct 2021, 06:40


    3        3+1      3+2      6        6+1      6+2      9        9+1      9+2
2S  22,83 %  35,42 %  47,30 %  11,97 %  18,53 %  26,04 %  6,61 %   10,27 %  14,85 %
3S  59,68 %  77,41 %  87,99 %  57,42 %  71,60 %  82,08 %  56,28 %  68,56 %  78,36 %
4S  84,14 %  94,93 %  98,52 %  88,86 %  95,55 %  98,38 %  92,11 %  96,55 %  98,60 %
Hi fahrgast,

I looked over your code and couldn't find anything wrong with it, so I ran the code for various scenerios and I actually got my original numbers not the table above?

For instance, when I run your code for
3S TN14, at least 6 in 6 rolls I get 70% (same as OP table) not 87.99?
4S TN14, at least 3 in 3 rolls, I get 92% (same as OP table) not 84.14% as in your table above
3S TN14, at least 3 in 3 rolls, I get 69% (same as OP table) not 59.68%
3S TN15 at 6 in 6 rolls, I get 48% which is the same as my original tables
4S TN 16 at 3 in 4 rolls, I get 94% which is the same as my origianal table
Etc.

Basically I ran 12-15 different scenerios using your code and came up with the numbers on the OP tables. So it seems like your code is actually returning the OP numbers and not the table above. Is there a chance that something happened in the translation of the code you posted (which seems good) and the final table run or that it represents a different TN ?

Would love to confirm the numbers as it helps when trying to figure out a fix.

Thanks!
 
fahrgast
Posts: 37
Joined: Sat 03 Jul 2021, 16:20

Re: Council Skill Endeavor math odds success rate

Thu 21 Oct 2021, 08:03

Sorry, hsi379. Last night I found a bug in my code and I fixed it, but I had no time to make a new table.

Thank you for doing it yourself. So your original numbers were correct, and your point stands the same.
 
hsi379
Topic Author
Posts: 33
Joined: Fri 10 Sep 2021, 19:24

Re: Council Skill Endeavor math odds success rate

Thu 21 Oct 2021, 18:57

Sorry, hsi379. Last night I found a bug in my code and I fixed it, but I had no time to make a new table.

Thank you for doing it yourself. So your original numbers were correct, and your point stands the same.
Thanks for confirming! Great we can feel confident with those numbers by multiple methods, and your code makes it much easier to run future scenerios.

Unfortunately, as this thread is showing the point of the 3 , 6 , 9 length = difficulty math not working is just one of many problems with Councils I think!

Even if we fix this issue, there is still the issue with the "step function" of adding and subtracting d6s. This is somewhat true throughout the whole game (+/- d6s are cruder than adjusting TN) but because of multiple rolls and the importance of multiple Tengwars in Councils/Skill Endeavors the issue just gets magnified.

Then, there is the seperate but related issue of the fact that Councils are not very strategic and from a 'game' perspective don't give the players many choices vs. say combat (stance, weapon, do something other than a strike, etc.).

I actually think a simple system with the right math can work. D&D4e skill challenges / Stalker0 Obsidian version https://www.enworld.org/threads/stalker ... -2.241440/ (even better, see pdf in post) are very abstract and are basically die rolls with a codification of success. There are also not many strategic choices but the math works, and LMs are encouraged to set different TNs for different skills which adds some variety. Obsidian also has tiered success --- failure, partial success, and success built in. And players have some limited resources that can use (e.g., powers which let them reroll skills) but not many.

IMO, the key to making these simple / abstract systems work is to vary the fictional positioning along the way so that the same skills are not always relevant and new challenges are presented so the "scene" doesn't feel static (all towards the same end goal). Having a penalty for "failure" or "success at a cost" that is interesting and a real narrative cost is also important.

Would it be great to also have more strategic choices for the player? Yes. But I think the primary value of these non-combat systems is in the resolution part. Something important is decided over multiple roles using an objective system instead of 1 roll or "X rolls and LM stops it when they feel like it", and the LM knows approximately what the difficulty actually is (when the math works) of the challange and can more easily match or adjust this to the fiction.

So, I'd love to start with a design where the math works and go from there. I think the key is setting the difficulty of the request relative to the kinds of things a 2S party could reasonably expected to do vs. say a 5S party, and then modify from there based on other factors. There is just too much difference in "power" with different dice pools and the multiple success mechanic.

So say start with the design principle of -- "equal level" situation should succeed 60% of the time without modifiers -- and go from there.

Trying to convince a woodsman tribe to guide you to a shuned cave not far from the edge of Mirkwood -- that's a 3S challenge -- so a party of average 3S skills has a 60% of succeeded without spending resources or gaining some other advantage (and not counting disadvantages of prejudice, etc.)

Trying to convince Theoden to lend a patrol of riders to assault X stronghold -- that's a 5S challenge. 3S don't have much of chance. 4S parties can maybe do it spending a lot of resources (hope) or working for an advantage. 5S can just walk up and have a decent chance or can almost guaruntee it using resouces -- which is fine since a party with an average of 5S skills are movers and shakers of middle earth now...
 
a2le
Posts: 23
Joined: Fri 03 Jul 2020, 19:11

Re: Council Skill Endeavor math odds success rate

Thu 21 Oct 2021, 19:25

... a council would be a silly and boring mechanic if a company could choose to roll always the same skill.
Well, that was impossible in the alpha, because they couldn't use the same skill more than once in the same interaction without spending hope, (see alpha p. 104 useful skill). The same rule has not survived in the final version, however, the general rule that characters "have only one attempt at anything that is resolved with a die roll" is still there (p. 17 repeating a roll), then rolling the same skill, again and again, is not a real option here.
 
gyrovague
Posts: 591
Joined: Tue 28 Apr 2020, 16:52

Re: Council Skill Endeavor math odds success rate

Thu 21 Oct 2021, 22:31

On my phone so can’t type an essay, but agreeing with a lot of what hsi is saying.

I will also acknowledge that is hard to accomplish all this without adding a lot of rule complexity, which I don’t think is in the spirit of TOR. To once again quote Jon Hodgson, the goal should be “simple rules with deep implications.”

Right now it’s simple rules with shallow implications, and so far my proposed solutions are complex rules with deep implications, which isn’t necessarily any better.

So my critique comes with the acknowledgement that I haven’t solved it. It’s hard.

Oh look somehow I typed an essay anyway.
 
Dunheved
Posts: 494
Joined: Wed 11 Mar 2020, 02:07
Location: UK

Re: Council Skill Endeavor math odds success rate

Fri 22 Oct 2021, 01:06

Apologies if I have misunderstood this conversation.
But one of the problems with the mechanic of this rule is that the Councils with greater challenge are actually easier (as in more likely) to be achieved overall, because there are so many rolls (so much "Time") in the most Challenging Councils. And the extra rolls give the players the opportunity to make up for poor initial rolls with some T result rolls later?

I have a thought about the 3, 6, 9, rolls mechanic: which I think should mean that the Challenge gets greater and not easier. I am trying to amend the RAW as little as possible here.

The "Time limit" is re- organised into bundles of three rolls.
If the Introduction Phase is Successful then the players can decide which of these bundles has Extra time allocated to it. Extra time, when allocated, cannot be saved for later use in a different bundle of three rolls. Each bundle of time is needed for a Round in the Council.
Round 1 Three Successes must be obtained. Else the Council ends in Failure. By default three rolls are permitted. However, any of the Introductory Success rolls can be use in this round if they were allocated here in advance. Round 1 makes up the first Bundle of Time.
Round 2 Three New Successes must be obtained. Else the Council fails. By default three rolls are permitted. However, any PRE- ALLOCATED Introductory Success rolls can be used in this round to increase the "Time" permitted in this second round. Round 2 makes up the second Bundle of Time and is needed if the Council is more Challenging than an Easy Council.
Round 3. Three FURTHER Successes must be obtained. Else the Council fails. By default three rolls are permitted. However, any PRE-ALLOCATED Introductory Success rolls can be use in this round to extend the "Time" available in a Difficult Council.

Of course an Easy Council reaches Full Success in Round 1. The extra Rounds are not needed. The next level of difficulty will need Success in Round 1 and it must be followed by Success in Round 2 before it can end in a positive conclusion.
The most challenging Councils will need all three consecutive Rounds: each building on the Success of the former Round before a Final positive conclusion is reached.

Organising things with this additional restriction will make the percentage chance of Overall Success greatest with the easiest council and least with the most challenging council.
 
fahrgast
Posts: 37
Joined: Sat 03 Jul 2021, 16:20

Re: Council Skill Endeavor math odds success rate

Fri 22 Oct 2021, 09:28

For the first problem, "the Councils with greater challenge may be actually easier", I have a simple solution: going back to the concept of Tolerance.

Let's say you need 3, 6 or 9 successes to achieve your goal in a council. You don't have a time limit, but you have a limited number of failures that depend on your Introduction roll (0, 1 or 2 failures allowed; this is the Tolerance). When you get more failures than allowed, the council is ended.

I have run some simulations with this rule and this is the result:

With TN 14:
     3 (T0)  3 (T1)   3 (T2)   6 (T0)   6 (T1)   6 (T2)   9 (T0)   9 (T1)   9 (T2)
2S   5.7691  33.926   50.6026   3.1106   9.8929  19.4487   0.6117   2.5547   6.2655
3S  47.0921  75.8932  89.9708  24.5909  52.577   73.624   12.8106  34.2288  55.9755
4S  78.1648  96.1748  99.3903  63.5202  90.4173  97.8801  51.6865  83.4265  95.324

With TN 16:
     3 (T0)   3 (T1)   3 (T2)   6 (T0)   6 (T1)   6 (T2)   9 (T0)   9 (T1)   9 (T2)
2S    7.3788  17.2593  27.809    0.7769   2.6979   5.8817   0.0761   0.3791   1.024
3S   29.766   55.38    73.2915  10.5481  27.4187  45.4146   3.7397  12.3988  25.0566
4S   62.1986  87.8602  96.4155  41.6066  73.3856  89.4211  27.8417  58.7829  79.8995

Who is online

Users browsing this forum: No registered users and 3 guests