A Question of Skill Check Mechanics

Locked

Which "skill check" mechanic option do you prefer?

Percentile Target Value
3
38%
Fixed 2D10 Skill Check with Opposed Stat Table Modifiers
5
63%
Other
0
No votes
I don't like any of these options!
0
No votes
 
Total votes: 8

User avatar
Tyrel Lohr
Vice Admiral
Vice Admiral
Posts: 1466
Joined: Thu Oct 04, 2007 3:48 am
Location: Lusk, WY
Contact:

A Question of Skill Check Mechanics

Post by Tyrel Lohr »

Recently, the internal review group for Second Edition has been locked in debate as to the best way to handle "skill checks" in the new edition. After discussing things internally (and chasing our collective tails a bit!), we've decided to put the question to you, the fans, to see what your feedback is on this subject.

First, before we talk about possible solutions, it might help to illustrate the problem. In VBAM 1E, each type of "check" -- be it for intel mission success, morale effects, etc. -- followed its own set of rules. However, in 2E, we desperately want to simplify this by providing players with a single, consistent manner for performing these checks.

Two options have presented themselves thus far:

Option #1: Percentile Target Value

This option works similar to the 1E Intel rules: you take one active statistic (offensive intel, in the case of intel missions) and divide it by a passive statistic or combination of statistics (such as intel mission difficulty) to determine the percentage chance of success/failure. However, instead of being a chance of failure (as per 1E Intel), this percentile target value is the chance of success.

Success and failure chances are then subdivided into major and minor results for each. A major success is scored on a D100 roll less than half the target number, while a major failure is scored on a D100 roll twice or more than that same target number. All other results are minor success or failures, as appropriate.

The advantage to this option is that it scales fairly nicely, and allows for fairly simple math to be used to calculate your target number for any given check. The disadvantage is that it is difficult to balance, and extreme results can leave a player with an extremely high chance of rolling a major failure. Alternatives have been put forth to address this hazard, most of which boil down to reducing major/minor results to a 50/50 split.

Example: Your empire is launching an intel mission against an enemy colony using 8 intel points. This mission has a difficulty of 3 and the enemy has 2 defensive intel assigned to that star system. The active stat is offensive intel, so you have a total of 8 activate points. The passive stat is mission difficulty plus defensive intel, which is 5. The check's target value is equal to 8 / (10 + 5) [where 10 is a fixed constant] = 53% chance of success (26% chance of major success), and a 47% chance of failure (0% chance of major failure, however).

Note: The playtest rules as written use this option, though it is not too late to replace it with something better.


Option #2: Skill Check Table

The second major option so far is the adoption of a fixed skill check table that players roll 2D10 against, plus modifiers, to determine mission success/failure, as well as the degree of success/failure (i.e., major or minor result). The chart is as follows:

Result Effect
7- Major Failure
8-11 Minor Failure
12-14 Minor Success
15+ Major Success

At average difficulty (+-0 modifier), players have a 45% chance of success (21% major, 24% minor) and a 55% chance of failure (21% major, 34% minor). This chart is then balanced around extremes of +-5. The minority chances at either extreme are 10-15%, with only a 1% chance of major success (-5) or failure (+5).

The modifiers for this roll are determined by taking the same combination of active and passive values from Option 1 and pitting them against one another on an opposed stat table (click here). This table uses a formula that subtracts the passive stat from the active stat and then divides the result by 2. This provides a +-5 spread for values of 0-10. However, to maintain a +-5 spread, a corrective measure must be taken on values greater than 10 to reduce them proportionally so that the largest number is 10. This is done by dividing both the active and passive statistics by the higher of the two divided by 10. This reduces the highest stat to 10 and the lowest stat to a proportionally lower number. (ed: There has to be an easier way to do this calculation, and I know I am doing a bad job of explaining it -- refer to the pre-generated table to see the actual effects).

The rules try to keep the active/passive stat totals to 10 or under, but this is not always possible. Take for example the interplay of Sensors and Stealth during detection checks -- a fleet could conceivably have a 100 or more points of the stat present. This is why extra math is required to constrain the values to a 10x10 grid.

Example 1: Your empire is launching an intel mission against an enemy colony using 8 intel points. This mission has a difficulty of 3 and the enemy has 2 defensive intel assigned to that star system. The active stat is offensive intel, so you have a total of 8 activate points. The passive stat is mission difficulty plus defensive intel, which is 5. Referring to the chart where 8 on the X-axis and 5 on the Y-axis meet, our modifier for this intel check is +2.

Example 2: Your empire is launching an intel mission against an enemy colony using 14 intel points. This mission has a difficulty of 3 and the enemy has 7 defensive intel assigned to that star system. The active stat is offensive intel, so you have a total of 14 activate points. The passive stat is mission difficulty plus defensive intel, which is 10. The larger of the two stats is 14, so we divide each by 1.4, giving us an active value of 10 and a passive value of 7. Referring to the chart where 10 on the X-axis and 7 on the Y-axis meet, our modifier for this intel check is +2*.

* This last example does raise a point that I just noticed that the reduction method does not always compute the same as a straight mathematical model, which does pose problems. However, in this case, I think the formula on the Excel sheet is better than the option's current wording because it allows for large values that are effectively the same to retain the most reasonable bonus/penalty. It is also worth noting that, while the provided table provides active/passive values out to 25, it could just as easily stop counting by ones at 15 or 20 and then start counting by fives until we can very large values. A formula will still be needed to calculate very large values, but a goal under this option would be to provide a single lookup table that would cover most active/passive values.


Let us know what your thoughts are on this matter, and whether you see either of these as being better or worse than the other, or if you have a solution of your own that might work better. Any check mechanic will have to accept a wide range of values, however, which is the largest impasse if finding a "best fit" option (intel and morale will tend to be in the 0-10 range, but other stats could be much, much higher).

Thoughts?
-Tyrel
Last edited by Tyrel Lohr on Fri Dec 04, 2009 5:37 am, edited 1 time in total.
[i]"Touch not the pylons, for they are the messengers!"[/i]
Gareth_Perkins
Captain
Captain
Posts: 242
Joined: Wed Oct 10, 2007 11:39 am
Location: Exeter; UK

Re: A Question of Skill Check Mechanics

Post by Gareth_Perkins »

This has made me think:
Tyrel Lohr wrote: Any check mechanic will have to accept a wide range of values, however, which is the largest impasse if finding a "best fit" option (intel and morale will tend to be in the 0-10 range, but other stats could be much, much higher).
It suggests that there almost has to be division somewhere in the calculation in order to consolidate modifiers down into the same range for every die roll regardless of the magnitude of the modifiers (or the type of die roll made),

It also begs the question, in these cases: will both the "attacker" and "defenders" stats be of the same magnitude every time?

If they are of the same magnitude every time then dividing the higher "stat" (plus mods) by the lower "stat" (plus mods) and rounding off (or down, or even up) will quickly give you a simple modifier, to which possibly further modifiers could be added, without needing the big chart (which BTW, I really don't like, despite generally preferring the 2D10 option),

All of that said, having a division in a frequently used game-function does bother me slightly - at least for numbers of magnitude 1-20 or so, the calculation is trivial,
Gareth Lazelle
User avatar
Tyrel Lohr
Vice Admiral
Vice Admiral
Posts: 1466
Joined: Thu Oct 04, 2007 3:48 am
Location: Lusk, WY
Contact:

Re: A Question of Skill Check Mechanics

Post by Tyrel Lohr »

Gareth_Perkins wrote:It also begs the question, in these cases: will both the "attacker" and "defenders" stats be of the same magnitude every time?

If they are of the same magnitude every time then dividing the higher "stat" (plus mods) by the lower "stat" (plus mods) and rounding off (or down, or even up) will quickly give you a simple modifier, to which possibly further modifiers could be added, without needing the big chart (which BTW, I really don't like, despite generally preferring the 2D10 option),
Straight up division is out because of the "dividing by zero" problem. You can handwave away division by zero as always being a maximum value at one end of the scale or the other, but that makes no logical sense when the difference between active and passive stats is 1 (such as 1 and 0).

Now, things could be simplified so that, instead of a formula, you just compared how many times larger one stat is from another. For example, you might receive a +3 modifier if your active stat is >= 3 times the passive stat, but < 4 times the passive stat. This would still be a division issue, so zero values would still require a special rule to resolve; however, zeroes could be treated as 1/2 or else use the difference between the two stats.

Using the Example #1 from my previous post with 8 offensive intel and a total difficulty/defensive intel of 5, this other alternative would give you a 8/5 = 1.6 modifier. If rounding naturally, this would go up to +2.

Let's assume then that we had offensive intel of 4 and a total difficulty/defensive intel of 7. At this point we would flip the sign and take 7/4 = 1.75, giving us a -2 modifier.

However, I just realized that stats that are the same end up being a problem using this method, as they generate +-1 results instead of 0. This could be fixed by subtracting 1 from the results. That makes the first example a +1 and the second a -1. I am also thinking that fractions in this case may actually need to be rounded up, so that a 14/10 skill check would result in a +1 (1.4 RU = 2 - 1 = 1 ) instead of a +-0.

-Tyrel
[i]"Touch not the pylons, for they are the messengers!"[/i]
Gareth_Perkins
Captain
Captain
Posts: 242
Joined: Wed Oct 10, 2007 11:39 am
Location: Exeter; UK

Re: A Question of Skill Check Mechanics

Post by Gareth_Perkins »

Tyrel Lohr wrote: Straight up division is out because of the "dividing by zero" problem. You can handwave away division by zero as always being a maximum value at one end of the scale or the other, but that makes no logical sense when the difference between active and passive stats is 1 (such as 1 and 0).
When are you going to get zeroed stats?

If the "defensive stat" is MD + Mods then MD simply has to be set at 1 or more,

For the offensive stat most actions are going to require a minimum stat of 1 to initiate the action, right?
Tyrel Lohr wrote:Now, things could be simplified so that, instead of a formula, you just compared how many times larger one stat is from another. For example, you might receive a +3 modifier if your active stat is >= 3 times the passive stat, but < 4 times the passive stat. This would still be a division issue, so zero values would still require a special rule to resolve; however, zeroes could be treated as 1/2 or else use the difference between the two stats.
That is functionally identical - even with the special rule for zeroed stats!
Tyrel Lohr wrote:However, I just realized that stats that are the same end up being a problem using this method, as they generate +-1 results instead of 0. This could be fixed by subtracting 1 from the results. That makes the first example a +1 and the second a -1. I am also thinking that fractions in this case may actually need to be rounded up, so that a 14/10 skill check would result in a +1 (1.4 RU = 2 - 1 = 1 ) instead of a +-0.
Subtract 1 from all modifiers after performing the calculation perhaps?
Gareth Lazelle
User avatar
Tyrel Lohr
Vice Admiral
Vice Admiral
Posts: 1466
Joined: Thu Oct 04, 2007 3:48 am
Location: Lusk, WY
Contact:

Re: A Question of Skill Check Mechanics

Post by Tyrel Lohr »

When are you going to get zeroed stats?
  • * Detection Checks: When Sensors and/or Stealth are zero.
  • * Morale Checks: When Morale or morale difficulty are zero.
  • * Intel Checks: When mission difficulty + defensive intel is zero.
For the offensive stat most actions are going to require a minimum stat of 1 to initiate the action, right?
These skill checks can be triggered automatically, not just from things like intel missions. Therefore there is a good chance that the active or passive stats may be zero.

That is functionally identical - even with the special rule for zeroed stats!
That is the point. The opposed stat chart works perfectly for comparing two opposed values to generate a skill check modifier. This formula just makes it a bit easier to calculate on the fly.

-Tyrel
[i]"Touch not the pylons, for they are the messengers!"[/i]
User avatar
DrakenFyre
Ensign
Ensign
Posts: 6
Joined: Tue Mar 31, 2009 3:58 pm

Post by DrakenFyre »

I've played variety of RPGs using both of these systems so here's my opinion on them.

The percentile target system would give you much finer granularity on success percentages but the math involved to caclucate the targets could end up being needlessly complex, and would be very difficult for anyone playing without the aid of a computer or calculator.

The skill check table does end up handwaving away some of the advanced detail the percentile system keeps but does make calculating and rolling targets much faster and easier.

Personally I prefer the skill check table system as many times I've found that a percentile system ends up turning the game into more of a math-sim than a game. In many cases in RPGs I've noticed that a lot of percentile based skill rolls end up getting handwaved away by the GM rather than having to calculate the target and make the roll.
Gareth_Perkins
Captain
Captain
Posts: 242
Joined: Wed Oct 10, 2007 11:39 am
Location: Exeter; UK

Re: A Question of Skill Check Mechanics

Post by Gareth_Perkins »

Tyrel Lohr wrote:
  • * Detection Checks: When Sensors and/or Stealth are zero.
  • * Morale Checks: When Morale or morale difficulty are zero.
  • * Intel Checks: When mission difficulty + defensive intel is zero.
Addressing those in turn:

* Detection: possibly a fair point - although don't all ships have some sensors (i.e. rating 1+)?
* Set morale difficulties >0 and that's gone. If morale is at zero presumably there is no point to making a morale check (the colony is already revolting)?
* Set mission difficulties >0 and this will never arise,
Gareth Lazelle
Locked