The mathematical comprehensibility of PE mechanics

[Ainamacar]

Just how easy to understand do you think PE's mechanics should be? This is a complicated poll because I think it's a fairly complicated issue. As mentioned in the recent update the designers are not constrained to use a d20 or similar math because the game doesn't have all the playability restrictions of a tabletop RPG. Nevertheless, having a system that can be largely understood without aid can help the player quickly make informed decisions both in play and while building a character. I generally dislike CRPGs that hide or obfuscate their mechanics because it leaves the player guessing about decisions, when at the very least I would expect the character to have some intuition about his or her own capabilities with respect to a challenge.

 

There is also an inherent subjectivity to this topic because what counts as "easy" math varies quite widely in the population. I am personally comfortable with a large degree of mathematical complexity (that being inherent to my day job) but that doesn't mean I want a game that uses Bessel functions, or even a lot of division. Besides, complexity naturally results as different mechanical elements come together, so simple "elementary mechanics" can lead to rich outcomes. That said, some math is numerically but not conceptually complex. Something like the normal distribution (i.e. Bell curve) is pretty easy to understand conceptually because it has so few parameters and they have a very intuitive interpretation. So while the math is hard in some respects the fundamental concept is contained in two parameters, and math on the parameters is easy. Therefore I imagine a computer game could use it extensively without becoming incomprehensible. YMMV.

 

Please note when taking the poll that when the game is providing information, I'm specifically asking about information given to the player before taking an action. Things like the probability of an attack hitting, and so forth. Displaying the actual math performed to resolve a task (e.g. showing what the "dice" rolled) is a separate issue, although I certainly hope the game makes that information available if desired.

 

Here are my choices, with comments.

 

1. What degree of unaided mathematical "comprehensibility" do you want in the rules?

I chose "Substantial", because I think the core mechanics should be understandable to pretty much anyone that puts in a little time, and because complexity should generally be emergent, rather than inherent. If the crafting system is more inherently complex I'm less worried because generally that is a rare event. However, the relationship between attacks and defense should be much easier to understand, because some form of that is probably the single most common event in the game. Even if the game provides that information explicitly (perhaps when mousing over an enemy) I think the math should be simple enough that I can make a reasonable guesstimate without doing that. After all, if there are 15 enemies I don't want the math to be so complex that I actually need to mouse over all of them to make sure I'm not making a terrible decision. Finally, an understandable system helps a lot when it comes to determining character progression. The basic effect and magnitude of a feat/perk should be understandable when taken, not only when it is used. Few things irk me more than when an ability says something like "improves your basic attack" and then gives me no information on which to judge whether the improvement is significant or not. If the system math is very complex the situation is not much better even if the exact effect is given, and even if the game provides very informative feedback during play.

 

I also think this is a good match for a game inspired primarily by IE games, which I think mostly falls under this category. The elements above are part of what I appreciated about those games. A large change here, while not inherently bad, strains (but by no means breaks) the resemblance for me. Others may feel differently.

 

 

2. To what degree should the game provide explicit information about possible results of an action? Should the degree depend on character skill?

I chose "Basic, None", although my opinion about whether it should depend on character skill isn't fully settled. For people who don't want to learn the math, something like this is extremely useful, and for me the real question is how much information should be provided. As much as I love diving into statistics, I feel that is generally better outside the game. I think the main goal should be to give players about as much information as one might expect the character would have.

 

For depending on character skill, I have mixed feelings. I love the idea of making a game that robustly handles what a character knows, and conveys that to a player, but I think a really good implementation of that would be both difficult and not necessarily in the spirit of the IE games. Moreover, something like the Awareness perk from Fallout 1/2 really changed the feel of the game for me. Not necessarily in a bad way, but I always felt it was inelegant that it was basically a switch. An analogy might be to karma/alignment/reputation systems. A really simple alignment system doesn't add much to the game, and can in fact encourage some ridiculous gameplay. A sophisticated reputation system, when it is finally implemented, can really add a lot to a game. I'm not sure designing an equivalent system to translate character knowledge into player knowledge is what I want the developers to spend time on in PE.

 

3. To what degree would you be willing to compromise system function to ensure it is mathematically "comprehensible" without aid?

I chose "A little". It's always a tradeoff, but I'm willing to tolerate a few more trouble spots to help global understanding. In addition, it's my opinion that understandable systems are generally easier to make work in the first place. The designers are human, after all. I didn't vote "non-issue" because I think it remains an issue when building a character, in complex tactical situations, etc. It mirrors my feeling about calculators: they are useful tools, but not a replacement for understanding.

 

If you got this far, thanks for reading. I hope this will foment a productive discussion.

Edited October 31, 2012 by Ainamacar

[Urist_McDorf]

Don't care, the game should tell me.

 

I love mathematically complex games, they're fun to figure out and play with. However, I'm most likely in the minority, and the complexity is largely vestigal to actual gameplay - if you can make a simple game mechanic be as fun and balanced as a complex one, you'll ensure that more players are satisfied than with a complex mechanic. Simple mechanics are also easier to design and balance, leaving the devs with more time for other important stuff (and time is what Obsidian often needs, judging by some of its games). Thus, I'm happy either way, as long as I can make an informed decision with the game's aid. Making it without the game's aid is nice, but not essential.

 

Substantial, none.

 

That's easy enough. As I've mentioned, I love playing with complex mechanics, and to do that properly mechanic transparency is required - the more, the better. Lack of character skill shouldn't obstruct that, since it's the player who's interested in the underlying maths, not the character.

 

Essentially none.

 

Refer to the explanation on the first point. Gameplay takes precedence over comprehension.

[Honn]

Don' care, None, and something else.

 

I generally don't care as long as I can clearly understand what the result of something is. I hate when I have to google a game just to figure out if weapon X does more or less damage (as an easy example) than weapon Y. That is just sloppy design. I want to be able to compare things, but if it is done by math or just very plain stats doesn't matter to me.

[JFSOCC]

Health bars go up, damage and attack scores must go up, xp bars must grow, enemy health must go down, attributes are better when they're high. Armor doesn't use thac0 because that's counter-intuitive when the rest of the scores need to be high.

 

Maybe it's a bit of an oversimplification, but I have had only marginal difficulty figuring out good builds, mostly they weren't so good, but I had no idea so they felt fine.

[Arundor]

Question 1:

I think the math can be as complex or as simple as Obsidian wants it to be, as long as they show it. The equations should be visible for anyone who wants to try to understand how they work, regardless of whether they are complex or simple. Even if it would take me an hour to calculate what the computer did in 0.001 seconds, I still want that option available so I can do theorycrafting.

 

Question 2:

All mathematical checks should be visible if the player wants to see them, regardless of character skill. Put a toggle somewhere the game options to enable or disable the visibility of explicit skill check data.

 

Edit: To clarify, by "checks" I mean things like skill checks or saving throws. We don't know yet if PE will have those specific mechanics, but that general sort of information is what I'm talking about here. Other sorts of information like opponent HP could be made visible based on character skill, like with the Awareness perk from Fallout.

 

Question 3:

The system should not be compromised at all for the sake of mathematical simplicity. If Obsidian feels they need to use complex math then they should. I plan to try to understand all the little details anyway, but nothing says it has to be quick or easy to do so.

Edited October 31, 2012 by Arundor

[Karkarov]

Question 1: The math should be simple enough that any educated human being in high school or better should get it basically at a glance.

 

Question 2: The game shouldn't tell me crap. I don't want to see a string of nonsense about how I roll a D20 and got a 15 on a bluff check plus my bluff skill VS the DC yada yada yada. It adds nothing to the game and is slightly off putting. Just because it is easy to understand doesn't mean you need to tell me EXACTLY what happened or what my percentage chance of succeeding is.

 

Question 3: No compromise, the game working as intended and being fun to play is vastly more important than me understanding how it works on my end. Like JFSOCC said, this isn't brain surgery. Bigger numbers = better, lower = bad. I am a fighter so I want bigger numbers in things that make me hit harder, hit more often, and make me take more hits before I go down. Figuring out what does that should be fairly simple without and big explanation, a simple tooltip on a stat that says "Strength: It makes you hit stuff harder and you can carry more things" will suffice.To borrow a line... "Common sense rules the day at the Friendly Arm Inn."

Edited October 31, 2012 by Karkarov

[rjshae]

I'd like to know just enough to assess the benefits of the equipment and combat tactics. At higher skill levels I'd also like to be able to assess how difficult a task is likely to be (with words like 'easy' or 'hard').

 

It would be nice if the internal mechanics were explained somewhere, but I shouldn't need to know that in order to play the game.

[Shadenuat]

1. Substantial

If something makes my head hurt but it makes game mechanics more fluid and complex, I'll eat my pills.

2. There is't a suitable option for me. Sometimes game mechanic should be pretty clear, like when I want to know how exactly my character's performance would be changed by strategic choices I make when changing stats, picking skills or weapons; but some things are best to let be in the realm of roleplaying - I'd prefer Diplomacy skill of 4 out of 5 say "You are able to persuade almost anyone into your way of thinking but the most strong willed", instead of "Any character if Will-save lower than X can be persuaded with a % chance of Y".

3. A lot. I know it contradicts with my pick at 1), but it's a gaming system and exists to be played with.

Edited October 31, 2012 by Shadenuat

[wanderon]

Don't know - don't care - send me the game and I will play it - almost always in the manner the devs seem to intend you to -

 

while it annoys me sometimes when the mechanics are hidden and unspecific (adds fire damage) (may slow target) (stun possible) I usually get over it and accept that the item/skill/whatever will do whatever it will do and it's up to me to deal with whatever that is as best as they let me.

 

In some cases it's actually a blessing in disquise allowing me to focus on playing the game and forget about the numbers in the background. This probably works for me becuase I am pretty much never the guy looking for the total build and often go the opposite way by gimping my character in some manner just to give the game a better chance to beat me.

[Matthiasa]

If they go beyond the math I can comprehend, they have bigger issues on their hands. Namely, the time required for such calculations for every event in a given the time it takes to run such that it can appear to be real time.

Even simple equations become computationally complex if they have to run a sufficiently high number of times.

That said, I going beyond what I know would not necessarily be a bad thing.

I am just not sure how say differential or difference equations would actually fit into the game. Solutions start taking too much time if to complex. Going to even higher math it only gets worse...

 

Either way I want to be able to see all off it and be given tools to play with them if the equations get to time consuming to solve manually.

And as you might imagine I am for 0 comprise at all.

If they feel the need to do a form of finite element analysis to get realistic damage results... more power to them.

Edited October 31, 2012 by Matthiasa

[Continuum]

The game should present all the information in a way that you don't need to understand math any more complicated than basic probability (e.g., your lvl 15 fighter Thog has a 65% chance to hit this level 15 goblin), with options to show more or less of the math involved. It should be enjoyable and logical to people who don't look at math, and that can be accomplished WITH solid math that ensures logical outcomes.

 

That said, many RPG players like myself enjoy juggling the gear and probabilities, and I hate it when the math in a game is treated as if it's top-secret. My undestand is that P:E is going to a be a classic style RPG, semi-turn based, with hit and damage outcomes determined by math rather than player skill at aiming. If I'm playing an action game, I don't expect to see any more math than I need to determine which weapon is more accurate and damaging; could even be a graph/image. If I'm playing what I would consider a hybrid action RPG like Fallout or Skyrim, I expect to have access to enough math to accurately compare items/spells/armor. If it's a turn-based game where every attack is determined by probabilities, I expect to be given enough access to the math to make informed decisions based on basic arithmetic/probability. The most common model that Obsidian has presented for P:E is the Infinity Engine games, notably BG and Torment. Those are basically turn-based games that happen on the fly, so I expect all the math to be readily available to me if I want it, including a combat log that I can look at if I want to do so. I hate guessing at how effective something will be.

 

Simply including the combat log and greater levels of mathematical detail as options can satisfy everbody. The effects of every action in this type of game IS decided by math, so it's there to be shown in whatever detail works.

[Grand_Commander13]

I think it's pretty important for the mathematics of the game to be able to be explained by a Project Eternity wiki and guides, so that people making a character will understand how effective their character will be and in what situation. I find it very annoying where you have to merely guess at what will make your character good, or how much making a certain choice will improve them.

[Continuum]

If they feel the need to do a form of finite element analysis to get realistic damage results... more power to them.

This is a good example... as Matthiasa says, the devs should use whatever math is needed to create good results. I don't need to be able to recreate the game in tabletop form with dice. But even if the actual result is mathematically complex enough to require calculus or such, there is generally a simplified version of the equation that can be presented to players in the form of basic arithmetic for decision making purposes.

[Ainamacar]

Voting has pretty much ceased (52 votes at the time of writing), and I think we can safely summarize the results. These, of course, should be taken with the usual internet poll grains-of-salt due to the highly self-selected nature of participants, etc. That said, there were no votes for "Other", so we can have some confidence that the responses adequately capture the basic spectrum of opinions.

 

The desired degree of comprehensibility is fairly evenly split between those who don't care so long as they can make a sound decision, and those who want at least the basic math of the game (if not more) to be understandable, with half of the latter satisfied even if only the basic system is easily understandable. Only 10% of respondents specifically want to avoid difficult to understand elements entirely. That suggests the path that would make the most people happy with this aspect (about 75%) is to implement a system in the neighborhood of "Marginal" and "Substantial", with robust in-game feedback for the sizeable number of people who want that.

 

The nature of that feedback, however, is not subject to any clear consensus. (I personally find this the most interesting result.) About 2/3 want some degree of information, but each of the six options is in the same ballpark. When comparing basic vs. substantial by summing over all the none/somewhat/largely options we see that basic information has a slight edge. Likewise, when comparing dependence on character skill of none vs. somewhat vs. largely by summing over the basic/substantial options we see that None has a slight edge. I don't think that edge is statistically significant. The only significant result I see is that the number of people who want at least some impact from character skill is about double those who want it not to matter at all. The 23% of respondents who don't want any explicit information will probably have to deal, especially given the result of the first poll.

 

The best bet might be to make a game option that offers No explicit information, Basic information, or Substantial information while also implementing a few interesting character abilities. There are two important requirements I would impose. First, a casual player on the Basic information setting who takes no information-related character abilities must have enough info to make sensible choices. Second, the game option must not make the character abilities irrelevant, and so they should provide qualitatively different information. A natural split might be to have the gameplay option concentrate on the detail of known information (like turning "Badly Damaged" into a specific number of hit points), and the character options to grant exploitable world-based information that would otherwise not be available. An example might be a "monster knowledge" ability or subsystem which could be used to support a very lightweight "called shot" system that would occasionally enable a player to take a special action, show a future enemy action, show what an enemy is casting, etc. It should avoid things that are easily circumvented based solely on player knowledge (like the "track" skill in many games that is pointless on a second playthrough or just by reloading). I don't know, just an idea.

 

Finally, the third poll shows a clear preference (among those that care at all) for not sacrificing functionality for understandability. Based on the first poll the small set of core mechanics had better be both highly functional and understandable, but beyond that well-tuned mechanics are the greater priority to the majority of respondents. I don't find that particularly surprising.

 

Thanks again to all those who voted, hopefully you and Obsidian have found it informative.

[TheDogProfessor]

Will Obsidian have to show their working-out?

 

Just imagine:

Longsword damage is modelled by the probability density function:

f(x)=(1/64)(6-x)(x-2)(x+2), 2<x<6

what is the probability that the longsword will deal less than 3 damage?

 

Pr(X<3)= DefiniteIntegral(2,3) (1/64)(6-x)(x-2)(x+2)dx

Pr(x<3)= 31/236

[Dersu42]

The game can use whatever mathematical formulas it wants to determine hit chances, damage, skill success, ect. What the game can not do however is leave me in the dark about what I should do to improve my chances, improve my damage, up my success rate.

 

The areas in which I have control, class selection, skill usage, equipment, character position, stat allocation and item consumption, need to be explicit in their effects. If there are two swords with various stats, and I have no way of determining easily within the game without lots of trial and error which sword is better... then don't show me the stats. Let me just pick the shiniest one and call it good.

 

Torchlight II is a good example of this. The way most of the magical modifiers in that game behave is completely bizarre and defies any kind of reasoning you might be able to apply to it from the very limited information you receive in game. This is not a problem on the lower difficulties as the game lets you steamroll enemies with little effort. However, on elite you need to eek out every advantage you can, and that is simply impossible given the inadequate descriptions.

[PsychoBlonde]

I like it when the math is simple, as in D&D: roll a d20, add modifier, compare to DC. Simple. What I like to NOT be simple is manipulating those various numbers. I want tough tradeoffs, like "I can't both wear the +5 armor item AND the crit resist item!" or "Well, I can increase my fire spell damage by 12% but that means my magic attack value goes down by 2".

 

I want ALL attributes to be valuable for ALL characters, so you are always wishing you had more stat or skill points. That being said, I hate, HATEHATEHATEHATE diminishing returns systems where the more you raise something the less benefit you get out of it, especially when (bleargh!) the only way you have of knowing whether you have "enough" attack or defense is some worthless tooltip that tells you you have a 75% chance to hit monsters "of your level" but everything you ever fight is somewhere between 3 and 500 levels higher than you are so god only knows what your defense numbers actually look like. I am not going to sit there for hours and let monsters beat on me so I can tabulate this crap and determine that when the tooltip says I have a "75%" defense value, this actually means I'm going to be *hit* about 3/4 of the time.

 

THAT is what pisses me off about a lot of games--that the numbers they give you mean NOTHING because they are *relative* to the numbers of theoretical enemies WHO DO NOT EXIST, while the enemies who DO exist have COMPLETELY DIFFERENT NUMBERS which you cannot see.

 

At least with the d20 system you know one thing for certain--your range of random variation. If monsters start to miss you, you know for a fact that nigh-invulnerable AC vs. those mobs is only 20 points away. Likewise if you're hitting sometimes, hitting consistently is no more than 20 points away. Every point you raise your AC or attack = 5% fewer or more hits.

 

That is what I want to see. I don't want or expect them to use the d20 system. What I want is to know, for a fact, that if I raise my dodge percentage by 1%, that means if I go and tabulate every time I get hit vs. every time I dodge, I will see a 1% increase across the board. Even if there are enemies that reduce dodge chance so I would theoretically need 120% dodge in order to avoid their attacks, if I go from 59% to 60%, they're going to be missing me 40% of the time instead of 39%.

 

Also, in my opinion, if a bonus is too small for you to detect just by "feel", i.e. when you raise your attack, it seems like you're hitting more consistently, it is not a big enough bonus for late-game. (Likewise for penalties.) If you have to drag a bunch of people in and sit them in front of computers and have them play alternatively bonus-on, bonus-off, so be it, but I want to be able to TELL when I pick up that mega staff. I don't want no piddly-ass 2% increase over my old staff.

 

Harrumph.

[PsychoBlonde]

However, on elite you need to eek out every advantage you can, and that is simply impossible given the inadequate descriptions.

 

*twitch* Third time I've seen this today. "Eek" is what you shout when somebody puts a spider down your shirt. When you're squeezing something for all it's worth, that's "eke out" whatever it is.

 

This has been a community service message.

[RoughOne]

Honestly, I think the perfect game would have the player figure this stuff out. My sword works pretty well, hmm, wonder if a mace would work better. Or I'm pretty GOOD (I think) with a sword, but if I had to change to a mace to beat up the guy in the tin can suit, would it be better, since I'm not that good with a mace? it's a game. Try things and see. I hate when I can look these things up and find the algorithms for how they are randomized, etc (Borderlands comes to my mind). Of course some indication of character health needs to be there, since we have no connection to the character in that sense. I recall a game whose name escapes me that would give direct feedback as to how well you held the bow by how tired you got holding the bowstring and how much it wobbled whenyou did. Perhaps clues like that. But, this is not that sort of game now then is it? This is no witcher.

 

Again, I prefer skill based systems in table top RPGs. The inner workings of some of which the player should never know. Good way to shoot the rules lawyers in the foot too

[Gatt9]

Overly complex math for the sake of having overly complex math isn't good game design. At most, a check should be a d100 roll with a Character's skill/modifiers added to it compared to a difficulty value.

 

Anything beyond that is excessive, and starting to enter Simulation territory by trying to model unneccessary (In this context) variables.

 

Going beyond that, requiring players to have advanced understanding of probability and math starts limiting your potential sales very rapidly. Once you get out of basic algebra land you're going to start limiting your potential market because you're demanding the Player have taken increasingly advanced math courses and/or sit there and calculate the probability that an 1/8" thick alluminum lockpick used on a lock manufactured from damascus steel would work, versus using the 1/4" alluminum lockpick. This is one of the major reasons why Wargames are a dead genre, because the Player was expected to have intimate knowledge of units and their qualities to the point where it ended up exceeding all but the most die-hard fans.

 

That type of stuff is fine in games aspiring to Simulation, but for something much more high level, a D20/D100 modified roll against a difficulty is plenty, with modifiers being just one number (Read: 3.x edition D&D style)

[Grand_Commander13]

How the heck do you get a tabletop RPG where people don't understand the skill system? That seems patently ridiculous, since everyone's read the same rule book.

 

Anyway, I am glad to say you likely won't get your wish: people should easily be able to look up exactly how the game's math works whether Obsidian tries to obfuscate it or not because it will all have to be easily addressed for modding. The only question is whether it will be simple enough for guide writers to give a formula that can be intuitively understood or if they'll need to include a graph.

 

I've played games where there was no way but trial and error to tell whether something was a good or bad idea; they aren't fun at all. Thanks to Project Eternity's focus on modding the real numbers for the game will come out quickly, but if the formulas are bizarre and hidden then lots of false ideas will take hold and people like me will have to spend time conducting tests for our guides when we should be able to just refer to a formula.

[teknoman2]

rule no1 of good design: keep it simple!!!

the more complicated you do something, that harder it is to do it, and the harder it is to use by the guy who will use it

it's not about math , but here's a design complication example that existed in some old games (maybe not as exaggerated and is unneeded: to loot, click on the corpse, press M for your character to move there, press L to open the loot menu, click on the item you want to loot, press space to loot it and press enter to confirm that you really want to loot it... then press esc to exit the loot menu.

if you think "why do all this when 2 simple clicks should be enough?" then think " why should they implement a system that you need a PhD in math to understand?".

an example would be "the lock strength is 25, my lockpick skill is 26, i have 100% chance of opening the door. if my lockpick skill was 5, i would have 5/25 or 20%chance to succeed". simple and efficient way to do it.

or my attack skill (the sum of all my stats, equipment bonuses, buff bonuses etc) is 17, the enemy's defense skill (the sum of stats, armor, buffs etc) is 15, i hit him with every attack. my attack skill is 10, his defense skill is 15, i have 10/15 or ~68% chance to hit him.

in the background there will obviously be some more or less complex formulas that calculate the attack (ie "attack skill = ((strength/2+dexterity)*weapon training-armor penalty+weapon enchantment+(buff-debuff)/4") and defense skills of each character, but knowing them is not necessary to the player, unless he plans to make mods. the player only needs to know the result of that equation.