Often we here on the internets take pride in our ability to do math. A lot of it. Often we will look at things in terms of DPR, a statistical figure calculated to determine what kind of damage output one can expect from a character's attacks.
When considering the implications of all of these calculations and the assumptions that have to be made just to make them, is it worth it to even consider this figure over actual gaming experience?
Do theory crafted calculations turn out to be more important when making a character than your experience at the gaming table?
| 1 person marked this as a favorite. |
Umm... their two completely different things. Actual game play experience differs to wildly outside of PFS play due to houserules, so the the best way to determine how the game actually works is by calculations. Especially since those calculations will apply to the people who play PFS or run their home games under similar RAW. So on the whole? The calculations are much more valuable for determining balance then actual play experience. Number don't lie. Experiences do.
What anzyr said because that is just solid logic and how the world works. You can figure everything out with calculations and formulas. Its why we can always mathematically determine who is going to be the next president in 3 years or how many murders there will be next year.
I knew (along with anyone else who was paying attention) with an incredibly high certainty that Obama was going to win the election (thanks to Nate Silver's site running the numbers). Because statistics do in fact work. The problem you are having MattR1986 is that your experiences are lying to you, just like all the Republicans who came up with a whole bunch of reasons why the statistics projected were wrong (even though obviously they were right). People deceive themselves all the times to try and make things match their expectations/wishes/etc., and that why you can't trust experiences when trying to determine objective truths.
| 3 people marked this as a favorite. |
Probabilities and math give just a baseline as to a game mechanic's parity. If it's wildly out of whack with the rest of the game rules, you can usually find out in the math before wasting time in gameplay.
Game experience will tell you if a mechanic "feels" like it should, or how often the mechanic will actually come into play vs how impacting it is, or if it steals the spotlight too much from something else.
When someone talks about "balance", they are (or should be) talking about the math. You can strive for balance in the math, but still look for unique gameplay experience across each and every game mechanic.
A longsword and axe are functionally balanced, but feel different in gameplay. This would be the simplest example of where math and gameplay can be used as two different tools for the final goal: a good game.
| 1 person marked this as a favorite. |
Theorycrafting says that my fighter is a weak character class choice.
Theorycrafting says that my friend's wizard is more powerful than my fighter.
Theorycrafting says our party rogue is worse at skills than our bard, and worse in a fight than our paladin.
My fighter regularly out damages the paladin.
The rogue has a different enough set of skills than the bard that their areas of expertise don't cross often enough for it to matter.
The rogue gets more kills than the paladin.
The wizard is enough of a generalist that his powers don't outshine the rest of the party.
All of our characters are specialised, but not particularly optimised (except the wizard, I guess, who spreads his spells across offence, defence, and utility).
Theorycrafting has it's place. That place is "long term statistical trends". And I have no doubt that it does it very well. As long as actual play experience is fun for everyone involved, though, and as long as everyone follows Wheaton's Law, theorycrafting doesn't actually matter once the dice start rolling.
Case in point: I had a switch-hitting fighter, once upon a time, called Ryn. He was good at what he did. But he wasn't great. One of the other players (who is no longer in our group) decided to bring along a character who could (and I quote) "Out-Ryn Ryn". It was a theory crafted optimised switch-hitting ranger build.
When this character first came out, I shrugged - if someone else wants to bring along a switch-hitter like mine, I'm not going to be bothered by it. As the weeks went by, though, it became clear (and abundantly so after the "Out-Ryn Ryn" comment) that the entire point of the character's build was to one-up me. I objected. The player (thankfully) switched character.
The point is that I didn't object to a theory-crafted character at the table. I didn't even object to the fact that the character was more combat capable than my fighter. I objected to the idea that he was specifically designed to be better than my fighter at the things my fighter was designed to be good (but not great) at.
|
What anzyr said because that is just solid logic and how the world works. You can figure everything out with calculations and formulas. Its why we can always mathematically determine who is going to be the next president in 3 years or how many murders there will be next year.I knew (along with anyone else who was paying attention) with an incredibly high certainty that Obama was going to win the election (thanks to Nate Silver's site running the numbers). Because statistics do in fact work. The problem you are having MattR1986 is that your experiences are lying to you, just like all the Republicans who came up with a whole bunch of reasons why the statistics projected were wrong (even though obviously they were right). People deceive themselves all the times to try and make things match their expectations/wishes/etc., and that why you can't trust experiences when trying to determine objective truths.
Tell that to Al Gore! The problem most people have with statistics in the political arena is that we have years of experience from before the last election where the number crunchers were wrong. That is because, at least when it comes to politics, many studies are purposely biased.
The question thus becomes not "do you trust numbers" but do you trust the person or organization that is pushing those numbers.
In LA we have a joke that you need to subtract at least 5 points from the democratic/liberal side of any LA Times poll. Can't even tell you how many times their numbers are just wildly wrong.
antecdotes are not data,
but all these antecdotes say you are wrong.
yay for old arguments made fresh.
honestly both are important. without the math your start and end points are going to be crap.
without actual play experience you will not know which parts of the mechanics will actually work as part of the game.
this is why paizo wants both as part of their development program.
And as to the actual purpose of the thread. Statistical analysis is helpful for deciding on feats and builds, but I pick more based on what I like to do as a player/character than what the numbers will say is optimal. Sometimes those things match up, sometimes they don't.
and you know what they say about numbers: probability says I will roll a 20 about every 20 times I roll. But I could roll a "1" a million times in a row. So the most optimal build "should" outperform a sub-optimal one but in real play it won't necessarily do so in any particular game.
The plural of "anecdote" is not "data".
There are certain things the math can't model, but can only make assumptions about. In real play, some of those assumptions may prove inaccurate, and you learn something important and may need a better model next time. So the answer is it depends on exactly how solid your math is. Someone good at the math should be fully aware of the limitations of their model and possible weaknesses that need to be tested with real data.
If someone is telling you in the real world that dice behave differently than expected, then the problem is likely with their own unreliable perception of reality rather than with reality. Or the fact that the sample size is inevitably too small to outweigh the randomness. People aren't perfect at observing patterns, because they sometimes like to see them where there aren't any. But yes, most people should be aware that there are some things to be learned about what will actually happen during a play session that you can't just calculate in a vacuum.
| 1 person marked this as a favorite. |
I prefer actual gameplay results. However, the reason is not that I don't want statistics. The reason is that the forum has so many statistics, and so much attention is given to statistics, that gameplay results often seem to not have a place.
Statistics give me an idea of how something could be used or abused in an abstract environment. They makes assumptions on the conditions of play, so that all can be constantly judged against the same conditions. I see some value in that. But, how often does actual gameplay adhere to those conditions? Not often, so I want both.
Theorycrafting says that my fighter is a weak character class choice.
Theorycrafting says that my friend's wizard is more powerful than my fighter.
Theorycrafting says our party rogue is worse at skills than our bard, and worse in a fight than our paladin.My fighter regularly out damages the paladin.
The rogue has a different enough set of skills than the bard that their areas of expertise don't cross often enough for it to matter.
The rogue gets more kills than the paladin.
The wizard is enough of a generalist that his powers don't outshine the rest of the party.
All of our characters are specialised, but not particularly optimised (except the wizard, I guess, who spreads his spells across offence, defence, and utility).Theorycrafting has it's place. That place is "long term statistical trends". And I have no doubt that it does it very well. As long as actual play experience is fun for everyone involved, though, and as long as everyone follows Wheaton's Law, theorycrafting doesn't actually matter once the dice start rolling.
Case in point: I had a switch-hitting fighter, once upon a time, called Ryn. He was good at what he did. But he wasn't great. One of the other players (who is no longer in our group) decided to bring along a character who could (and I quote) "Out-Ryn Ryn". It was a theory crafted optimised switch-hitting ranger build.
When this character first came out, I shrugged - if someone else wants to bring along a switch-hitter like mine, I'm not going to be bothered by it. As the weeks went by, though, it became clear (and abundantly so after the "Out-Ryn Ryn" comment) that the entire point of the character's build was to one-up me. I objected. The player (thankfully) switched character.
The point is that I didn't object to a theory-crafted character at the table. I didn't even object to the fact that the character was more combat capable than my fighter. I objected to the idea that he was specifically designed to be better than my fighter at the things my fighter was...
All I got out of this was "And then I realized those theorycrafters who said Fighter was weak really knew what they were talking about. Amazing how can they figure that out just by running the numbers and not needing to have my actual experience to draw on."
@ Rub-Eta: Having all the numbers might be tricky in real life, but in Pathfinder the number of variables is significantly smaller. Also, we don't need to quantify abstract concepts like love in PF. Just AC, to hit, skills, damage, HP, etc.
@ Pyrrhic Victory: Well yes you need to find people who won't lie to you, that helps. But numbers themselves are never wrong. And luckily its much easier to compare PF numbers then political polls.
@ Abraham spalding: Not really. Actual play experience won't help determine if something is balanced at all. At best actual play experience can only tell you how "understandable" the mechanics are. It won't actually tell you anything else that theorycraft can't.
not what i was trying to say.
the mechanics and balance can be fine but also be boring or onerous or simply not fun when used at the table. those are the points that playtesting can help with.
as such for me playtesting is as much for the... fun aspect as it is for the number crunching. remember the gunslinger was originally a fighter subclass. it worked as such and balancewise not much changed but making it a separate class was definitely a good move.
What anzyr said because that is just solid logic and how the world works. You can figure everything out with calculations and formulas. Its why we can always mathematically determine who is going to be the next president in 3 years or how many murders there will be next year.I knew (along with anyone else who was paying attention) with an incredibly high certainty that Obama was going to win the election (thanks to Nate Silver's site running the numbers). Because statistics do in fact work. The problem you are having MattR1986 is that your experiences are lying to you, just like all the Republicans who came up with a whole bunch of reasons why the statistics projected were wrong (even though obviously they were right). People deceive themselves all the times to try and make things match their expectations/wishes/etc., and that why you can't trust experiences when trying to determine objective truths.Tell that to Al Gore! The problem most people have with statistics in the political arena is that we have years of experience from before the last election where the number crunchers were wrong. That is because, at least when it comes to politics, many studies are purposely biased.
The question thus becomes not "do you trust numbers" but do you trust the person or organization that is pushing those numbers.
In LA we have a joke that you need to subtract at least 5 points from the democratic/liberal side of any LA Times poll. Can't even tell you how many times their numbers are just wildly wrong.
It's still closer to accurate than the New Yorker film critic who was said to have proclaimed "I only know one person who voted for Nixon", in the year Nixon won 49 states.
| 1 person marked this as a favorite. |
First, neither is really important. This is a game. =)
Second, both have value to the decision process.
The statistics have value. But only to the degree that the number crunching is accurate, the time the cruncher wants to take, how much the cruncher really understands statistics.
Most of the numbers I've seen on these forums really show that the person doing the crunching really doesn't understand probability at all. There are some that do understand it, really put in the time and effort, and check their work. But that is a fairly rare combination.
In addition there is no real way to accurately take into account all the 'other stuff' like tactical acumen, working as a team, differing goals, hostile GM, etc...
There was a guy at a PFS event a while back crowing about how his build was gonna do 1.3% more damage than anything else. He had less than a full page of hand written calculations that he spent maybe 10 minutes writing down.
All that other stuff that isn't controlled or accounted for has a much larger swing than his 1.3% improvement.
Quick glance showed at least 3 math errors right off.
The number crunching is really only useful to the extent of broad strokes. "Build X could reasonably expect to do approximately the same damage as build Y, if the circumstances don't favor one over the other."
That's about as definite as you can get.
Personal experiences are valuable if you feel you can assume the other persons circumstances are close enough to your own to be meaningful.
|
*** The calculations are much more valuable for determining balance then actual play experience. Number don't lie. Experiences do.
This. Play experience is always anecdotal. If Joey just has bad luck with dice, it might seem like his Paladin is just a terrible character. Joey is going to think that since his Paladin can never hit anything, that Paladins are bad. People are wired to see fault in themselves and will almost always externalize blame, so Joey's experience is not a good measure of the Paladin class' strengths or weaknesses. Looking at the math behind the Paladin can give us an idea of what's going on with the class. If it turns out that the probability of Joey rolling that poorly was roughly 10%, than we know that the issue lies with Joey or his dice, not with the class.
That's not to say that there aren't some factors that are hard to quantify but easier to evaluate through play experience. For example, a damage dealing class with extremely high mobility might be functionally more powerful than the base math would indicate. If, for example, the class has a teleportation based movement mode that allows him to quickly and efficiently set up flanking and gain bonus damage and accuracy, that might shift the numbers a bit. Functionally though, that variable can still be accounted for mathematically. We just factor in that functional bonus to hit and damage as being a more reliable effect with a greater chance of occuring and adjust the projection accordingly.
Play experience can help expose flaws in the mathematical projections of a class, but anecdotal evidence is not in any way a gauge of a class' actual power or effectiveness.
It ties into another discussion I've had about how a class can have the capability of dealing a thousand points of damage a round with great defenses and still not be a good class. Does it have a reliable way of delivering that damage? What are the class' capabilities other than withstanding a hit and dealing out a truckload of damage? Playtesting can reveal that something that appears valuable on the surface actually has a deeply inherent flaw, or that something that was meant to be a bit of fluffy fun is actually an incredibly potent option, but that play experience isn't undermining or devaluing the math; it's revealing components of the equation that you may not have accounted for initially.
RPG Superstar 2012 Top 32
|
| 1 person marked this as a favorite. |
One gamer points to a sheet of maths and says "This is how things will typically play out."
Another gamer points to last week's game and says "This is how things will typically play out."
Then we all try to decide which is more valuable in figuring out how things will typically play out.
One gamer points to a sheet of maths and says "This is how things will typically play out."
Another gamer points to last week's game and says "This is how things will typically play out."
Then we all try to decide which is more valuable in figuring out how things will typically play out.
I'm sure there's a step about ignoring numbers or misinterpreting/misreading in there somewhere. I could swear that's a step in there.
he does not need numbers and facts he has the truthiness in his gut.
One gamer points to a sheet of maths and says "This is how things will typically play out."
Another gamer points to last week's game and says "This is how things will typically play out."
Then we all try to decide which is more valuable in figuring out how things will typically play out.
Last Week's Game is too small a sample size to draw any conclusions.
|
RPG Superstar 2012 Top 32
|
I'm sure there's a step about ignoring numbers or misinterpreting/misreading in there somewhere. I could swear that's a step in there.One gamer points to a sheet of maths and says "This is how things will typically play out."
Another gamer points to last week's game and says "This is how things will typically play out."
Then we all try to decide which is more valuable in figuring out how things will typically play out.
It's assumed within all three sentences. :)
| 1 person marked this as a favorite. |
Neither is more important, ultimately, for game design, but they tend to be most useful for different questions. Statistics may tell you whether or not something has the same mechanical power as another option, but they can't tell you whether it's usable or cumbersome in play. They also can't tell you whether or not a trade-off between power balance and usability is worthwhile from a player's perspective. Those questions require gameplay experience.
"There are lies, damned lies and statistics." -Mark Twain
The fact is numbers can very easily be lies and deceiving. Don't let someone try to fool you into the logic of their "empirical data" when their statistical quantification is as complex as 5th grade math.
Right. Except when you can compare them one to one. Which we can do in Pathfinder and it isn't even hard.
| 2 people marked this as a favorite. |
I've got a PC I play regualrly. He's a Bard/Ranger/Witch. As you can guess, he's terribly unoptimized. His DPR is abysmal. His abilities are Str 10, Dex 13, Con 11, Int 16, Wis 16, Cha 14. He's taken permanent Con drain.
But you know what? He's the character that frightens all the other players. Not because he'll beat the players, but because he's so resourceful and unpredictable that nobody's really sure what he'll pull off, next. He ends up collecting lots of odds and ends, and disseminates what the party needs precisely when they need it. Everyone is convinced he's capable of far more than they've seen him do, and he's pulled the party out of a jam through quick thinking and unconventional tactics many a time. From stymieing a ghost with a rope, to burning down a mansion with a well-placed spell, to bartering a game of souls with a vampire, or answering a ton of riddles in succession, he is consistently a force to be reckoned with.
Part of the reason he is so successful this way, though, is because he isn't focused. He doesn't have all his eggs in any one basket. Sure, I imagine retraining him properly in having a better set of capabilities on paper, but the truth is, his madcap force-of-nature style is just what the party needs, no matter what it costs in traditional measurements of effectiveness.
The way I see it, there are two measures of game skill: build skill and play skill. Both are important, and one does not preclude the other. Build skill is where you'll see the math shine. That's how you'll make an effective character, under the assumption of knowing what it is you'll face. Build skill is key to being able to do what you want.
However, there's also play skill. Do you know how to set up your flanks while minimizing AoO? Which square is the optimal place to cast the Burning Hands from? Is that pillar in the middle of the battlefield beneficial or a hindrance? And you can only test play skill through actual experience.
And yet, there's a second type of play skill--story skill. Yes, you know the NPCs, the lay of the land, distances to get from point A to point B. But can you get the NPC's army to support your efforts? Can you set up the ambush in the correct pass? Can you figure out when the spy will strike?
If you've got low build skill and play skill, you'll die fast--like the one elven wizard with a 6 Con who was first over a barrier into a band of orcs everyone knew were there. If you've got high build skill but low play skill, you'll wonder why you're having trouble setting up your nova, or you'll find yourself in a situation with fewer resources than you expect and crumble in fear. High play skill can overcome low build skill--like my character above--but it's a hard, painful, choose-your-battles-wisely road.
So both the statistics and the anecdotes are important. The statistics show how to set up a build, but the anecdotes showcase a player's ability to succeed with whatever he's got, not whatever he's planned. If you stick with one, you'll find problems you didn't prepare for. You need both statistics and anecdotes to know whether you're accomplishing what you hope to.
He's arguing with what he thinks you said.
He's saying that numbers can lie and be deceiving. I'm saying that is true, unless you compare, say AC 35 to AC 44. Those numbers don't lie in the context of PF. AC 44 is simply better. Unless he was saying "Yes, we can use numbers to reach truthful conclusions" I don't think I'm arguing with what I think he said, but rather with what he actually said.
If our characters were identical fighters, except that you had Greater Weapon Specialization and I'd opted for Blind-Fight, you would have the numerically higher number: but I'd be hitting more often (and thus doing more damage) if we both had to fight a bunch of monsters in the dark.
I think that's drawing too much from two sentences myself.
Besides, we both know both of the half-orc fighters had dark vision anyway.
Where is mplindustries to tell us all how vital strike is a mathematically weak feat at best? Ah these boards are a revolving door of annoying stereotypes whose names change but whose personalities remain...or get picked up by someone else.
Where is mplindustries to tell us all how vital strike is a mathematically weak feat at best? .
What's wrong with someone stating facts with contingency statements for variables?
|
|
Considering all the claims made on gaming websites that insist they can use a few equations to prove beyond a shadow of a doubt that "X class is overpowered, Y class, is objectively inferior, Z class can be played in manner A or manner B and anything else is stupid, blah blah blah," the fact that a lot of the people who do this can't even agree with each other, and that I tend to see all these claims slashed to ribbons every time I actually go play (go ahead - tell the frequent-Fighter-player and frequent-Rogue-player I play with that their classes suck), I can say with utmost certainty that the answer is "no."
Math is obviously very useful, but Pythagoras was wrong - the Secrets of All Things are no more encoded in math than in any other language (yes, it's supposedly the "Universal Language" through which Terrans will no doubt establish First Contact, as beautifully-demonstrated in Carl Sagan novel-turned-Jodie Foster movie Contact, but that's precisely because it's artificial - and even then, this assumption led to a bit of a letdown when the first pulsars discovered originally inspired hopes of having found a signal from intelligent extraterrestrial life, so regular were its signals). With regards to gaming, "probabilities and projected statistics" are useful within reason - I was able to use it to devise and verify a revised DC formula that made the 3.5 Tome of Magic's Truenamer satisfactorily effective. However, some people appear to have become trapped in a narrow mental groove, which is the farthest thing from what RPGs ought to do.
Often we here on the internets take pride in our ability to do math. A lot of it. Often we will look at things in terms of DPR, a statistical figure calculated to determine what kind of damage output one can expect from a character's attacks.
When considering the implications of all of these calculations and the assumptions that have to be made just to make them, is it worth it to even consider this figure over actual gaming experience?
Do theory crafted calculations turn out to be more important when making a character than your experience at the gaming table?
I think they each have their place - it may well be that one build has a higher DPR than another but the second is more fun to play. That doesnt mean that either consideration is less useful, just that they're answering different questions. (One about game balance or perhaps utility, the other about what you should play if you want to have fun).
One thing I've never seen addressed (although presumably it has been discussed amongst theorycrafters many times) is that DPR is usually presented as a mean number and doesnt capture the spread of damage. For example: an attack doing 5-16 (4+1d12) is better against monsters with higher hit points (say 15 average) whereas one doing 9-12 (8+1d4) is better against monsters with low hitpoints (say 8 average), even though they have equal average damage. Similarly if fighting a horde of low hit point creatures you'd prefer to hit often for not much than rarely for a huge amount - even if the DPR were identical.
I'd be curious if there are some other statistics around besides DPR which capture this sort of thing.
Considering all the claims made on gaming websites that insist they can use a few equations to prove beyond a shadow of a doubt that "X class is overpowered, Y class, is objectively inferior, Z class can be played in manner A or manner B and anything else is stupid, blah blah blah," the fact that a lot of the people who do this can't even agree with each other, and that I tend to see all these claims slashed to ribbons every time I actually go play (go ahead - tell the frequent-Fighter-player and frequent-Rogue-player I play with that their classes suck), I can say with utmost certainty that the answer is "no."
Math is obviously very useful, but Pythagoras was wrong - the Secrets of All Things are no more encoded in math than in any other language (yes, it's supposedly the "Universal Language" through which Terrans will no doubt establish First Contact, as beautifully-demonstrated in Carl Sagan novel-turned-Jodie Foster movie Contact, but that's precisely because it's artificial - and even then, this assumption led to a bit of a letdown when the first pulsars discovered originally inspired hopes of having found a signal from intelligent extraterrestrial life, so regular were its signals). With regards to gaming, "probabilities and projected statistics" are useful within reason - I was able to use it to devise and verify a revised DC formula that made the 3.5 Tome of Magic's Truenamer satisfactorily effective. However, some people appear to have become trapped in a narrow mental groove, which is the farthest thing from what RPGs ought to do.
I've never seen people who argue against generally accepted information that people with system mastery agree on using anything that be considered an objective viewpoint in even the vaguest sense. All your observation indicates that there are people who argue against the truth and there's a lot of those running around real life, but that doesn't change the scientific truth no matter how loudly they talk. So much like any respected biologist will tell you that evolution is totally a thing, any respected person with system mastery will tell you that Fighters and Rogues are incredibly underpowered and that full casting is the strongest thing in the game. Why? Because its mathematically true.