It is not so much that numbers are more important, but many of us play differently, and the numbers are not biased like our experiences can be. In addition the numbers help us to find what is possible, even if we don't use them to their full advantage at the table.
|
| 2 people marked this as a favorite. |
I'm a rocket scientist, specializing in deterministic and non-deterministic M&S (Models & Simulations) of missile systems from the sub-system level up to the architecture level. When we prepare for a flight test (where we live-fire an actual, instrumented missile), we spend months simulating the holy hell out of that scenario and crunch, crunch, crunch data to extract our best guess at how things are going to work. I've never worked a flight test (in nearly 20 years) where we hit all of our predicted metrics.
AFTER the test, we go back and plug in all the measurements and initial conditions we measured and rerun our sims. We can recreate a flight test with fairly accurate results. This tell us our models are valid. All the models are really good for is bounding our expectation. What actually happens when the rubber meets the road is usually within the boundaries, but nailing down the exact outcome before the test is very difficult.
I kinda look at theorycraft the same way. If your houserules and the other players are your initial conditions, you can use theorycraft to bound your result, but it's not going to reliably give you an exact or near-exact prediction of what will happen in your game. A rule bent or forgotten, a monster played outside expectation, or an ally that takes an unorthodox course of action, all these things can affect the results.
Personally, since my job amounts to "do math all day," I try to do as little as possible in my hobby. In other words, I don't theorycraft, I just play and have fun.
-Skeld
They're both important, but they're bot situational.
"Man, weapon focus bites, i never miss by just one..." is something that can be covered by statistics. How much its increasing your damage is determined, over the long haul, by math.
"Wow, rogue sneak attack hurts..." is something you need play experience with. No formula is going to tell you how often you sneak attack or full attack sneak attack.
I'm a rocket scientist, specializing in deterministic and non-deterministic M&S (Models & Simulations) of missile systems from the sub-system level up to the architecture level. When we prepare for a flight test (where we live-fire an actual, instrumented missile), we spend months simulating the holy hell out of that scenario and crunch, crunch, crunch data to extract our best guess at how things are going to work. I've never worked a flight test (in nearly 20 years) where we hit all of our predicted metrics.
AFTER the test, we go back and plug in all the measurements and initial conditions we measured and rerun our sims. We can recreate a flight test with fairly accurate results. This tell us our models are valid. All the models are really good for is bounding our expectation. What actually happens when the rubber meets the road is usually within the boundaries, but nailing down the exact outcome before the test is very difficult.
I kinda look at theorycraft the same way. If your houserules and the other players are your initial conditions, you can use theorycraft to bound your result, but it's not going to reliably give you an exact or near-exact prediction of what will happen in your game. A rule bent or forgotten, a monster played outside expectation, or an ally that takes an unorthodox course of action, all these things can affect the results.
Personally, since my job amounts to "do math all day," I try to do as little as possible in my hobby. In other words, I don't theorycraft, I just play and have fun.
-Skeld
I think with your level of experience you will agree with me then when I say the number of measurements and conditions that one needs to simulate in pathfinder are significantly smaller then those required for real world use. And thus much easier to get correct a significantly larger portion of the time.
Statistic measurements are useful, as noted upthread, for baselines and certain general evaluations.
But you cannot (and the vast majority of theorycrafters utterly and totally fail to) properly or adequately account for the massively chaotic variables brought in by the human element--which unfortunately for the theorycrafters, is kind of necessary to make the games happen.
Innumerable variances in player skill, tactics, experience, willingness and ability to cooperate, and in GM skill, tactics, experience, and sense of fairness, and overall types of play styles, campaign styles, campaign settings, encounters designed, downtime permitted, houserules used, etc. etc. etc. etc. are going to make theorycrafting only valuable in a very limited way. Players and GMs, ones at least good at what they do, have to make a lot of judgement calls based on the situation they are in that stats-crunching is not going to help them with at all. Those judgement calls often have to draw from actual experience and from intuition, not just from system mastery.
And when I choose my group, I am going to prefer to be with a group of people who I trust to be skilled at making those judgement calls, not with a bunch of number crunchers---the number crunching skill comes with time and we've usually got at least one person really good at it to help guide us when it's actually useful, but the ability to adapt to actual human play is not a gift all gamers have, and those without that gift I have no desire to play with whatsoever.
I also think the varience is really high in what can happen.
Sometimes they both agree like power attack with a two handed weapon being extremely useful agianst constructs with damage reduction you cannot bypass.
| 1 person marked this as a favorite. |
I generally take "theory crafting" more seriously than gameplay experience. That said, I read and appreciate both. Why?
When I read theorycraft I can backtrack the numbers. I can look over the builds, I can consider them on their merits (or lack thereof). I can see for myself whether or not I agree with the methods used to reach the conclusion the theorycrafter has arrived at. I also generally find that Theorycraft generally try to encompass more options in their comparisons, while gameplay experiences can and will vary (at times dramatically) based on the person playing. This is not surprising since the Theorycrafter generally has more time to consider what to do, while the person at the table is working in a fairly narrow timeframe.
That said, this can also lead to an "armchair general" state of mind where WBL translates to perfect gear for the character build, that you will have unlimited downtime to craft your own items, or that the wizard always has the right amount of the right spell memorized for the situation, no matter the situation.
Let's consider an example where a personal experience and theorycraft disagree. Pirate did a DPR comparison of weapons here. For the sake of the argument, let's assume his numbers are correct. The best weapons for melee full attacks from a DPR standpoint are the falcata, the elven curve blade, the scythe, and the falchion. Out of the top 25 weapons, only the scythe has the option to deal piercing damage. Similarly only the Earth Breaker deals bludgeoning damage. Extrapolating from these numbers we can conclude that slashing weapons are generally superior for doing full attack damage.
My party is currently in a Skull & Shackles campaign, and were originally equipped with slashing weapons. The fighter was TWFing with kukris, the rogue had a greatsword, and the ranger carried a dwarven waraxe as a backup melee weapon for his primary weapon, a longbow. However, a significant amount of the fighting has taken place underwater, where slashing/bludgeoning weapons deal half damage and bows are essentially useless. My party adapted by equipping some spears they found early on, and the majority of their damage output is now done using piercing weapons such as spears, tridents, daggers and so on.
I have gameplay data that can be used to conclude that spears do more DPR than falcatas. Does my personal anecdote invalidate Pirate's calculations? Of course not. My campaign has an unusual factor in play (underwater combat) that Pirate is unable to accomodate for since he doesn't know my campaign, nor is he crunching numbers for my campaign - it is meant as an observation on the DPR of weapons without outside influence. Other factors could be a campaign heavily focused on destroying skeletons (Bludgeoning weapons leap ahead) or campaigns where carrying martial weapons is a capital offense (leading to the rise of Simple weapons).
These are examples with an obvious factor in play and I don't think many people are going to disagree with my conclusion that my anecdote does not invalidate Pirate's theorycraft. Well done theorycraft can comment on the state of the game. That said, Individual campaigns can and will introduce factors that can throw theorycraft off. That does not invalidate the value of theorycraft.
Next, let's consider a different example based entirely on gameplay experience.
Theorycrafting says that my fighter is a weak character class choice.
(...)
My fighter regularly out damages the paladin.
(...)
All of our characters are specialised, but not particularly optimised(...)
Again, let us assume that Chemlak's basic promise is correct - that fighters regularly outdamage paladins.
When I see observations like the one listed above, I don't have the luxury to look over the math and method used to reach the conclusion. "The fighter regularly out damages the paladin" - why is that?
It could be because the party is primarily fighting neutral opponents and the paladin rarely gets to use Smite Evil?
It could be because the fighter player is generally better at optimizing character builds?
It could be because the paladin player is having a streak of bad dice rolls?
It could be because you have a set of houserules in play that benefit the fighter more than the paladin?
In this case I have a sample size of 1 (Chelax) and no way of knowing how he reached his conclusion other than his personal opinion, and no idea what campaign-specific factors were in play. He might very well be correct, but I have no way to reflect over or validate (or for that matter, invalidate) Chelax's argument. I learn nothing from his statement other than that in his party, the fighter does more damage than the paladin.
This is not meant to be insulting or dismissive to Chelax's opinion - his post is a convenient since it was made in this very thread and I read it immediately before making this post.
Finally, let's consider a third personal anecdote where gameplay data and theorycraft agrees:
My 4th level party recently fought two weresharks. Two players were unable to attend the session, so the party was made up of a fighter dualwielding kukris (1D4+7/1D4+5), a switch hitter ranger using a bow (1D8+2), and a rogue fighting with a greatsword (2D6+7).
No one in the party had silver weapons.
The fighter was unable to penetrate the DR 10/Silver of the weresharks on his individual attacks. The ranger ran into similar problems with his bow. The only character able to significantly hurt the weresharks was the rogue due to the advantages of fighting with a two-handed weapon. The fighter was reduced to providing flanking and using the Aid Another option to maximize the chance the rogue could hurt the weresharks.
Based on my personal anecdote, I conclude that THF is better than TWF or archery for penetrating Damage Reduction.
A person theorycrafting the same topic would most likely agree that THF is a better option than TWF for penetrating DR, but point out that Clustered Shots actually makes archery the best method - not THF.
If I were to read two posts arguing that THF is better than TWF for penetrating damage reduction I would most likely value the Theorycraft post over the Personal experience post both since I can track the numbers and consider them for myself, but also because it acknowledges options that might not have been relevant in the personal anecdote.
|
|
I think with your level of experience you will agree with me then when I say the number of measurements and conditions that one needs to simulate in pathfinder are significantly smaller then those required for real world use. And thus much easier to get correct a significantly larger portion of the time.I'm a rocket scientist, specializing in deterministic and non-deterministic M&S (Models & Simulations) of missile systems from the sub-system level up to the architecture level. When we prepare for a flight test (where we live-fire an actual, instrumented missile), we spend months simulating the holy hell out of that scenario and crunch, crunch, crunch data to extract our best guess at how things are going to work. I've never worked a flight test (in nearly 20 years) where we hit all of our predicted metrics.
AFTER the test, we go back and plug in all the measurements and initial conditions we measured and rerun our sims. We can recreate a flight test with fairly accurate results. This tell us our models are valid. All the models are really good for is bounding our expectation. What actually happens when the rubber meets the road is usually within the boundaries, but nailing down the exact outcome before the test is very difficult.
I kinda look at theorycraft the same way. If your houserules and the other players are your initial conditions, you can use theorycraft to bound your result, but it's not going to reliably give you an exact or near-exact prediction of what will happen in your game. A rule bent or forgotten, a monster played outside expectation, or an ally that takes an unorthodox course of action, all these things can affect the results.
Personally, since my job amounts to "do math all day," I try to do as little as possible in my hobby. In other words, I don't theorycraft, I just play and have fun.
-Skeld
That's definitely true for deterministic modeling, but not necessarily for non-deterministic models, which are just based on statistics. In fact, you can boil down a very complex process into a few statistically modeled variables. While modeling (theorycrafting) might involve a fewer number of parameters, the variability of those parameters and complexity of interaction between can be much greater for the single reason that there are humans in the loop and humans are difficult to predict.
-Skeld
|
I believe what this boils down to is the difference between probability and conditional probability. Namely, the kind of theory-crafting to be found on massage boards usually comes with a certain set of assumptions about the kind of game you are going to play - for example, the amount of gold your character has is assumed to be the one alluded to in the Core Rulebook. I think that by this point, people managed to just about perfect their understanding of the theoretical model that matches this set of assumptions.
However, in any given game, there is a different set of baselines that it follows. It could be incredibly hard to pinpoint them exactly (unless, of course, the players and GM take extra care to consider every single little detail on the table). Sometimes certain rules of the game are changed, and sometimes it's just a question of style - an aggressive group of players who don't like in-game rests would be ill advised to pick many spellcasters since their resources run out rather quickly, and martial classes will consistently over perform. On the flip side of that, the GM also has a huge impact - for example, some GMs like boss fights, others like to use many different creatures with different kinds of attacks, etc. There was even this one GM I knew who'd almost always use a combination of flying, invisibility and ranged attacks against his PCs, making any melee type pretty much useless.
So, what we have is probabilistic projections of how a "universal game of Pathfinder would look". Here we are using simple statistics. However, in each game there is more relevant information to consider, turning what we need to accurately predict it to conditional statistics. As I'm sure everyone knows, there are two quite different things (the chance of getting heads on a coin flip is 0.5. The chance of getting heads on a coin flip that resulted in heads is 1 - very different results).
This, I believe, is the reason that theory crafting does not accuratley perform well as a prediction model. I'm sure, however, that for every home game there exists more accurate models based on added information.
So ultimately, statistics matter more than experience. You just have to be intelligent about choosing the correct model you are using with your statistics.
Stats are great for comparing like things. They don't tell you things like which class is better. With math you can compare fighter and paladin DPR, and that will tell you who has higher DPR in the circumstances assumed. That doesn't tell you which class is better.
People like to make logical leaps like Wizard > Fighter because they can make twinked caster builds that will one-shot 99% of all encounters, completely ignoring the fact that most groups don't want to see that at their table or ignoring campaign realities that make those tactics far less useful than casting haste so that someone can swing a greatsword more.
Things like "That's only your games! Personal experience isn't universal!" are sometimes not valid complaints. For example, many guides will berate feint builds for rogues, because they should just flank. That's a great perspective from a math point of view, but completely forgets that walls and corners exist, or that flanking may be a stupid thing to do. Game experience can make certain "statistical truths" seem stupid.
An experienced player can typically spot weaknesses in theorycraft builds before they are even tested.
For example, using headband of Stat boost (like Charisma) to qualify for a feat tree (or something) that is essential to the build; i.e. a headband of charisma to qualify for eldritch heritage (abyssal) for the inherent stat boost. If someone were to dispel, or sunder, that headband, not only would you loose the charisma boost, you would also lose the benefits of Eldritch Heritage, so you would lose the strength boost too.
Experienced players can often times look at something, and see issues before needing game play tests because they know something will be wrong even before it hits the table. As examples, see the Swashbuckler and Investigator playtest threads; many people spotted issues with class features, without needing to actually build a class and test it. Like the swashbuckler being even more MAD than a Monk, or the Investigator's Studied Combat/Studied Strike being effectively useless as presented unless you maxed Int above all others.
To me, Theorycraft is more valuable than experience because theorycraft is repeatable. Someone having the experience where they roll 5 natural 20's in a row is extremely difficult to repeat, and would skew data with extreme examples.
Theorycraft, on the other hand, represents the expected average that a build will perform to. The only way I would consider experience more valuable than theorycraft, is if the experience were to run a scenario something like 50 times, to see how the build, feat, spell etc, compared instead of just a one-shot lucky natural 20, or unlucky natural 1 or something.
Theorycraft also completely removes the human element from play. If 30 people give you test feedback, with each one having various houserules or forgotten rules, or mis-read rules etc, about a something being designed, are those experiences valuable in the context of the game, considering the game they are playing are different than what the book says?
If you were a scientist and you created a formula to test something, and someone who was testing your formula to confirm your results changed something in the formula and then concluded it didn't work, how does that help you verify your formula? It doesn't; because the other scientist wasn't testing your formula. What if they forgot a step? Does that help? No. What if they misunderstood one of the steps? Does that help? No.
Theorycraft tests the rules without variance, though they can include it if necessary (if flanking then this, if hasted then this, if righteous might then this). That's what it's good for and it more accurately gauges the actual mechanics being tested, than someone playtesting it at a table.
However, I would find experience valuable if they were to outline every minute change to the game they could possibly be using, and then record every aspect of their play for review to see where something worked, and more specifically, why it worked, instead of saying, "It did/not work."
| 2 people marked this as a favorite. |
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.
There is only one Ryn.
I'm going to have to ask you to cease and desist at once.
Sorry mattR, I couldn't resist reworking your sentence for comic effect:
It becomes a problem when people break down these anecdotes as being more significant than they really are. An example of this type of thinking is Bob has 1 cookie and Joe has 2 and Joe says I will ALWAYS have ONE WHOLE cookie more than you! Ya he does at the moment, but in reality it's just standard variance. The anecdote is misleading. Taking these anecdotes as dogma as a baseline for the value of your character is just bad IMO. "You only did 22.5 damage this session? Start over I say". In one session you are unlikely to accurately be close to your 40.2 DPR. Over the course of 4000 rounds? Probably would be pretty close. No one ever accounts for the extreme variability of these anecdotes.
I would say statistically sizing up your numbers with your comparable CR should suffice in making sure you aren't a weenie (if CR is relevant in your campaign)
It becomes a problem when people break down these numbers as being more significant than they really are. An example of this type of thinking is Bob has 1 cookie and Joe has 2 and Joe says I have one HUNDRED percent more cookie than you! Ya he does, but in reality its just one cookie. The statistics are misleading. Taking these DPRs as dogma as a baseline for the value of your character is just bad IMO. "Your DPR is only 22.5? Start over I say". Its just rough, theoretical abstractions of what your character is likely to do as an asymptopic goal over a long period of time. In 4 rounds you are unlikely to accurately be close to your 40.2 DPR. Over the course of 4000 rounds? Probably would be pretty close. No one ever accounts for the extreme variability of these numbers.
I would say sizing up your numbers with your comparable CR should suffice in making sure you aren't a weenie (if CR is relevant in your campaign)
DPR calculations are never taken in a vacuum. They are always then compared to enemy health totals and such.
22.5 DPR is STELLAR for level 4, for instance, since most things won't have more than like 50 HP at best.
It's terrible at level 10, when 100+ HP is commonplace, to say the least.
|
Consider this when trying to determine whether gameplay is more important or theorycrafting with pure numbers is: The Paizo staff has playtests to test the viability of new classes. While these tests are running, they want you to play with their new toys, not just say that it looks weak or strong based on the numbers. If that's all they wanted they'd just do it all themselves. Numbers are important to consider, but so is actual playing experience.
You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells? What about if they rearranged their stats from a 40.2 DPR to 30.8 that seems bad right...unless maybe it was to make sure they had an adequate AC/saves/hp. etc. etc.
22.5 DPR doesn't matter in any of those cases. Its of course, still terrible at level 10 for anyone who's actually trying to do damage. Which when DPR matters.
You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells? What about if they rearranged their stats from a 40.2 DPR to 30.8 that seems bad right...unless maybe it was to make sure they had an adequate AC/saves/hp. etc. etc.
What about when that Rogue rolls 5 natural 20s and happens to roll 3/4 damage on all of his sneak attack dice? Does that mean the Rogue is OP because he can drop a shitton of damage with some lucky rolls?
What about when that Monk rolls absurdly high on his disarm checks and disarms a marilith of all her weapons? Monks OP right?
How about when that Paladin is smiting Renchrazzzerolok, Archfiend of the 7th laywer* of hell and rolls under a 5 on every attack. Paladins suck amirite?
*[Edit] you know, I didn't spot this typo until after I hit submit, but I think it works better this way.
=====================
You keep trying to compare apples to oranges to prove your statement. If you are playing a highly social/political game (ala Song of Ice and Fire), then having a high DPR isn't necessary and a theorycraft character in such a game isn't going to build for high DPR. If one were to bring AM BARBARIAN into Westeros, he's going to feel pretty useless when there are almost zero combat encounters (though he's probably the only character guaranteed *not* to die).
At the same time, if you have lots of SoS spells with high DCs, or utility spells then of course you're not measure DPR, you're a full on f%+@ing caster! Casters don't do DPR, that's what fighters, and barbarians and paladins are for.
When you're measuring DPR, you are measuring the effectiveness of a combat class that uses weapons, not goddamn spells!
One cannot use apples to make orange juice, and then complain the orange juice doesn't taste right.
For what its worth, the game becomes more static as you level because your static modifiers are what increase more than the actual variance. You roll more D6s and have a larger static modifier attached, you don't have a larger D20 with extra numbers. The exception to this is when you don't get to roll, such as whether or not you get a full attack or sneak attack off. Slightly harder to get an average with that, but you can still estimate the value for getting the full attack off every time.
But to continue on. Combining everything into one figure as a baseline can still be very misleading.
"Well even though I have a 20% chance to hit, with my 6d6 damage my DPR is higher than yours!
To be fair, 20% is pretty awful chance to hit and no one does 6D6. They do 3 attacks with a different chance to hit each doing 2D6 + 24 each at 11th, instead of one attack with 2D6 +9 at one.
As much as I hate to bring World of Warcraft into it, this is something Blizzard got right.
Apparently, the majority of their game balancing is performed by looking at in-game statistics rather than mathematical models. They look at damage actually performed, therefore taking into account the things you don't get on spreadsheets, like average player time taken to respond and hit a button and people choosing that power because it has a nice shade of green or is more fun to use, rather than the one that does the most damage. They look at how popular a class is in PvP (if it's too popular, they nerf it because the reason for the popularity is likely that it's too attractive a choice.), they look at the composition of raid groups and times taken to down bosses.
They look at people showing them graphs and figures and spreadsheets and averages and calculations, and give them a metaphorical equivalent of "that's nice, dear."
At the end of the day, if what you're worried about is how it affects the players in the field, then actually examining how it affects players in the field is likely to give far better results than sitting at a computer doing DPS/DPR calculations. Human beings (other than the extreme minmaxers who you should never balance your game around if you want it to sell to the average player on the street) tend to have all different kinds of reasons for making their in-game choices. Some will pick things based on effectiveness, some will choose something because it reminds them of something they saw in a movie or read in a book, some just think an option sounds cool.
If you're worried about how the game will be when played by computers and robots, then by all means bring out the math.
MMOs work a bit differently than games Table Top though. The classes that can't keep up are liable to stop being invited to raids and thrown into the trash can as far as players are concerned. I know I've seen that happen quiet a few times and in several games I've played. They also tend to be weighed mostly in DPS rather than utility, not that they can't work otherwise or that table tops can't be weighed totally that way.
Edit: MMOs can also be patched and hotfix and everyone is going to have to adhere by that, unlike tabletops which are an abominable mess of errata you may never see and every house can have its own rules. This is a boon and a bane mind you, but I'd say its mostly a boon. It does create a bigger urgency for balance though, and any needed changes are much more likely to come rather than thrown under the bus, though this varies with company.
Occasionally you see an outlier utility need such as a mage's ability to teleport making a fight so much easier or a particular tanks ability to blockade in a certain way, but that's an outlier usually and not the best way to plan things because it enforces and idea that players should hunt that down and shun others for not being it.
Another thing is cookie cutter builds. You tend to have a wider variance of builds in a game like Pathfinder than you will in most any MMO. Things also tend to be streamlined in an MMO compared to a table top. I'd say this is flavor rather than good or bad, but its much easier to care when every arms warrior is doing roughly the same when doing their rotation properly rather than a tabletop where there is a massive variance in what attribute distribution, feat choices, race, etc. they have.
At the end of the day, if what you're worried about is how it affects the players in the field, then actually examining how it affects players in the field is likely to give far better results than sitting at a computer doing DPS/DPR calculations. Human beings (other than the extreme minmaxers who you should never balance your game around if you want it to sell to the average player on the street) tend to have all different kinds of reasons for making their in-game choices. Some will pick things based on effectiveness, some will choose something because it reminds them of something they saw in a movie or read in a book, some just think an option sounds cool.
Well something like wow has the advantage of both storing the data for later use, AND having an enormous sample size to pull from.
The downside for most humans looking at the games are small sample size and sampling errors. You're really going to remember the one time that rogue popped off a full attack sneak attack and one shotted the boss. You're not going to remember all those times he spent moving and getting in one measley sneak attack (if that)
You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells?
Not taking into account things that are irrelevant to the discussion (since terrible DPR is terrible DPR regardless of what else you specced for...it just means you're good at SOMETHING ELSE) =/= Looking at things in a vacuum.
The general assumption that people who aren't arguing and being contrary for the sake of being contrary is that DPR discussions are supposed to compare, shock of all shocks, people who are trying to DEAL DAMAGE (per round, even).
As much as I hate to bring World of Warcraft into it, this is something Blizzard got right.
Apparently, the majority of their game balancing is performed by looking at in-game statistics rather than mathematical models. They look at damage actually performed, therefore taking into account the things you don't get on spreadsheets, like average player time taken to respond and hit a button and people choosing that power because it has a nice shade of green or is more fun to use, rather than the one that does the most damage. They look at how popular a class is in PvP (if it's too popular, they nerf it because the reason for the popularity is likely that it's too attractive a choice.), they look at the composition of raid groups and times taken to down bosses.
They look at people showing them graphs and figures and spreadsheets and averages and calculations, and give them a metaphorical equivalent of "that's nice, dear."
At the end of the day, if what you're worried about is how it affects the players in the field, then actually examining how it affects players in the field is likely to give far better results than sitting at a computer doing DPS/DPR calculations. Human beings (other than the extreme minmaxers who you should never balance your game around if you want it to sell to the average player on the street) tend to have all different kinds of reasons for making their in-game choices. Some will pick things based on effectiveness, some will choose something because it reminds them of something they saw in a movie or read in a book, some just think an option sounds cool.
If you're worried about how the game will be when played by computers and robots, then by all means bring out the math.
The difference is Bizzard is actually able to track all that, regardless of peoples exaggeration or storytelling etc. Paizo does not have a database that keeps track of every Pathfinder game ever played, so they can monitor the statistical average of what class is being picked, what abilities are being used etc.
If Paizo could do that, then reporting table experience would be incredibly valuable. But the can't. No human perfectly remembers every aspect of the game they played, nor is WoW approach even 10% the complication that Pathfinder or other TTRPGs produce. Classes in WoW only have a handful of choices to choose from, where as Pathfinder you might have scores, or even hundreds, of choices to make at every level of the game.
@MrSin, maybe your experiences differ than mine, but in MMOs, character builds tend to be far more similar than in Pathfinder. Every class in Pathfinder can have dozens and dozens of options to choose from, so each one can be wildly different than others; not so much in an MMO. Usually, in an MMO, you've only got a handful of abilities you can use at one time, so your choices for use are very limited. This means builds are far more likely to be similar to one another than in Pathfinder. Look at games like LoL, Smite, DOTA, WoW, characters have very limited options, sometimes being restricted to only choosing their gear and nothing more.
Sure, some things are going to be universal (deadly aim/power attack for martials, glitterdust/haste/teleport for casters etc) but the way classes get built in this game can be, and are, incredibly different than one another.
@MrSin, maybe your experiences differ than mine, but in MMOs, character builds tend to be far more similar than in Pathfinder.
Oh yeah, that was a typo. Changed that just now. You definitely get more variance in a table top than an MMO. MMOs also have more players trying to push you towards the cookie cutter than a tabletop. Tabletops at worse try to artificially enforce certain ideals, but that's probably an entirely different conversation.
And I'm talking about putting too much weight on DPR. Obviously you don't want to do 1d6+2 at 10th level, but I see people measure the character's worth by that artificial DPR stat i.e. looking pretty much at DPR and not the larger picture.
That's because DPR is the easiest method of measuring combat effectiveness. You can do maneuvers, but they can get very tough to land at higher levels, and the problem is that maneuvers still leave enemies alive to deal damage. Dead or destroyed enemies deal no damage. Grappled, Triped, Over-run... the enemy can still deal damage.
As a martial character, your job is to deal damage, as a caster, your job is to make it easier for martials to deal damage. If your DPR isn't up to snuff as a martial, then you have a bad build, unless you have some other meaningful way of preventing enemies from dealing damage. Perhaps maybe you've got an absurdly high attack bonus, and relatively low damage, but you've got something like Dazing Assault that lets you stop enemies from attacking.
The only characters for whom we measure DPR, are, oddly enough, character's who's job it is to deal DPR. One does not play a Paladin in order to be a skill monkey, one plays a Paladin to be a holy badass slayer of evil. You're not a badass evil slayer if you can't hit the broad side of a barn or cut your way out of a wet paper bag.
You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells? What about if they rearranged their stats from a 40.2 DPR to 30.8 that seems bad right...unless maybe it was to make sure they had an adequate AC/saves/hp. etc. etc.22.5 DPR doesn't matter in any of those cases. Its of course, still terrible at level 10 for anyone who's actually trying to do damage. Which when DPR matters.
When you are trying to deal damage, your attack bonus might matter, and your damage might matter, but your DPR surely won't.
When you are trying to deal damage, your attack bonus might matter, and your damage might matter, but your DPR surely won't.You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells? What about if they rearranged their stats from a 40.2 DPR to 30.8 that seems bad right...unless maybe it was to make sure they had an adequate AC/saves/hp. etc. etc.22.5 DPR doesn't matter in any of those cases. Its of course, still terrible at level 10 for anyone who's actually trying to do damage. Which when DPR matters.
Don't suppose you could expand on that could you? A high attack and no damage could be pretty awful, and a high damage but low attack could be pretty awful too. DPR is just calculating that against a particular AC(Avarage for the CR usually) I thought. I'd think its the part that matters most!
You say they aren't and just took one in a vacuum. Is a 22.5 DPR terrible at 10th if the person has a really high bluff/diplomacy in a very social game? Or if they have good SoS abilities? Or if they have a lot of buff spells? Or a wide range of utility spells? What about if they rearranged their stats from a 40.2 DPR to 30.8 that seems bad right...unless maybe it was to make sure they had an adequate AC/saves/hp. etc. etc.
I think you're getting too hung up on the idea of DPR being the only relevant maths in PF. There are other ways to judge effectiveness based on math, and rather ironically a lot of the abilities you mentioned in your counterargument are even easier to calculate.
For example:
In a very social game, what would you rather play - someone with +2 diplomacy or someone with +22 diplomacy? Do you think these two numbers give us a *general* idea of who may be better at diplomacy? Of course they do!
Similarly, that caster: if he wants to make a monster fail a save, should he have a DC of 19 or a DC of 20?
And finally, that fighter: if he wants to bring that monster to 0 HP, what is better - 10 DPR or 20 DPR?
My rhetorical questions are loaded, of course. Theres a fundamental difference in the statements I am giving: they include some constraints on expected performance. Stats are only helpful when asked within these constraints. The point you should be arguing is that sometimes people forget to ask 'what are the relevant goals I am trying to achieve with this calculation?' and end up creating a DPR monster for a social game.
Of course, that someone could make a DPR monster for a combat game and be doing the right thing. Not only this but if they are answering a specific question, the maths lets them answer that question with a certainty trial-and-error cannot provide. You are currently trying to argue against these points and it isn't doing you any justice.
And I'm talking about putting too much weight on DPR. Obviously you don't want to do 1d6+2 at 10th level, but I see people measure the character's worth by that artificial DPR stat i.e. looking pretty much at DPR and not the larger picture.
You would be surprised the amount of people that have told me that that is perfectly acceptable DPR for a rogue who can't get a flank.
NOTE: It's not.
Sure it is, if monsters encounter at that level have 17 hit points
You still don't kill them, while the fighter is laughing at everyone who told him that great cleave and lunge were stupid feat choices.
Sure it is, if monsters encounter at that level have 17 hit pointsYou still don't kill them, while the fighter is laughing at everyone who told him that great cleave and lunge were stupid feat choices.
Oh to be Enlarged with Lunge, Power Attack and Great Cleave in a room full of 1 HD Goblins...
Oh to be Enlarged with Lunge, Power Attack and Great Cleave in a room full of 1 HD Goblins...Sure it is, if monsters encounter at that level have 17 hit pointsYou still don't kill them, while the fighter is laughing at everyone who told him that great cleave and lunge were stupid feat choices.
This happened to me once. It was glorious.
Then the DM read the challenge rating rules and realized you don't just add together individual CR to make a higher level encounter. IE 30 CR 1/3 goblins isn't actually a CR 10 fight.
Then I was sad.
|
|
| 1 person marked this as a favorite. |
Oh to be Enlarged with Lunge, Power Attack and Great Cleave in a room full of 1 HD Goblins...Sure it is, if monsters encounter at that level have 17 hit pointsYou still don't kill them, while the fighter is laughing at everyone who told him that great cleave and lunge were stupid feat choices.This happened to me once. It was glorious.
Then the DM read the challenge rating rules and realized you don't just add together individual CR to make a higher level encounter. IE 30 CR 1/3 goblins isn't actually a CR 10 fight.
Then I was sad.
Every player should have a moment that awesome, but without knowing what's going on behind the screen.
-Skeld
Every player should have a moment that awesome, but without knowing what's going on behind the screen.
-Skeld
Oh, but the GMs face can make that moment much more delicious.
Every player should have a moment that awesome, but without knowing what's going on behind the screen.
-Skeld
Don't worry, I've had better moments that were even RAI-legal. Still agree and appreciate with the sentiment :)