Benchmarks

[Korahir]

I am really curious what numbers are considered benchmarks regarding average dmg output, AC, saves, etc. by you playgrounders. I'll try to give 2 specific scenarios and you give me your opinion of what you expect when you sit down at the table. i.e. A beatstick with AC xx and Hitpoints xx, a dedicated dmg dealer to do xx average damage per round, a skill monkey having xx skills with xx ranks in it. etc.

Scenario 1:
Let's say you agreed to play with people you never played before. You were told the power level would be "normal" or "mildly optimized" and that the others are experienced in playing 3.5. You'll be starting at level 5. What numbers do you expect to see?

Scenario 2:
Again you agreed to play with people you never played before. You start at level 20 in an heavily optimized campaign. What are your benchmarks?



I also don't care if you describe a Player or GM perspective. I just want to get a general feel for your benchmarks.

[Fouredged Sword]

To expect to have your AC useful, you should aim for level +13 for characters who don't expect to avoid half of incoming attacks and AC of level +23 for characters who expect to rarely get hit.

[Palanan]

I notice this thread didn't attract much attention, which is a shame.

There are a lot of generalities and "everybody knows"-es bandied about the Playground, sometimes supported with intensive math and often not. I've always wondered if the really extensive number-crunching had been done years ago, on 339 or whatever, and all the Playground regulars are just so familiar with it that they don't bother citing it anymore.

One of my early builds on the Playground was criticized because it didn't do "big people damage," although this term was never strictly defined. (There were no halflings involved.) I'd love to see something like what Korahir is asking, but quantized for every player level, or maybe every two or three levels.

What should an acceptable fourth-level barbarian be dealing per round, as opposed to a Whirling Spirit Frenzy Lion Fudge Neapolitan build? What should a basic eighth-level duskblade expect to dish out, absent Iaijutsu Focus? And so on. If someone's ever gone through and done those numbers, I'd love to see them.

[Piggy Knowles]

My go-to is Optimization by the Numbers (http://community.wizards.com/go/thread/view/75882/19869122/Optimization_By_The_Numbers). This is a handy tool that comes up with average figures for AC, HP, Fort/Ref/Will, attack bonus, etc., for monsters at every CR.

This lets me figure out where I need to be as a character. If I'm claiming I'm defensive, what does that mean? What percentage of attacks are going to hit me on average at each CR? If I claim to be a damage dealer, what do I expect to be reasonable damage?

For example, if my goal is primarily to deal damage, the minimum I consider acceptable is that it takes me two rounds to take out an enemy under decent circumstances (ie, assuming I can get off a charge if I'm a charger build, assuming full attacks, etc.). I like to shoot for the OHKO, but two hits is my benchmark. So, I know that I will want to be averaging at least 205 points of damage per round by ECL 20, because the average CR 20 opponent has 410 HP. (Actually, to be safe, I prefer figures around 250, to account for DR and regeneration/fast healing.)

If I'm not handling most enemies in two hits, I have to either seriously reconsider my role (maybe damage dealing is secondary to some other build component, which is why I'm low on that end), or I have to adjust things accordingly.

Obviously different builds will have different benchmarks. A build focused on harrying attacks and debuffing isn't going to care about taking someone out in two hits, but it might care more about what someone's Fort save is, for instance.

[Korahir]

thanks a lot for the link, piggy. this helps me a lot.

Still i feel like there is a lot of personal feel to that. You mentioned you are going for the OHKO with the maximum two rounds on damage dealer. How do you feel about specialists that can OHKO a lot of things but completely fail against others (precision dmg springs to mind)? What is the "benchmark" here?

What probability do you shoot for considering the average saves? Is 50% if the ability has multiple uses? Is 20% fine if you can do it every round?

[eggynack]

There's probably no numerical value that I'd assign to various optimization levels. Numerical differences make far too small of an impact on overall power level for me to use benchmarks in that manner. My general benchmark, such as it is, is to determine the tiers of every other character, and settle in at around the level of the highest tier character in the party. That's generally just because higher tiers mean more options, and I like piles of options all the time.

[Fouredged Sword]

Optimization is art, not science. It by it's nature is about having a feel for the gaps in your character, and a good sense of your strengths.

The optimization by numbers thread is great for determining ballpark, but there is so much more to building a mechanically great character than numbers. That said, good numbers is a great place to start the conversation.

[Telonius]

There are a lot of generalities and "everybody knows"-es bandied about the Playground, sometimes supported with intensive math and often not. I've always wondered if the really extensive number-crunching had been done years ago, on 339 or whatever, and all the Playground regulars are just so familiar with it that they don't bother citing it anymore....
If someone's ever gone through and done those numbers, I'd love to see them.

Part of the problem is likely the demise of the d20 optimization boards, as well as the accidental purge (back in 2009) here. Lots of old info (most of it not really useful, but some real gems) have vanished into the ether since the forums got started. Some of the threads were saved and resurrected, but not all of them. It's actually pretty rare that anybody saves these things to their own computers, and reconstructing something like that can be a real pain in the butt.

[Piggy Knowles]

thanks a lot for the link, piggy. this helps me a lot.

Still i feel like there is a lot of personal feel to that. You mentioned you are going for the OHKO with the maximum two rounds on damage dealer. How do you feel about specialists that can OHKO a lot of things but completely fail against others (precision dmg springs to mind)? What is the "benchmark" here?

What probability do you shoot for considering the average saves? Is 50% if the ability has multiple uses? Is 20% fine if you can do it every round?

As others have mentioned, it's not an exact science. I just use the Optimization by the Numbers to make sure my build can actually do what I think it can do. If I'm claiming to be a damage machine, I ask myself, how much damage do I actually do? What does that represent in terms of actual enemies I'm taking out? Sometimes I see builds that claim to be damage machines, but they're only putting out maybe a hundred points of damage at ECL 20 - that means it's going to take five full attacks under ideal circumstances to take out an average CR 20 opponent!

I also have some other personal philosophies as well. I tend to focus more on offensive numbers than defensive ones, instead preferring to focus on immunities and non-AC defenses. I have also done some of my own tallying of things like common resistances, or common senses beyond standard vision/darkvision. For example, here's a spreadsheet I did showing SRD enemies with alternate senses (https://docs.google.com/spreadsheet/ccc?key=0Am8aHMxh9PPJdHRxWlpqVkpnZzZxSUxDYk1YMVc0X 2c&usp=sharing) - handy for knowing what to expect if I'm, say, playing a stealth character.

In general, though, it's really not this hard science where you can say, "a character focused on save-or-dies needs to have saving throws of exactly DC 29 or greater, a character focused on defense needs to have a minimum AC of 42" or whatever. Instead, it's a question of deciding where you want each build to be, and then double checking the numbers to make sure you're actually getting there.