What is Practical Optimisation?

[GilesTheCleric]

What does the term mean to you? At what point does something stop being PO? And, what does it take to reach PO in terms of system mastery?

[Red Fel]

What does the term mean to you? At what point does something stop being PO? And, what does it take to reach PO in terms of system mastery?

A thing has crossed from Practical Optimization (PO) to Theoretical Optimization (TO) when any of the following occurs:
Its use at the table results in concussion by literature
It requires at least one degree in a mathematical field to utilize
An explanation of it takes more than five minutes
The application of it to a single action takes more than five minutes, or results in a NI numerical result
That's too many books!
The resultant build could bench-press a celestial body in a setting where one does not do such things
Any combination of the above.
It's like the "rule" of obscenity - "I know it when I see it." Look, if my build is just designed to do a single, non-game-breaking thing really well, it's likely PO. If my build cracks continents with damage, TO. If I have the equivalent of Batman's utility belt, somewhere in between. It's a scale, not a perfect bright-line rule.

[JNAProductions]

It really varies, person to person. If you ask Emperor Tippy what their PO is, the answer is going to ridiculously high-level TO for most tables.

[legomaster00156]

I must concur. PO and TO are extremely subjective to each person and each table. Chain-gating efreetis for endless free Wishes may be banned at one table, but at another, it may be something every wizard does sooner or later, leading to a race of efreetis with burning hatred for all mortalkind, that now seek to break their own doom of perpetual wish enslavement.

[Alent]

Heya Giles! Long time no see! :smallbiggrin:


It's like the "rule" of obscenity - "I know it when I see it." Look, if my build is just designed to do a single, non-game-breaking thing really well, it's likely PO. If my build cracks continents with damage, TO. If I have the equivalent of Batman's utility belt, somewhere in between. It's a scale, not a perfect bright-line rule.

I find this a good description. "Optimization I can use at my table without violating the gentleman's agreement of play" is also probably another good answer.

In terms of system mastery, it strikes me as a level of skill that lies past theoretical optimization, but doesn't require you to actually master TO, just understand some relevant TO. It's that old difference between "knowing if you can" and "knowing if you should." I'd picture it as an erratic sawtooth progression, where you learn some TO, learn how to practically apply or when not to apply it, learn some more TO, etc.

[AvatarVecna]

Practical Optimization, to me, tends to not involve questionable rules interpretations, infinite/technically finite loops, and other highly looked-down-upon tactics and abilities. Because that answer is kinda vague on what exactly counts, it's better for me to approach it from the other direction: namely, if something isn't Practical Optimization, it's Theoretical Optimization, so what is TO? To me, Theoretical Optimization is defined by the former word: Theoretical. This is something that works in theory and not in practice. If I could see it flying at my table, or that of my DMs, it's Practical.

Wizard 5/Shadowcraft Mage 5/Incantatrix 20 with every feat after ECL 21 being Epic Metamagic or a metamagic feat, so they can apply any metamagic they know to any applicable spell for free? Practical. That same build DCSing their Heroics feat to gain a permanent feat of any kind, and doing so to gain effectively all the metamagic feats they want for free? Theoretical.

Barbarian 25/Half-Orc Paragon 3/Hulking Hurler 2/War Hulk 10 with Natural Heavyweight/Reckless Rage, with every epic feat being Great Strength, with an item of Continuous "Giant Size", with a Tome Of Str +5, and with a Belt of Epic Giant Strength +126 crafted specifically for them by their Epic Artificer friend, hurling a 10 mile diameter sphere of solid iron at somebody and dealing 187 trillion d6s of damage to them is PO. A Hulking Hurler/Cancer Mage dealing NI damage, or a d2 Crusader, or a crit-fisher using Aptitude Weapons with Lightning Maces and Roundabout Kick for infinite attacks, is TO.

Only examples, BTW.

[Cosi]

To be cheeky, PO is "anything you've been allowed to do in a practical game" and TO "is anything you've only theorized about".

To be slightly more serious, it doesn't really matter. You could probably figure out a line somewhere (probably something to do with infinite loops or being able to defeat any printed enemy), but ultimately the thing matters is whether a build is past whatever power level the game is supposed to be at regardless of whether that excess power is "practical" or "theoretical".

In general, I don't think you can have meaningful normative discussions about optimization absent an optimization target.

[Flickerdart]

If you can get away with it, it's PO. If you have to say "But by RAW...!" it's TO.

[Gnaeus]

I wrote a guide. In my sig.

[Echch]

I find this a good description. "Optimization I can use at my table without violating the gentleman's agreement of play" is also probably another good answer.

I think this is the view I have too. In addition, I would also say that a build is only TO when the actual combination is utilized. A Sorcerer that has learned Heroics, Shun the Dark Chaos and Embrace the Dark Chaos isn't TO. A Sorcerer that uses Heroics, Shun the Dark Chaos and Embrace the Dark Chaos in combination with each other for metamagic feats is TO. A Sorcerer that uses the three spells to get the Open Minded feat for a few minutes because Wieldskill isn't on his spell-list is somewhere in between.

Or how Gnaeus put it in his guide:

Showing your DM that you can play responsibly with your toys gives him little incentive to take them away. Learning that line between a good character and too good a character is at the heart of Practical Optimization. There is nothing wrong with keeping a couple of overpowered abilities in your back pocket and only using them to prevent a TPK.

EDIT: Yeah, looking a Gnaeus sig is a good idea.

[Segev]

Yeah, as others have said, the line varies from table to table. The crux of it is: if your DM understands what you're trying to do (i.e. you didn't pull a fast one, and he didn't fail to grasp the implications when you explained it), and he allows it, it's practical optimization.

If you're relying on "but it's the RAW" to try to insist, or you had to pull a fast one to slip it by your DM, or you just plain know that your DM won't allow it once he knows what's going on, it's theoretical optimization.

Well, actually, that second one is "munchkin optimization." Theoretical optimization is "I have no intention of using this at a table; I'm just seeing what the RAW will let me technically get away with, hypothetically."

Note that TO can become PO if you find a DM who thinks it's cool and will allow it. What is horrifically laughable TO at one table might be perfectly acceptable PO at Emperor_Tippy's table, for example.

[Bucky]

Is Tome of Battle TO because my DM won't let me use it?

[Segev]

Is Tome of Battle TO because my DM won't let me use it?
I suppose, yes. I mean, any builds you make with it are purely theoretical, since you can't play them, right?

[Gnaeus]

Is Tome of Battle TO because my DM won't let me use it?

Likely. I mean you could keep a couple of concepts in mind in case you ever came across a game where you could play, or you could find an online game that uses it. Or you could try politely and privately to point out the merits of the system, (ideally right after the wizard casts a spell that totally nerfs a boss fight).

It could even be way too strong for a group. I mean it usually isn't, but if your group is all fighters and monks and your caster players are incompetent, it is reasonable that ToB could blow combats out of the water just by being hard to screw up.

But yes, PO, like "balance", can really only be evaluated in the context of a specific group or game. It's one of the reasons why arguing tiers is so frustrating, because no two tables will draw the exact same lines. I've been in a group for 10 years and I still have to check sometimes. There's no way I could know what was PO at Segev's table or Flickerdart's.

[Hecuba]

Practical Optimization: Having some goal for your character which you can use your character sheet and your system mastery to pursue.

Theoretical Optimization: A set of tactics which are consistent with the rules of the game and will likely solve an in-game problem (or potentially any in-game problem), but which can be reasonably expected to subvert the functioning of the game if implemented.

[Grim Reader]

Its PO if it'll fly at the table I'm building it for.

Its TO if it won't, or if I'm not building it for a table but for fun.

[Deophaun]

In theory, there is no difference between TO and PO.
In practice, there is.

- DM Berra

[ExLibrisMortis]

Practical Optimization POt is a subset of Optimization O that applies at a given table t. Theoretical Optimization TOt is the complement of POt in O.

Elements of POt are characterized by their allowedness, that is, the table's Dungeon Master DMt has (implicitly) approved them. Hence, we get POt = {o in O : o is allowed}.

Elements are allowed based on various things, all of which can be formulated as house rules H, which are subsets of O that are banned, that is, H = {o in O : o is banned}. In that view, POt is O \ (H1 ∩ H2 ∩ ... ∩ Hn).

A big open problem in Optimization research is the question whether every POt can be finitely specified by a set of houserules, that is, is n always finite? A certain subset PDM* of the set of Dungeon Masters DM seems to suggest that the answer may not be as obvious as it intuitively seems. Stay tuned for more news!


*PDM stands for Psycho DM. Elements of PDM represent some of the most difficult to characterize elements in DM, and their study is ongoing.

[Esprit15]

Sometimes it is a simple matter of implementation. An optimized necromancer who sends their undead army to deal with anything and everything is theoretical optimization. The same necromancer who has their army maintaining their keep, with perhaps two of their strongest undead minions traveling with them is a lot closer to practical optimization.

[daremetoidareyo]

Statistics!
If any given sample of a hundred DMs in the population, if 85% or more would allow your trick in their game, it's PO. That's two standard deviations away from a perfectly average DM. Any trick more banned/nerfed than that is on the TO side of the spectrum.

[Doctor Awkward]

What does the term mean to you? At what point does something stop being PO? And, what does it take to reach PO in terms of system mastery?

In short: Theoretical Optimization is what occurs in the vacuum of a homogeneous campaign setting overseen by an arbitrary and wholly indifferent DM who believes that all fluff is mutable and does not care about anything other than the Rules As Written being followed to THE LETTER.

Practical Optimization is everything else.
It's things that are likely to fly at an actual gaming table assuming the DM has some sort of setting or fluff restrictions in mind, would prefer not to have to plan sessions around certain classes, or would like to see a (largely) balanced game where everyone has fun and a chance to shine.

Where one ends and the other begins is generally dependent on the type of game a particular DM wants to run (for the most part excluding obvious things like The Omniscifier, various infinite damage loops, and drowning yourself to recover from infinitely negative HP).

So does that mean that TO is pointless? Not necessarily.
The purpose of TO (especially things like Pun-Pun) is as a thought exercise. It's basically like destructive play-testing. Learning the system well enough to smash it to utter pieces gives a DM perspective. It allows you to look at a thing and know how broken it really is because you have a basis for comparison.

In addition, if you do end up at a table with players who think it would be a lot of fun to build stronger characters, having greater knowledge of the system will help you build stronger opponents to compensate for this (this is basically the entire point of the Red Hand of Doom handbook).

[Deophaun]

Statistics!
If any given sample of a hundred DMs in the population, if 85% or more would allow your trick in their game, it's PO. That's two standard deviations away from a perfectly average DM. Any trick more banned/nerfed than that is on the TO side of the spectrum.
Quibble: That little clause means that an outlier sample that just happens to contain all of the most restrictive DMs determines the difference between PO and TO for everyone.

It's the distribution of samples that would best determine PO/TO from your method, not the composition of any sample.

[D.M.Hentchel]

Okay I have a slightly different standard than everyone else it seems.

TO: Anything that seeks the most powerful/diverse set of abilities for a given scope (basically any optimization)

PO: As above with the added clause of, "does not trivilize or over-shadow other players in the game"

Basically making other players useless is where I draw the line, though an exception probably comes up where one player refuses help and has a very weak character.

[barakaka]

I've done the TO build with Anima Mage and Unfettered Heroism + Action Point / Metamagic Persist abuse for a ton of immunities... at level 14. In the campaign I was in, it was completely normal and to a degree required. We regularly fought monsters around CR 18-24. So in a sense, for that campaign, it was PO.

I'm of the opinion that PO means:

Whatever optimization (or lack thereof) one must do to reach the same level of power as the rest of the party, while not overtaking another's role in the party.
NEVER outperform someone in the thing they like to do.
ONLY do things that the DM can handle without going nuts.


Some classes (Truenamer) need to use some "broken" feats and items from "broken" books in order to match the power of a Wilder, or Warblade. Would YOU deny a Truenamer the resources he needs to get to tier 3 in this case?

Conversely, a Wizard should pick very dumb/supportive but fun spells each day if the rest of the party is a bunch of Monks and Truenamers and the DM is new. Leave battlefield control and action economy out of your build so you don't ruin anyone's day.

[Psyren]

It's like the "rule" of obscenity - "I know it when I see it." Look, if my build is just designed to do a single, non-game-breaking thing really well, it's likely PO. If my build cracks continents with damage, TO. If I have the equivalent of Batman's utility belt, somewhere in between. It's a scale, not a perfect bright-line rule.


It really varies, person to person. If you ask Emperor Tippy what their PO is, the answer is going to ridiculously high-level TO for most tables.

I'll take these, and a side of fries to go please

[Komatik]

One way I've seen it put at the WotC boards - TO is optimization that is never intended to see the table (sans Tippy's, I guess). It's just an exercise on how much you can squeeze out of the system hopefully breaking it into pieces in the process. PO is optimization intended to make a playable, competent character.