So, many people agree that the CR system is merely a very loose guideline, given how much of a difference proper optimization can make in terms of character capabilities.
To this I say, It could be made much more useful (or even work as intended) if the way of determining a party's effective level (EL) were better.
Note that any system we come up with will require more bookkeeping.
Also note that if you don't like bookkeeping, why the hell are you playing this game?
For the EL system, I suggest a table very similar to the table for Monster Statistics By CR.
Some things can be omitted. For example, since running out of hit points means that something has gone horribly wrong, and many PCs have defenses other than hit points, hit points shouldn't factor into party level.
Neither should spells per day or anything else that's more of a resource to be tracked over multiple encounters. If tons of spells per day increased their effective level for a nova, then their effective level would be much lower on the later encounters where they get hammered for their poor resource management.
And neither should consumables with a gold-piece cost, even though those technically count as "a party's resources." This is because those resources are not replenished on a daily basis, and come out of the treasure accumulated over the timeline of the campaign.
This EL system will take 4 PCs as the baseline. Note that gunslingers, magi, maneuver specialists, and many other unusual classes and builds will muck it up, which is why it is a rough draft.
If no character seems to fit the highest slot (the one with no multiplier) on the following list, use the highest value in the party. Never multiply the highest value of a statistic in a party. If a multiplication would raise a character above the highest value in the party, cap the value as equal to the highest value in the party. Use normal math when doing multiplication here.
----------------------
Armor Class: This will add up:
the AC of the party member most likely to soak up hits (fighters, battle cleric, eidolon, paladin, etc)
1.2x the AC of party members that attack in melee (rogues, TWF rangers, backup clerics, Arcane Strike bards, barbarians).
1.5x the AC of party members that primarily attack at range (primary spellcasters, etc)
and average them. The reason for the multiplication is because, since the others are getting attacked less often, they usually have a lower AC, and would otherwise reduce the average result by an undue amount.
-------------------------
High Attack: This will add up:
The attack bonus of weapon-using characters (melee or ranged) whose attack bonus is the result of permanent or near-always available feats and buffs such as Weapon Focus, Weapon Training, and (at higher levels), Rage and Judgments. If no character is optimizing, the highest consistent attack bonus takes this slot.
1.2x the attack bonus of characters such as paladins, rangers, and backup clerics that are no slouch in combat even unbuffed and can boost their attack significantly in certain circumstances, as well as fighters that are not using all their feats to advance their weapon attack bonus.
1.5x the attack bonus of such characters as bards and low-Str clerics without battle spells, that can employ weapon attacks if they must, but are typically doing something else and can only last a few rounds if forced to absorb all attacks.
2x the attack bonus of characters that rely on touch attacks, such as full arcane spellcasters (use the higher of either melee or ranged attack bonus).
And average them.
----------------
Low attack: This will add up
The attacks of full-BAB optimized characters forced to use weapons they weren't built for (greatbows for barbarians, greatswords for archery rangers).
1.2x the weapon attacks of characters that typically have certain buffs (rage, judgments, melee-focused divine spells, flanking bonuses) that can't use them.
1.5x the weapon attack of characters with 3/4 BAB using weapons they aren't built for at all (such as armored melee clerics using crossbows, or crit-focused finesse combatants using greatclubs)
2x the weapon attacks of characters ludicrously out of place wielding that kind of weapon (basically, any single-classed full arcane spellcaster using a weapon attack)
and average them.
----------------
Average damage, High: This will add up
The average damage of a full attack from characters intended to make full attacks, such as two-handed fighters or archery rangers. Calculate this damage using the formula here, using a creature of a CR equal to the character's HD.
The average damage to one target by a spellcaster's (or other character focused on an ability balanced around DCs) most gimmicked-out ability, such as the chain lightning spell of a caster with Magical Lineage (chain lightining), Spell Perfection (chain lightning), Greater Spell Focus (evocation), Greater Elemental Focus (electricity), School Power (evocation), and Empower Spell. Use the Poor save of a creature of a CR equal to the caster's caster level. If the gimmicked spell targets 1 creature/3 or 4 levels, multiply the damage by 1.1, or 1.2 for 1 target per 2 levels, or 1.3 for 1 target/level. If the caster's most damaging spell is not optimized, multiply the damage by x1.5 (after adjustment for number of targets). If their damaging spells are negligible (such as fireball with no metamagic at level 20), don't count the character when averaging out the damage.
If the caster uses a spell that doesn't deal damage but takes the target out of the fight, such as Flesh to Stone, use the Hit Points column of the Monster Stats by CR to determine damage dealt on a failed save, and use 0 as damage dealt on a successful save, and calculate damage based on that.
If a caster's signature ability requires specific circumstances that are violated somewhat frequently (i.e. dominate person versus nonhumanoids, electricity versus demons, or any mind-affecting power against undead or constructs), cut the damage by 10, 25, or 50%, depending on the specificity and how often it comes up in the campaign.
and then average them
-------------------
Average damage, Low: This will add up
The average damage of a single attack from characters intended to make full attacks, calculated from the formula above.
The average damage of a DC based damage dealer when they cannot use their gimmicked ability to maximum effect. Typically, this is (3.5 times caster level times the chance of the target failing their save) plus (3.5 times 1/2 caster level times the chance of the target making their save). Use the Good save of a monster with a CR equal to the caster's CL to determine chance of success.
If a caster's ability takes a monster out of a fight completely, or requires specific circumstances, follow the same guidelines as above.
And average them
-----------------------------
Primary Ability DC: This will add up
The DC of a caster or other DC-dependent character's most powerful go-to ability. For the Dazing chain lightning abuser, this would be their chain lightning DC. For a caster not specialized in a single spell but instead a school, use the DC of their highest-level combat spell of that school that allows a save. If they are not specialized at all, use the DC of their highest-level combat spell that allows a save.
1.2x the DC of an ability that is not particularly high but comes up very frequently, such as Dazing Critical on a TWF fighter using 15-20 weapons, or a Monk's Stunning Fist ability.
And average them.
-----------------
Secondary Ability DC: This will add up
DCs keyed off an ability score that is not that character's highest score
The DC of a caster or other DC-based character's ability that is not optimized at all, such as a combat spell from a school they have no specialty in and is three or more levels below their maximum
The DC of a critical feat on a character with a threat range of 19-20 and not using two weapons
And average them.
-----------------
Good save: The average save bonus of the party in the saving throw category in which their average saving throw bonus is highest.
-----------------------------------
Poor save: The average of the party's lowest save bonuses, regardless of save type.
-----------------------------------
This will then be cross-referenced against a table just like the Monster Statistics by CR table that will tell you their effective level.
