As I've discovered, the commander is REALLY front-loaded. I haven't looked at it that much at higher levels, although my gut says that it would be fine if it had only 1 or 2 cohorts max, if only because 3+ cohorts is just a plain old ton of actions that are still meaningful with 2/3 casting.
I'm going to be jumping around a bit in this analysis. Sorry for the disjointed thoughts. I've got a lot of them and I'm just trying to get them all out and recorded.
Okay, looking at the commander at 1st level, two front-loaded things jump out. That's a big problem, since those two things are all of the commander's class features. The end result is, in short, numbers that are too high at 1st level.
1) Auras go up to Cha bonus. This is your standard sort of frontloading that you see in a lot of classes, where you have a class feature you can dip into that doesn't deeply care about continuing with the class that granted it. The aura choices themselves have a few balance issues relative to each other, but that's a separate matter I'll address later if I remember. Some of the auras are stupidly good at low levels but just don't scale so well, and that's something I will be addressing here.
2) Cohorts. This is the other type of frontloaded class feature, and it's a bit subtler. You've got a class feature that relies on class levels for its scaling and really does need that scaling to keep up at higher levels. Its problem is that at the level it shows up, it's really really really powerful, and it doesn't drop down for a couple of levels until the scaling system catches up. Specifically, the fact that you get a full fledged 1st-level cohort at 1st-level, which the scaling formula implies to be the appropriate power level for what you should get at 3rd.
The cohort is a bit harder to evaluate than the auras. You see, auras and their effects are pretty straightforward. You're giving a bonus to your party just like it's written out. The bonus is going to start at +4 (there's no competition for stats against Charisma, so you will stick an 18 there), and it's going to increase by around +1 every 3-4 levels, give or take, with a little jump at the end when you start using tomes. Depending on your level of optimization, this could go a lot higher (ask Soro_Lost, he loves to talk about Charisma scores in the 60s or higher). The biggest problem here is that the numbers end up in a reasonable place, but just start too high. This isn't just because of the ability score-based scaling, which starts much higher than its prolonged scaling would imply, but because of what this gets applied to.
I'm going to reference the
marshal class, from the Miniatures Handbook. Say what you will about the class being too weak (it is), it did have a decent sense of what sorts of numbers can go on what level of scaling. The marshal's auras are divided into minor and major.
- The minor auras are lesser effects that don't directly impact combat power (ex: bonuses to skills) and more limited effects that generally appear with smaller bonuses, but can afford larger ones here due to being more situational (ex: bonus to AC vs. AoOs) or otherwise applying to only a subset of what they normally do (ex: bonus to a single type of save). It's not even uncommon for bonuses of these sorts to apply fully based on an ability score when found as class features (ex: a paladin's divine grace granting +Cha to all saves), but those only apply to a single character, rather than the entire party.
- The major auras are your more generally applicable, always useful combat bonuses. This includes things like bonuses to attack rolls, AC, or all saves. Major auras have a much more limited scaling, based solely on class level. Similarly, you often see these sorts of bonuses as class features in larger magnitudes, but as with minor auras, those apply to individual characters, whereas this is a class feature that applies to each of the at least 4 people in a party.
This division of bonus categories is an important thing to note, and was one thing that the original marshal did mostly right. There's a huge difference between +4 to all saves and +4 to a few skill checks. Presently, all of the commander's auras are lumped into a single pile. With the exception of a couple that scale at reduced rates (fast healing, caster level) a couple that don't scale at all (alliance, dogpile), and a few that provide bonuses that are inherently on different scales and are adjusted appropriately (reflective damage, SR, resistances, and movement speed), all of the auras give a straight bonus equal to the commander's Charisma modifier. Everything acts on the same scale. This means that +1 damage is valued as highly a +1 on attack rolls, +1 AC, +1 to all saves, +1 to a trio of skills, and +1 max AoO/round. Some of those comparisons are quite reasonable. Others, not so much. There needs to be an adjustment in the scaling between the various auras.
Back to the ability score-based scaling. Remember, these auras are applying party-wide. Taking the +all saves aura means that not only do you get divine grace, but everyone in the party gets it, too. Worse, they get to use your optimized Charisma score, rather than their own. Getting +Cha to all saves is fine. Giving a small to moderate bonus to all saves to the party is fine. Giving a big bonus to the whole party in a limited set of circumstances is fine. Combining the best of all of those is problematic. I'm just using saves as an example because it's easy to understand, easy to break down (all vs. Fort or Reflex or Will), and has a solid comparison point (divine grace @ paladin 2nd), but the same principle holds true to most of the other bonuses that auras can grant.
Now, one last thing about auras before I try to move on to cohorts. Some of the auras just produce way too big numbers at 1st level. This is strongly related to what I just talked about, but there are a few specific nuances to it I need to address. I've already discussed how +Cha can give bonuses across a party that are much larger than they should be, but most of what I've talked about have been about bonuses on a d20 roll. Some types of numbers, especially at the lowest of levels, have a much narrower band of reasonability. In this case, I'll have to dissect the specific auras that relate to this.
- Energy Shield: Reflective damage is cool. A little bit is neat and useful. Unfortunately, at 1st level, hit points are on a much narrower band than a d20s 1-20 range. Typically, hp goes from 6 up to about 15 at 1st level. You get a few outliers at CR 1 like the 29 hp sack of meat that is a 4 HD zombie, but on average you get 12-13 hp for CR 1 enemies (according to
Optimization by the Numbers, and level 1 encounters generally deal with larger quantities of weaker, lower CR enemies, too. 18 Charisma deals 8 damage when hit, enough to take down the average CR 1/2 creature (6.6 hp) and takes off half the hp of the strongest one (16 hp). Your typical CR 1 enemy loses after hitting you twice, too. Your typical CR 1 enemy is doing noticeably less damage than that on each of its hits, too. Most enemies will only be averaging 4-7 damage per hit at CR 1. Even the 18 Str elite array orc fighter with a greatsword, sitting at the top of the damage curve, only deals 13 damage per hit, just over 50% more than you do reactively. If he doesn't drop you in that single hit, he's still going down to the second burst of reactive damage, so at best he gets a pyrrhic victory as you both bleed out. One of the deadliest enemies for the level has a non-zero chance of losing if you stand there and do nothing. Most any enemy that needs 2+ hits to drop you loses automatically.
- Toughness: Damage reduction is a difficult thing to evaluate. It comes up so rarely on the players' side, but even then usually in very small quantities. It has a similar issue of quantity as energy shield, being very potent at level 1 but losing that potency as higher level enemies come with higher amounts of damage. Small amounts of DR have very little effect, but the defensive benefits become more and more pronounced the higher DR rises until all or nearly all of the damage is prevented. It's similar to AC in this regard, except that instead of working on a 20-point scale, at 1st level DR typically works on a 4-7-point scale, (although outliers can go as high as the low 10s). Having DR 4 will halve or negate the damage that most CR 1/2 and CR 1 enemies deal. Even our 13 avg. damage orc from before loses significant enough chunk of damage to guarantee that he'll need at least 2 hits to drop your party's barbarian or warblade, and give very good odds for even the squishier commander with only a d8 for hp.
The exact effect of a cohort at 1st level is much harder to analyze, as it has numerous factors that must be taken into account. The numbers, too, are less intuitive. At 1st level, the arcane, divine, and psionic foci pretty much amount to 1 spell/power per day. It's not a direct thing, but we do kind of know at this point roughly how effective one casting of a wizard or cleric spell or a psion power is (roughly 50% chance to win one of the 4 encounters in a day is what I typically estimate it as). Not unreasonable for a companion-type class feature if the rest of the class is comparably potent (which it would be once 1st-level auras get addressed). Stealth and wild foci are mostly about skills, and since wild's wild shape doesn't show up at the level range we're evaluating, it can be safely ignored. A commoner 1 with +2 damage or 2/day invisibility is very reasonable as a 1st-level companion, so those actually seem quite fine at first glance.
Combat focus is where the numbers really start to come in. It adds a pile of them. Hit points, attack, damage, DR, a feat, and a maneuver. There's a lot to look at here, and I don't think I can do an effective, accurate analysis for this sort of thing without actually statting some things out. I'm getting a bit tired, and generating those numbers is a lot of work, so I'll come back to this later.
Before I go, though, I don't want to forget about some of the front-loading scaling I mentioned earlier. At higher levels, the cohort keeps to 2 levels behind the commander. That's fine. At 2nd-level, the cohort is 1 level behind the commander... warning signs, but not necessarily wrong. At 1st level, however, the cohort is right at the commander's level. Alert! With combat focus, you've got a 1st-level fighter for a pet at level 1. If the commander stays home and reads a book in bed, he's already contributed his share to the party at level 1, just with the cohort tagging along. It's similar to the druid with its animal companion. The commander has the same class chassis, even, just with auras instead of spells.
I'm not sure what the fix is, especially since a major part of the premise of the commander is having cohorts. My gut ideas are either to remove the cohort at levels 1 and 2, or to remove the focus entirely until level 2 or 3, leaving the cohort just as a CR 1/3 commoner at 1st level. I'm not sure how well those would work, and I suspect I'll know more and/or have other, better ideas once I get the deeper stats analysis done.
Also, the cohort level/effectiveness just scales unevenly at low levels. That's just off-putting. It's 1, 1, 1, 2, 3, 4, .... Those three levels sitting at level 1 just don't work. As a creature unto itself, it's virtually impossible to make it appropriately powered at all three levels (commander levels 1, 2, and 3) despite the cohort staying identical across them.