I agree with all these points, especially delaying the release one week.

v0.91 bugs found.  Med size random map, self designed sovereign: First: scroll that initiates new quests not working. Second: When a beast master stack tames a beast, beast is left on former lair/goody hut.  If player moves into this square, reward is not offered.  If player waits a turn, moves tamed critter to another square, and then stack moves into empty lair, reward is offered.  Same as preceding except if there are two or more hostile stacks in the square with the lair.  Tamed critter will be in same square as hostiles.  Hostiles (in lair) seem to ignore newly tamed critter (for one season, anyway - never left tamed critter there).

Indeed, it should be a bonus for someone specifically specializing in exp gain, not a one hit wonder you always take. Wouldn't mind seeing potential completely gone though. 

"General" tree:  When I first started playing, I kept thinking the 'general' tree was about military generals.  For new people playing the first time, change this label to "generalist" would more clearly indicate what it is about. 

That would make it a clear non-brainer first choice......

Why not restrict path choices from Henchmen making them only have access to the general tree.  This would prevent over abuse of Henchmen by lowering their base effectivity.

I think it would be better to place the global enhancements (merchant, administrator, ...) in the general tree and only the sovereign should be able to select them. In this case they could be stronger, but they would not be overpowered, because they can not stack.

Based on the experience numbers for champions, +25% experience taken at level 3 only pays for itself at level 12 (i.e., at the point when you'd normally have gained enough experience to be level 12, you'd actually be at level 13). This is not a significant advantage, nor is it so much better than the other choices you could have that you absolutely must take it before anything else. Granted, you will reach some levels slightly sooner than you would otherwise, but you won't be a full level ahead until you'd normally have reached level 12.

You also don't get a second "free" level until some time after level 20 (I don't know exactly when since I don't have experience numbers for levels after level 22).

Being level 13 when you'd normally be level 12 is such a minor advantage that I'd seriously doubt the utility of taking Potential - an extra 1 to 4 health (depending on race and class) and an extra 1 or 2 accuracy, dodge, spell resistance, and spell mastery isn't going to make much of a difference to you at that point.

This is, however, still better than what Potential I currently is, which doesn't pay off until level 20, or the subsequent Potential traits, which don't pay off until your champion would normally be level 18 (Potential I and II) or 17 (Potential I, II, and III). Mix them with Knowledge and that stays at about level 17.

Edit: Assuming my curvefit is correct, Potential 1 as it currently is will not give you a second extra level (read: the first extra trait you earn, once you account for taking Potential instead of something else) until you'd normally have reached level 41. Thus, in a normal game the only benefit you'd see from taking the Potential trait is that you would have 1 extra level and 1 trait that grants no direct benefit to anything. This isn't a good pick currently for anyone. Taking all three potential traits and/or knowledge is useful if you're getting large bonuses from leveling (mage with Flame Dart, or the lightning bow, or something like that), since you at least get a few extra levels before you'd stop, but only 1 extra level by the time the map runs out of experience (or by when you run into the level ceiling, assuming you've tried making a superhero) isn't worth it.

Even at +25% experience, you only get your second extra level when your champion would normally be level 25.

Wow. I had no idea the exp curves for levels were so steep. Thanks for pointing it out so clearly, joeball.

It's not very intuitive that +25% xp actually means something like +5% in terms of levels.

Every Friday and Monday on my twitch channel

I've been generating and revealing maps(challenging, dense everything, medium), running a few turns to watch the AI(normal); after about a dozen or so maps I have a few observations.

It seem the AI has much better initial placement, typically clear for 3-4 spaces, with mostly medium stacks nearby. Did see a deadly stack within 4 spaces in several of the maps, maybe a quarter, but no closer than that.

The AI sometimes will just sit a turn or two, before moving or settling. They do seem to pick better spots, watched with joy as several times they settled the spot I would have chosen. One even spent 3 turns moving to a 3/3/3. Some AIs cleared the immediate area of goodie huts before settling, others did not.

There don't seem to be enough quest locations generated to justify Heroic's 2 point cost. Maybe it should just be a flat faction wide bonus to Hero xp, or faction XP, but IMO it's too expensive as is.

Again, JMO, but maps have too much water. For a game where water is only a barrier, except in the scenario, there should be more land. Ideally, an option for an all land map, along with a setting for river and lake density, would rule.

I played a lot of  sandbox .91, several sessions of 4 or more hours, couple of times a lot more, zero crashes. Been regenerating maps and watching the AI for the first few turns since .92 released, probably 3 dozen maps or so by now, zero crashes.

My PC is nothing special, Win 7, Geforce 9200, 4 gigs of RAM, everything up to date driver wise. Game has been very stable for me.