Organic Levelling
I enjoy the narrative structure of Elder Scrolls and RPGs in general, so, while I recognize that Skyrim is a game, I try to avoid disupting the narrative structure of the game for levelling. For instance, imagine in your favorite cop show, if the detective had stopped in the middle of the crime scene, shot himself in the foot, repaired the damage with his amazing medical skills, and then repeated it 5 more times. You'd think he was insane, and in any case it disrupts the narrative. Organic levelling is a system I've devised to try to capture what I see as a narrative-friendly way of "practicing" skills. I also happen to think it extends the enjoyment of Skyrim - there are an AWFUL lot of dungeons, caves, lands to explore, enemies to fight, etc. If you "grind" to maximum values in key skills too early, you're MUCH more likely to be bored with your combat tactics on the 45th dungeon than the person that is continuing to get new perks, spells, etc, to that point and beyond.
 Basic Principles
- No deliberate "grinding"
- If it wouldn't make sense for an in-universe character that has no conception of skill levels or game mechanics, it should be avoided.
This means no deliberate casting spells repeatedly near enemies to gain skill, no deliberate damaging of oneself, no healing of opponents, and no sitting for hours letting a mudcrab attack you in order to improve your heavy armor skill.
However, it is fine in a fight you are confident you are going to win (e.g.any fight with a wolf past level 5 or so), to use "less practiced" skills to kill the wolf. In the same way a tennis player playing in a less important match against a weak opponent might use the opportunity to work on their weak net skills, a reasonable adventurer might conclude that a pack of wolves is a good opportunity to practice with that bow even if he's mostly been using swords, or vice versa.
Deliberately seeking out fights is also reasonable, to a point. An adventurer on their way from Rorikstead to Whiterun might well hunt any deer encountered along the way, but probably wouldn't zig-zag all over the plains looking for every last one (who needs that much meat anyway?). An adventurer with a full load of loot she is carting back to the local town has far less justification for hunting deer if she can't even carry the spoils.
Casting beneficial spells on NPCs, or non-combat spells on yourself is reasonable if they could be of some aid (the NPC is a follower, or you expect combat). Casting Stoneskin in the middle of a bar, or Courage on every patron of the bar, is probably violating the basic premise unless you have some sort of reason to be doing so even if skill levels did not exist.
In general, the basic idea is to level what you actually use during the majority of your active adventuring time. Play as if skill level increases were a random occurrence not linked to any specific action, but generally known to require active use of the skill.
 Crafting Skills
These are an exception, since they must be deliberately practiced to get any significant skill progress at all. More later.
The intention of this section is to keep track of what potions should be picked up from alchemy vendors and which should be hand made, assuming you are an alchemist. The comparison numbers are rough, but assume 4 unboosted max alchemy enchant (i.e. total bonus of 100%). Numbers would be slightly different with bigger bonuses, but you can't reach high enough to really affect the comparison without exploiting the Fortify Restoration glitch. Some effects are not really worth obtaining in any way, but are listed here in case someone actually does want to use them.
- Aversion to Magic (Made = 40%, Bought = 100%) - It's worth buying the weaker variants of this, as they will STILL be more powerful than you can generally make.
- Aversion to Fire/Frost/Shock (Made = 90%, Bought = 100%) - You can probably just reach equality with store-bought with looped enchants, but it's still just good to buy these. For most of the game they will be stronger than you can reasonably make.
- Fortify Enchant (Made = 28%, Bought = 25%) - While you can make better at the top end, it requires really high enchant skills to get there (and a lot of expenditure on Grand Souls). For most of the game, store-bought will be comparable or better.
- Fortify Barter - going off of memory, these are pretty weak made. To verify when I have the game in front of me.
- Fortify Destruction (Made = 150%, Bought = 50%) - These are worth making even with lesser alchemy skills - unfortunately the ingredients are somewhat hard to come by.
- Fortify Conjuration (Made = 150%, Bought = 100%) - Somewhat weak effect, but still clearly better than store-bought.
- Fortify Illusion (Made = 120%, Bought = 100%) - These are reasonably comparable, though the extra 20% can help on Expert and Master Illusion spells.
- All other Fortify Skill potions (Made = 120%, Bought = 50%) - These aren't quite as good early, but by the mid-game you can make better potions if you've been working on Alchemy.
Generally I've noticed that poisons (barring special ingredients) tend to be noticeably weaker when they are made by the player, even with the Poisoner perk. This suggests that Poisoner is possibly a poor perk to take at this point, or that it might be tweaked to increase its strength (or introduce multiple ranks) when the Creation Kit comes out.