I do.

sicilian and king's indian are 2 of the most theory intensive openings

Yeah, you can just wing it in the French and the Ruy.

easiest? queen's gambit

hardest? catalan opening, queenside expansion variation

Traxler's is definitely the hardest, as there are like 1000 different things you have to memorize or you lose because you are down a bishop and a lot of times you are forked, so you have to keep doing checks

True

The Catalan is strategically more difficult than most.

That's one reason it's popular at the top level.

ruy lopez or

King's Indian Defense: Fianchetto, Debrecen, Classical Main Line, 8.e4 a6.

i copy pasted because it is to much to right

The OPs question is key for thinking about the opening.

First it depends on what you mean by average. You need to have some idea of your goals in chess... a target elo, and target your repertoire to that elo. For me.... I usually target 2100-2200ish. That's sort of the extent of my chess aspirations.

Then you can use database statistics to tell what works at different elo levels, and think about why that is / design your repertoire accordingly.

Here are some criteria I usually look for - 
- applies early pressure 
- requires a precise response / not entirely obvious how to respond
- somewhat uncommon
- still solid if they do handle it correctly

At lower level applying pressure early generally pays great dividends. This is more true when the opening is uncommon. At master level alot of lesser known lines will be played correctly by both players, but below master.. players will bungle the opening reliably. When the position is unusual they'll often bungle the middlegames too.

So I like to play some lesser common sicilians - I've played the nimzowitsch, taimanov, classical. And I also play the Van Geet with the Veresov. Against d4... I think the QGA is good, though the slav is also very good (not the semi-slav). I also think the tarrasch is very good. All these apply pressure early.

Top level players have often noted the tradeoff between solidity and novelty, but the nice thing about targeting your repertoire for lower levels is that tradeoff is often less severe or in some cases nonexistent.

For example, the Najdorf is great opening, but below masters it has one of the worst winrates of any sicilian... there's too much theory, it's too common, and black doesn't have ways of decisively punishing mistakes during the opening phase the same way as in other sicilians, because the move a6 is by itself a passive waiting move.

I avoid very common lines - i.e. Ruy Lopez, italian, scotch, Najdorf, QGD (for the most part), semi-slav, KID... Especially if they don't apply pressure early, which many of them do not because they're more theoretically ideal lines... I think you're just missing an opportunity by doing that.

Grunfeld features some very wild play.

Russian Variation etc.

Only engines get it correctly, because tactics alternate on every move.

Grundfeld for me also...

Defends against kingside attacks, has a strong classical presence in the center and has a very active piece development.

sargossa becuz no 1 playz it

As a fellow average player, probably the hardest openings to play from my experience are pretty much, any "modern" thing, such as KID, Pirc, Grünfeld, Benoni, Benko, some sicilians and such; i find them to be both, hard to play from either side, as they tend to be not intuitive at all, or most of the time, it feels like they are justified from rather brute calculation than anything else, and it's requiere a very precise play, as they tend to punish badly any sort of small mistake in the opening-Middlegame, and maybe even endgame; same applies with other openings from the white side, such as The Reti or Catalan, both of which took me a while to get used to.

The Traxler is crazy too.