I noticed you and Andrew seem to be talking past each other here. Andrew is talking about Wesnoth in player vs player mode. You seem to be talking about the scenarios. From a balance perspective these are two completely different games.
For anyone reading here who hasn't tried both modes, the PvP mode is a fairly good and balanced standard wargame. You can play against the AI, which is decent but not great, or against other players online.
The scenarios on the other hand... what everyone is complaining about is true. But this is a problem with the scenarios, not with the game. In order to make them "interesting", pretty much every scenario has a bunch of gotchas in it, like if the objective is to get your general to the other side of the board in 25 turns, on turn 10 some event will happen like a dam breaking and blocking off the main path or a new enemy appearing on your flank. On a first time play through, it is very hard to anticipate what gotcha will appear and how to mitigate it. To make it worse, most scenarios are ten or more battles long, and you need enough survivors and gold from the early battles to win the later ones. So, if you barely win an early battle you've basically lost the scenario, but you won't find that out until later.
To sum up, when people are loving on Wesnoth, they are talking about the game, and mostly play it in PvP. When people are hating on Wesnoth, they are talking about the scenarios, which are truly unbalanced. Tell people to try the PvP first.
Of course I was talking about p2p! You can always savescum in campaign, so why complain at all? :-D
Damn. Now I need to install Wesnoth and play some games... .
Needless to say you could win games of NetHack by doing the metagaming strategy. The problem is that it wasn’t going to help you win the current game in front of you, the one you’re playing right now. And that’s where some players (like myself) part ways with metagaming. We are interested in winning the current game. That is, taking every game as a unique puzzle to solve and attempting to win. So a game that randomly decides “haha you lose” is unsatisfying.
(I mostly used save-scumming in the campaign, due to overwhelming odds anyway, so I'm speaking about luck in multiplayer games)
The original point of comparison here is Heroes of Might and Magic which does not have this problem since the game is designed as a campaign with many battles and the leveling up aspect only applies to heroes which are essentially generals that do not directly participate in battle.
That said, that Double Suns campaign with desert elves is quite great and I enjoyed it, even if it uses rather different settings from the base game.
Other games manage to incorporate a lot of randomness while keeping the basic interactions (combat etc) deterministic and predictable. For an example of this, see Slay the Spire. In that game the cards all have completely predictable effects. The randomness comes from the cards you’re dealt. Furthermore, you can control the randomness of card draw in numerous ways, potentially making your deck very reliable, possibly at the expense of full potential.