So you're already a strong builder of decks thanks to the past two tips. I've only got one left to give you, one that will (hopefully) keep you improving forever. It's only listed #3 because I like the narrative structure of it being last - IMO, it's actually the most important of all.
Tip #3: Don't Play Every Game to Win
Don't click away! I promise this is still about optimizing your win rate and getting better at deckbuilders. This isn't me saying "remember to have fun, it's just a game. :)" Because while yes, that's true, I know that's not why we're here.
This tip should IMO apply to everything you do in life, but deckbuilding is a good place to start because their design rewards experimentation much more than other games.
Dedicate games to experimental strategies that you'd normally avoid. In order to grow, you must view cards/strategies from new angles, in case you discover something better than what you're doing now.
i understand the fear of trying something strange
There's a kind of tension for players who are concerned about playing well, and it's this: if you care a lot about victory and defeat, and you're grinding out lots of games to try and improve, you'll naturally start adopting patterns of play. Developing patterns is mostly good: save mental energy by formulating frameworks and strategies to follow, right? You had a great run with a specific card, so every time you see that card, you remember what synergies generally worked and try to replicate them.
But adopting patterns creates problems when you start to accept too many things as a given. For example, "thin decks can be good with certain cards" can easily become "always keep your deck thin". Or "This card combo is pretty strong" becomes "I see this combo is available, so I'll ignore other potential strategies".
My theory is that some games cause humans to generate a lot of inaccurate 'rules' whenever game variables have a massive influence on your outcome but are ALSO so numerous that it's hard to account for each variable at once.
For example: what if your OP card combo five games ago only worked because there wasn't another card in the available pool that your opponent would have otherwise grabbed in response to your strategy? That one game 'proved' that your strategy was strong, so now you're stuck with a big head and a big miscalculation of power level.
No good.
If you're like me, it's ridiculously hard (and kinda boring) to look at every single card combiniation available and ascertain all of the theoretical synergies and counters that are possible. And to do such a thing every turn would make your playgroup hate you. It's arduous to perfectly logically think your way through this stuff. (Which is by design - if you could, the game would be "solved" and become stale.)
experimentation doesn't always pay off, but it's always worth trying
So what do you do instead? The answer is, essentially, the scientific method: just play through a bunch of games and try out different strategies in different contexts, and make internal note of your observations. If you keep your attention open, you'll start to formulate (almost subconsciously?) the nuanced contexts that make a certain strategy better/worse in different scenarios. And that's how you get better.
Whenever you play to win, you'll find yourself kind of "locking in" a set of rules for yourself to give you what feels like a consistent game. But if you never have experimental games, if you never try out something new, you'll never open yourself up to evolving layers onto your strategies. On the obvious level, you might learn that you undervalued a certain card. But on a less obvious level, you might learn that you had overvalued your go-to combo and can produce an equivalent effect by doing some other thing.
Finally: yes, if you can give yourself "win conditions" such as having fun or seeing how far you can take some janky strategy, you'll probably just have more fun with the game overall and avoid burning out. So that's nice too.
Now get in there and do some science! :)
+: Seriously, Keep Experimenting
This one's going to be short because, by design, this third tip is something that you'll probably be doing 'forever' if you really want to get good.
In any half-decent game, you'll never really stop benefiting from experimentation. Yes, after a point you'll get diminishing returns. But if you always do the same things, you'll always get the same output. Sure, you won't get lower results by staying the same... but you give up the chance of higher results too.
At a higher level of play, your experimentation just gets more complex and nuanced. For example, maybe you'll start testing the quickest turn that the opponent's strategy can adapt to a change in your own strategy when doing x deck vs y deck with abcd cards in the pool.
You get the idea - there's always something to experiment with. Just introduce more contexual variables and you'll have a new experiment.
BUT. This only works if you keep an open eye to your results. Trying an experiment multiple times is important to see the same strat across tons of contexts. Taking internal note of what you see - more than once, ideally - is the key to improvement.
I guess just in general, that's how you learn in real life too.
Anyway, I hope you enjoyed reading this series, and hope this information can be useful to you. And I especially hope that I've made you excited to play some deckbuilders!
Good luck out there. :)