Mathematically, the odds are better than 98.24% on every game.
However, experience differs due to variance. Some folks have huge luck and others have the opposite; same as any other game. There are also multiple approaches to probability, and one in particular helps demonstrate why Lucky Bit in particular gives some people such a negative impression.
Consider that the odds of hitting the very edge win are 1 in 32768. Well, if you don't plan on playing 16,000 bets you can basically assume that for your instance, the odds are actually zero. They aren't, of course, and that's the fun, but when evaluating an expected return, it's a safe estimation. The same logic can actually be applied on down the outer bets: the second payout has 1 in 2048 odds so if you don't have 1,000 bets in the forecast you can safely assume you won't hit that slot. Striking the third slot from the forecast is appropriate for less than a few hundred bets, which describes most players. So what you do is calculate the game odds without the fringe wins to come up with a "reasonable expectation" for a lower number of bets, and the answer is actually quite surprising.
For blue and green, striking the outer bets impacts the expected return by less than 1%. The big impact from playing these lines comes from the bitcoin network fee on small bets.
For yellow there is a notable drop in the expected return using this method; a few hundred bets can expect more than 90% returns.
For red the drop is significant. The expected return is under 85% for less than 1,000 bets and below 50% for a dozen bets.
Let me re-iterate that. Your expected return on a dozen bets on red is
less than 50%. Notice that should any of that dozen bets hit one of the outer 3 slots it's a guaranteed double-up or better, and a single x9 landing paired with any non-x0.2 hit is a profit, for that same dozen bets. The chances of either of these scenarios is actually higher than your intuition might suggest, and we didn't even include the more-common profit instances of several x2/x4 landings within a given dozen bets. But hitting all x0.2 is not uncommon either; there's a very high expected variance within those dozen bets. Variance means bigger winners and losers - and that's why the game feels harder to win than it is. Anyone that's won 30
BTC probably doesn't think Lucky Bit has bad odds - it's all perception when it comes to this sort of thing, and there really isn't a mathematical model for how a game feels.