Let's start with a rules recap for those who just want to know how to play:
On a turn, you roll the six dice up to three times, keeping and re-rolling dice as you wish, to simulate a dive. Dice show the numbers 1-5 and a treasure chest, and when you stop rolling, you place shells one of the levels that you reached. If, for example, you rolled 1-1-2-3-5-5, you place two shells on level 1, one shell on level 2, or one shell on level 3. You can't place shells on level 5 because you didn't dive through level 4 to get there. If you roll a perfect sequence of 1-2-3-4-5-chest, then you grab one of the five random treasure chests, which are worth 5-8 points, then take another turn. If you don't roll a 1, then you don't place any shells at all!
When the fifth chest has been claimed, the game ends immediately. Alternatively, when someone places their last shell, each other player takes one final turn. On each of the five levels, whoever has placed the most shells scores the main treasure token for that level; whoever has the secondmost shells takes the secondary treasure token, and in a game with 4-5 players, the player with the thirdmost shells also scores. Ties are broken in favor of whoever has the most shells on the level immediately above the one being evaluated. Players sum their points from tokens and chest, and whoever has the highest score wins.
Well, that depends on why you're rolling that die and what you need for a result. If you have a one-in-six of succeeding, then you're might have this rod of hope buried under the expectation of failure, ready to give a shrug of acceptance but sure to explode with a yell should that lone face end up on top. For a one-in-six chance of failure, you'll be chanting "C'mon, c'mon, c'mon, c'mon, c'mon" until that die stops rolling and does what it's supposed to do — lock in your victory. Don't even think abou— ARE YOU KIDDING ME?!
Once you add more dice to a roll, you have more possible outcomes — and if a game designer is doing their job, they are going to make you care about some outcomes more than others. On top of that, they're probably going to give you the chance to re-roll dice. That's the key to the success of Yahtzee, a.k.a. Yacht, and the hundreds of dice games that incorporate Yahtzee's "roll up to three times while locking in whatever dice you want" structure. The first roll is random, and while you hope for something magical, that roll is generally laying the groundwork, suggesting possible outcomes and poking you with options. You think about what could happen, then re-roll, unable to force it to happen but hoping all the same. If it didn't work, well, you have one final chance to bring it all home, possibly on the aforementioned one-in-six.
Into the Blue uses the Yahtzee structure, but like so many other dice games, it improves on that standard by making you care about different things over the course of the game. Sure, ideally you roll that 1-2-3-4-5-chest because that nets you free points and another turn, but to make it happen, you'll need to get stupidly lucky. Most of the time, though, you'll place one or more shells on a level, and if you have the lead on that level or the tie breaks in your favor, you mentally lock in those points. Those are your points. Yours.
Never mind that they're not actually yours until the game ends and your majority holds. A brilliant feature of Into the Blue is that due to what you need to roll to place shells on a level, you can rarely be sure that a majority at the deeper levels will stand. On the fifth level, for example, you can place at most two shells at a time (with a roll of 1-2-3-4-5-5), so a player can rarely dominate that level, which means it's almost always in play — and if that level is competitive, then you also care about what you have on level four since that will determine who gets 10 points and who has to settle for only 5, and the points on level four are 8 and 4, so that's not shabby either.
Each time your turn comes around, you're looking for certain results on the dice to fend off an attacker on this level or to grab available points on that level or to tip a tie-breaker towards you, which will likely win you points elsewhere at the same time. You have a shifting catalog of results that you want to see — whether you're rolling or someone else is — and you care about each turn because you never know what might happen.
As in other majority games, you want to win levels with as small a margin as possible so that you can place additional shells somewhere else and ideally clamp on points there, too, but this is a dice game, after all, so who knows where you might end up? Will you place another two shells on level 1 because you failed to roll a 2 yet again? Will you place nothing because the 1s are proving too elusive — or are you trying to cheat the odds and not re-rolling "enough" dice, whatever "enough" means? Sure, constant failure can be frustrating, but that's also part of the joy that comes with playing a dice game: Unexpected things happen and make a particular playing of the game memorable.
Knizia does all the right things in Into the Blue, using dice to their maximum effect by giving you and your fellow players a tangled web of desirable objects that will have you trying to place a foot on every string to keep it under your control. You'll probably fail to do so, but you won't know for sure until the dice stop at game's end.
For more examples of gameplay and a remarkable streak of luck that you might never seen on your table, check out my video overview: