If all players are experienced, there can be complex signals passed via draw and discard decisions. No doubt reams could be expended on this subject, but here are just a few simple observations:
Obviously, you should watch for whether the player after you picks up the card you discarded. This is usually a clear signal that their crew includes this color. Sometimes, however, it can be a signal that they are planning to change the tavern to this color and then sail.
Conversely, when the player discards a color, it usually means that their crew is not that color. Of course, it could be a duplicate number, so the first discarded card of a color doesn't tell you much for sure.
What does it mean when a player discards a low valued card? Sometimes it means that they are holding a very low score. However, it is quite possible to be holding a hand of 20-some points and still not be able to do better than discarding a 4 or 5. It is less possible to have a bad score and discard a 2 or 3. It's possible to bluff by discarding a low card, but I think that this is rarely profitable.
It is fairly often a good idea for the very first player to sail immediately. This is because each player can drop an average of about 10 points on the first turn, for instance by drawing a card that matches their crew and discarding a high valued card, or by changing the tavern color to match more cards in hand. Sailing when the other players have not had a chance to do anything confers a large advantage. Of course, getting skunked early costs a lot of points as well.
I wrote a program to simulate many first turns and check the outcome of sailing. It uses a very good, but not totally optimized, strategy* for lowering its score before sailing.
There are two measures of whether it is a good idea to sail. The first is the probability that you are not skunked. However, it is possible to be skunked and still have the average of the other players' scores be greater than your score. Which measure is more appropriate depends on what the score is so far. For instance, if you're in the lead near the end of the game, it doesn't help you if the average score of other players increases if the person who skunks you moves into first place as a result. Fortunately, the odds are not very different between these two measures.
| # of players | Sail with ≤ this and on average your score, after all is said and done, will be less than your opponents' average | Sail with ≤ this and you will be skunked less than ½ of the time (more detail) |
|---|---|---|
| 2 | 23 | 26 |
| 3 | 20 | 21 |
| 4 | 19 | 19 |
| 5 | 18 | 17 |
| 6 | 17 | 16 |
| 7 | 16 | 16 |
And here's how often it happens, according to my program:
| Score | # of players for which this is a cutoff (cross reference to above table) | Probability of sailing with ≤ this if the discard stack is seeded (Rio Grande rules) | Probability of sailing with ≤ this if the discard stack is not seeded (my rules) |
|---|---|---|---|
| 16 | 6, 7 | 14.5% | 12.8% |
| 17 | 5, 6 | 17.8% | 15.8% |
| 18 | 5 | 21.4% | 19.2% |
| 19 | 4 | 25.4% | 22.9% |
| 20 | 3 | 29.7% | 27.0% |
| 21 | 3 | 34.2% | 31.2% |
| 23 | 2 | 43.6% | 40.3% |
| 26 | 2 | 57.8% | 54.3% |
(These probabilities can all be bumped up another 0.3% if you use an inhumanly optimized method for choosing when to draw.) In the moment of taking the first turn, of course, knowing these latter probabilities isn't useful, but perhaps knowing the likelihood of getting hit by this when you are not the first player could modify your cross-round strategy. I mostly display these probabilities to illustrate why 2 player games are silly and why not seeding the discard stack is reasonable.
I have not developed anything better than a vague feeling about what scores are good to sail with later in the game. As stated above, each player can drop an average of 10 points on the first round. This means that by the second round, a good score to sail with has dropped dramatically. It does not drop by as much as 10, though, since score reductions often depend on the tavern and can be undone by the next player.
I would guess that in the second round, scores 5 points less than those listed above are good bets. So in a 4 player game, sailing with 14 or less is probably ok. The first player again has some advantage in the second round — having had twice as many turns as everyone else — but not as much. From there on, it becomes more difficult to drop points, so the rate at which a "good score" decreases slows down and the turn order matters less and less.
In a very long round, it is quite possible to sail with as little as 6 points and still be skunked. However, being skunked isn't nearly as bad so late in a round, since you're only likely to be picking up 20ish points instead of the 50 or more that you risk if you sail early.