Toward the end of the game, you want to get rid of as many tiles as you can so that they don't count against you. Or do you?

Yesterday, with less than 20 tiles to go, I got a few points for extending "RE" (considered a word in our group) to "rein" by adding "IN." That seemed like a bad move, because this opened up two columns for my two opponents who each got a multiple of the points that I did. Given the chance to play this again, I would exchanged tiles, or even passed.

Does there come a point toward the endgame where you "play not to lose" rather than "play to win?" That is, should you refrain from making moves that help you if you suspect that they will help your opponents more? Should you sometimes "burn" tiles and turns and let your opponents make moves that help you?

I'm assuming that you can't immediately end the game in your favor by using up all your tiles, and thereby preventing your opponents from responding.

  • I'm not sure that the specific examples you give are things that are different in the end-game vs during most of the rest of the game. As in, "refrain from making moves that help you if you suspect that they will help your opponents more" is a good idea at any time during the game. But there are probably still some things that change during the end-game, especially if you can run out the tiles and end the game while you have the lead.
    – GendoIkari
    Jun 6, 2016 at 20:59
  • 1
    @GendoIkari: Earlier in the game, if you open up things for your opponents, maybe they will do the same for you, and you'll have more chances to recoup. Whereas if this is the last two or three rounds, you make a move, your opponent responds, you may not get a "second" chance. Things change as you get closer to the "edge," either of time or of the board.
    – Tom Au
    Jun 6, 2016 at 21:06
  • A note that among tourney Scrabble players, what you described is a pre-endgame, not an endgame, situation. Endgame in Scrabble refers to when the bag is empty, so the tiles available to you and opponent (in a two-player game) are all known. In an endgame, a definite theoretical best sequence of plays for both players can be determined, because of the perfect information available. Pre-endgame is when about 1-2 moves are left for each player to make before endgame is reached. It requires different strategy. You may want to edit the misleading question title if that is what you meant.
    – Ricky
    Aug 17, 2016 at 3:31

3 Answers 3


There's a huge amount of information out there about Scrabble end-games - far more than I can summarise here - but most of it concerns two-player games. In those situations, it becomes possible to plan out the last few moves with a great deal of accuracy, since once the tiles are gone from the bag you can work out exactly what is on your opponent's rack. The key becomes maximising the points differential between you: so a 5 point move where the best response is 10 points would be preferable to a 10 point move with a possible response worth 20 points. You also need to factor in the letters that might be left on your rack at the end, since these tiles are worth double to your opponent (once when they are subtracted from your score and again when they are added to theirs).

In a 3 or 4 player game, as you describe, it is much harder to control the flow of play and work out what the other players are going to do. Therefore, it's much more about just maximising your score on any given turn. Still, there are a couple of things you can do in the endgame to help your chances:

  • Get rid of any high-scoring tiles from your rack as a priority.

  • Go for longer words if possible so that you can be the first to empty your rack.

  • As @Becuzz suggests, try not to leave any blatant double- or triple- word opportunities for the player following you.

  • If the bag is empty and you can't see the Q, you know one of your opponents has it, so try to block any tempting spots they could play it out.

Depending on the situation, some of these may be mutually exclusive, so you need to make a judgement call on which is a priority. E.g. if you have six tiles and an opponent only has two left, they will probably be able to clear their rack next go, so your priority should be to get rid of any high-scoring tiles from your rack.


The app Quackle has the useful feature of giving its valuation of those plays it lists. By comparing its valuations of various options in various circumstances I found out that Quackle values:

  • leaving 1 tile in the bag at +10
  • leaving 2 tiles in the bag at +7

The point is: once there are not many tiles in the bag, provided that there's a decent chance that you can go out, if you play, draw tiles, and leave 1 tile in the bag, your opponent must either

  • score 0, or
  • empty the bag, giving you the first opportunity to go out.

Another endgame tactic which comes into play slightly earlier is that if the number of tiles in the bag is 7 or not much larger, and you have an awkward tile like Q or J or V which would hinder your chances of going out, seriously consider exchanging because it might be your last opportunity to. In the endgame, try to go out. (I went out in 60% of my club games.) Alternatively, it might happen that it's a good idea to play so as to prevent your opponent going out -- but for you to know that, and to know what to try to do, you need to either track tiles or do a grand totting up of tiles so that you can work out your opponent's rack.


In general, this is a real strategy in many board games, especially those with multiple rounds. Sometimes you need to play in a way that is purely for your immediate benefit (get rid of as many tiles as possible and gain as many points as possible along the way). Other times a way to (indirectly) come out ahead is to cause another player harm (ie. don't play your two tiles for negligible gain when that opens up a lot to an opponent who may be able to capitalize on it). Either way, it's a bit of a gamble as to which will be more helpful. Without perfect and complete information, you will never know for sure what is best.

Some examples:

  • Say you are playing Uno and the player next to you has one card left. You suspect it is a green card. It would be a really bad move to switch the color of the discard pile to green. Unless the card they have really isn't green. Stopping another player from winning is a way to help yourself. Problem is you don't know for sure if what you do will stop them, so it's a gamble.
  • Playing hearts (or some other game where you gain unwanted points between rounds). It may be beneficial in the long game to help someone besides yourself win a round in order to cause another player to end up at a disadvantage. (ie. player A is doing really well and player B is about to go out and end the round. If the round ends now, player A will get a lot of points, making it harder for them to win and thus easier for you.)
  • In Scrabble, don't play words / letters that are easy to extend off of for negligible gain. Playing the word "the" next to a double word score tile is probably a bad idea because any one of a few letters will let your opponents extend that and get much more than you got out of it.

Ultimately, these kinds of plays are more for a bit of a long term strategy or to stall the game (if the game works like that). But ultimately it can work, you just won't really know every time.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .