The moment you put a meld on the table, opponents learn which suits you love and which ranks you're building around. If you time this right, they never get to use that information.
When you meld 5♥-6♥-7♥ on the table, three pieces of information leak out at once.
One, you're strong in hearts. Two, you're working around the middle ranks in that suit. Three, you might be one card away from extending the run in either direction, so 4♥ and 8♥ suddenly become risky discards for opponents.
Every opponent watching the table now has a defence plan. Don't feed hearts. Especially don't feed 4♥ or 8♥. If they were going to discard a heart soon, they'll hold it a turn or two to check what happens.
That's the information cost of every meld. This tip is about controlling when you pay it. On Play Rummy, you'll notice that human opponents change their behaviour the moment you meld. Sharp bots do the same.
Tip #11 says meld the moment you can. This tip says, sometimes, one turn later is better. Both are right in their own way. The trick is knowing when.
Melding early is safe. It removes deadwood risk. That's the default. But there are moments where holding your meld for one extra turn hides your suits long enough that opponents drop the exact cards you need next.
This isn't the same as "hoard your melds." Holding a meld for many turns is expensive. Holding it for one carefully chosen turn is cheap. The intermediate move is to recognize the difference.
The most common version of this tip is the one-turn hold. Here's how it works.
You have a completed meld ready. But you're close to a second meld. If you hold your first meld one more turn, you might complete both and drop them together.
Take this hand. Your first meld is 6♣-7♣-8♣. Ready to go. Your second combo is 9♠-9♥ waiting for a third 9. If you meld the clubs right now, opponents see clubs mid-range. Then next turn if you draw the 9♦, you meld the second group and now they see two of your plans in two turns.
If instead you hold the clubs and draw the 9♦ next turn, you can meld both groups in the same turn. The information hits opponents all at once, one turn later than it would have.
Is that useful? Sometimes. The information is the same total amount. But when it arrives later, opponents have fewer turns to defend against it. If the round ends in three turns, opponents got three turns of defence data instead of four. That's a small edge.
Say your first meld is worth 21 points of deadwood if it stays in your hand (three cards averaging 7 points each). If the round ends this turn, you lose those 21 points.
The chance the round ends on this exact turn is small, usually 5 to 15 percent in a mid-round position. So the expected cost of a one-turn hold is 21 times 10 percent, roughly 2 points on average.
The benefit is that opponents don't defend your suit for one turn. That's worth something. It might save you a card that would have gone into the discard pile safely. Hard to quantify, but usually worth 1 to 5 points of expected value.
A one-turn hold is roughly break-even to slightly positive. A two-turn hold is usually negative because the expected round-end cost grows fast. Three-turn holds and beyond are almost always bad.
If you're a new player, this tip is not for you yet. The bigger risk in your game right now is not melding at all. You're sitting on completed groups and losing points because you don't spot them.
Get comfortable with tip #11 first. Meld as soon as you have three matching cards. Do that for a hundred rounds. Then come back to this one.
Timing your first meld is a decision you can only make well once melding is automatic. Otherwise you'll use timing as an excuse to sit on melds you should have played already.
Here's a short checklist for the moment you have a completed meld in hand.
Hold one turn if. You have a second combo one card away and the outs are still live. You're playing against sharp opponents who visibly change their discards after you meld. Your deadwood total is low enough that a one-turn hold doesn't push you over 25.
Release now if. You have no second combo close by. Your deadwood is high (over 25). Opponents seem passive or new, so they won't defend anyway. It's late in the round and any turn could be the last.
Default to release. Hold only when you can name the reason.
Expert players sometimes time their meld not to hide info, but to shape what opponents believe.
Say you have a run of hearts ready and a set of Kings ready. If you meld the hearts first, opponents lock down on hearts. If you meld the Kings first, opponents lock down on face-card ranks across all suits. These two defensive shifts leave different cards open for your third meld.
If your third meld target is another face-card group, meld the hearts first so opponents don't tighten face cards. If your third target is another middle-suit run, meld the Kings first so opponents don't tighten mid-ranks.
This is subtle. Most players won't notice you're doing it. But over a long session, feeding opponents the meld order that leaves your third target open adds up.
Once you meld, the very next discards from opponents tell you how they're reacting. Watch for two patterns.
Passive response. They keep discarding whatever they were discarding. Same suits. Same ranks. Their hand is far from a plan that intersects yours, and they don't feel the need to defend. Good sign for you. Push forward.
Active response. They suddenly stop discarding cards in your suit. They hold cards they'd have thrown yesterday. They start discarding from a different rank cluster than before. That's defence. They see your plan and they're playing around it.
Active response means the next card you need might not come from the discard pile any time soon. You'll have to lean on the stock. Adjust.
Using this tip as an excuse to hoard. A one-turn hold with a clear reason is smart. A five-turn hold because "you don't want to give away info" is a bad habit hiding behind a good idea. Track how long you're holding melds. If it's more than one turn, you should be able to justify each extra turn.
Ignoring deadwood total. The hold is only cheap if your deadwood is manageable. If you're already at 35 points of deadwood and holding a meld pushes you higher, you're gambling in the wrong direction.
Holding against passive opponents. If your opponents aren't watching the table, the info you'd be hiding doesn't matter. There's no defence to duck. Release the meld.
Timing melds against a partner in team play. Some Rummy variants have team elements. Hiding info from your partner is silly. In those variants, meld as soon as you can regardless.
Play ten rounds where you consciously mark, at each turn, "I could meld right now, but I'm choosing to hold" or "I'm melding." Write down your reason for each hold.
After ten rounds, look at your log. How many holds were productive? How many turned into wasted turns because the round ended before your second combo materialised?
You'll quickly find that maybe one in three holds was actually worth it. Recalibrate. Hold less. Release more. That correction is where the real intermediate-to-expert jump lives.
Timing your first meld is a real skill. It's not "hoard your melds." It's the ability to pause for exactly one turn when you have a good reason, and to release without hesitating when you don't.
The default is still melding as soon as you can. The exception is a one-turn hold when you're chasing a second meld or facing a sharp defender. Two-turn holds and up are usually mistakes.
On Play Rummy, this tip only matters against strong opponents. Against beginners, meld early and let the round unfold. Against experts, timing is a small edge that adds up over dozens of rounds.