Count Winners, Count Losers, Count Both

The question of whether to count winners or losers confronts every declarer and commonly pops up as a question asked of experts. The quick answer is, use the easiest count, the one that clarifies the hand best for you. It won't matter, provided your head is screwed on straight, given the inviolate number of 13 tricks per hand. If you count winners (competently), you're necessarily counting losers, and vice-versa. Unfortunately, declarers' heads are not always screwed on straight, so I suggest counting both, at least if you're analyzing a full hand at leisure to see how it might have been played better. In play, doubtless many declarers would find counting both a bit too much of a task.
I have often recommended the count of winners in any type of contract, since that will tend to tell you how to play the hand to bring potential winners to fruition. However, I don't recommend an inflexible policy on this matter, and will grant that the count of losers will often be the most convenient count. I only point out that when you count losers, you must count out from one hand or the other, accounting for every card in that hand. Every card must be either a winner in its own right, or go on a winner in the other hand, or be counted as loser. Can you live with that? Are there any ways of cutting the number of losers down? Answering those questions is the task before you.
But what you don't want to do is to count losers in one suit in one hand and losers in another suit in the other. That will commonly bring grief, as discussed here. So lemme give you a hand that is equally amenable to counting winners, to counting losers from the closed hand and to counting losers from dummy:

K Q J 6
K J 9 4
A 7
A 9 6
10 8 2 7 4 3
A 7 5 3 2
Q J 9 8 4 K 2
8 5 4 2 Q J 10 7
A 9 5
Q 10 8 6
10 6 5 3 Contract: 4 hearts
K 3 Opening lead: Q of diamonds

This hand certainly filled the bill (of being amenable to the count of winners, of losers from dummy and count of losers from the closed hand). Only one wrinkle kept this hand from being the flattest of the session -- and perhaps one of the least interesting. That, of course, is the 4-1 split in hearts. But it proved to be a costly wrinkle to those who weren't careful, or perhaps it would be better worded, who were too complacent about knowing they'd make their contract, for the cost to those who came up with only 10 tricks was a horrendous 54 matchpoints! I've seen declarers lose less for going down in a makable contract, and quite often find that those who miss a slam, over which partners so commonly get uptight, lose far fewer matchpoints.
Well, lemme go through the counts. The easiest, I would say, would be to count from dummy. Let's see. Spades are solid. We're always losing a heart and we're always losing a diamond. Which leaves only the club suit where we've got a loser we can do something about. On a 3-2 trump split, that third round of clubs would be virtually automatic. So whatever the opening lead, let's get out the trump. Oh! West shows out on the second round. Okay, we'll hafta ruff that club before finishing with hearts. Here's one declarer who got a diamond opening lead, knocked out the A of hearts and got a club return, taken with the K. Should declarer be afraid of a singleton club (and thus suffer a ruff on the second round) or a doubleton (and thus an overruff on the third round), he's got to ask himself what alternatives he has, which here are none! The third round of clubs must be ruffed or accepted as a loser. There's no other card that can take care of the third club. To be sure, declarer has a few protections. One is the number of clubs they have. With 8 cards in the suit, the opposition figures to go the third round without a ruff. Also, if there is a 6-2 split in the suit, the short hand might be East, playing second hand to the third round and thus before declarer has to make a decision. And finally, this is duplicate. If the cards just aren't distributed so as to make ruffing that card feasible, he's got to remember that the people he's competing against are facing the same situation. Anyway, after taking West's club shift, declarer wins with the K, goes to the A of clubs, and ruffs. He then takes a second round of trump, noting the bad split, and so having taken care of all losers in dummy, declarer finishes drawing trump, and claims, conceding a diamond loser. One overtrick, losing only a diamond and a heart.
Or counting from the closed hand: Forget clubs. We have a K doubleton opposite the A. Spades are still solid, now offering a fourth winner on which to throw a loser from the closed hand. We note that same loss of a diamond. Nothing can prevent that from winding up a loser. But please note now that if we're forgetting clubs, since every club in the closed hand is accounted for, we have the 3rd and 4th diamonds to concern ourself with. Well, we surely can throw one on the fourth round of spades, which means we've got to ruff the fourth round of that suit, which means that on the 4-1 split in hearts, we've got to ruff before finishing the drawing of trump. Otherwise, we're going to have two diamond losers, in addition to one in hearts.
Which brings us to a count of winners: Four spades, two clubs, one diamond and how many hearts? Well, three if we simply lead out hearts after knocking out the A. That's 10. Can we produce another trick? And the answer to that should be obvious. We can get a fourth heart winner by ruffing in one hand or the other before drawing all trump, which is feasible, given good communication in spades. And of course, it should be realized.
Opening lead the Q of diamonds. Win in dummy and knock out the A of trump. West wins, hits his partner with the K of diamonds and that worthy leads a trump! Hm-m-mmm. Okay. We run three rounds of clubs, ruffing the third in the closed hand, where we now have one trump facing two in dummy, the number East also has. Now cash out all winners, starting with trump, of course: That's four trump winners (one on the lead from East, one on the ruff of a club, and two being the number in dummy). Run all side-suit winners and that makes for 11 tricks.
Take your pick. But bear in mind that it's not always that simple. Here are some traps to watch out for: (1) I have already referred to the not too rare mistake of counting losers in clubs in the closed hand, of which there are none, with K doubleton opposite the A, and then diamond losers in dummy, of which there is one (with the A and a spot card), and from that merrily drawing trump and . . . and . . . wondering why you have three losers! No, if you play "from" dummy, you've got to take care of the third club, and if you play from the closed hand, you've got to cash the A and lose a diamond before drawing all trump, or even drawing three rounds, so that when you lose that diamond, a defender won't be able to wipe out your trump holding on a simple trump lead.
Particulatly with the 4-1 trump split is it necessary to keep your mind on doing what you can do with all the cards in this hand or that hand. As mentioned above, if trump split 3-2, the necessary ruffs would be virtually automatic, since you need only one in each hand. After three rounds, you'll ruff one club in the closed hand, ruff one diamond in dummy, throwing a second diamond on a long spade.
Other traps that declarers fall into are losing the potential to sluff losers by not watching entries, failing to keep stoppers, thus letting the defense cash out losers, running too many trump leads, as mentioned above, or not enough trump leads, blocking a suit and thus missing the potential of valuable cards. Oh, there are many ways of screwing up a hand, which is why I've got about 25 or 30 categoriesl.
See also You should be glad. You should be glad about these subtle wrinkles offered by the cards as opposed to the simple cashing of top winners. For two reasons: First they make the game a little more interesting, wouldn't you say? Do you want a game where you just cash top winners? Or one where you're asked to use your noggin a bit? And secondly, these wrinkles are what allow you to move ahead of your indifferent and careless peers.
Just about everybody was making the contract, with the upshot that an overtrick became a very valuable trick. Those who were content with 10 tricks got 18 (on the basis of 100), while those who picked up an 11th winner got 54! Fifty-four matchpoints for saying, I'd better get a ruff before drawing all trump?
I've often pointed out that the disparity in scores by virtue of missing an overtrick (or two) is directly proportional to the ease in which it (or they) can be picked up. If an overtrick can be realized only by some criss-cross, upside-down squeeze, maybe one or two will surge ahead, but you'll have company. Easy to pick up? You're gonna lose a lot when you don't make sure you pick those tricks up. Like 36 out of 100!