Most matchpoint advice is about protecting a good score. Take your overtricks. Don't go down in cold contracts. It's sound advice, and it quietly assumes the board is still alive when the play starts.
Sometimes it isn't. Sometimes the auction already handed you a bottom, and the only question left is whether you notice.
What a lost board looks like
Say you and partner hold a 4-4 heart fit and 26 points between you, but the auction wandered and you landed in 3NT. The lead comes, dummy goes down, and you count nine easy tricks. Plus 400. Feels fine.
It isn't. Most of the room is in 4♥, and on this layout 4♥ makes ten tricks for +420. Your +400 beats nobody who found the fit. On a matchpoint scoresheet, where each pair you outscore earns you a point and each tie earns half, a steady nine tricks is worth close to nothing.
So what does a tenth trick cost you? If some risky line produces it, you're +430 and suddenly you beat every pair in hearts. If the same line fails and you go down one, you're minus 50. Which is also close to nothing. The disaster you're being careful to avoid has already happened; you just haven't seen the traveller yet.
That one hand holds the whole idea. Once your contract sits below the field, the safe line locks in the bad score and the gamble is nearly free.
Why this feels wrong at the table
Everything in a declarer's training pushes the other way. Secure the contract. Never throw away a sure thing. At IMPs that instinct is exactly right, because the gap between +400 and minus 50 is ten IMPs, and nobody loses sleep over 30 points. Rubber bridge agrees.
Matchpoints doesn't care how many points you scored. It cares how many pairs you beat, and that single fact changes which risks are expensive. It's the same reason matchpoints reward precision so ruthlessly: ten points can separate a top from a bottom.
The emotional problem is bigger than the technical one. Going down in a cold 3NT feels humiliating, and nobody enjoys explaining it to partner. But partner will be staring at the same bottom either way. One version of it had a chance.
The reverse: a board that's already won
This cuts both ways, and the second direction gets forgotten more often.
Imagine you stretched into a vulnerable 4♠ that you suspect most pairs won't bid. Ten tricks is +620. The pairs resting in 3♠ score +170 even when they take ten. You're ahead of all of them the moment your contract comes home, so forget the overtrick. If a safety play guarantees ten tricks at the cost of an eleventh, take it without a second thought. The extra 30 points buys nothing against a field sitting at +170, and going down turns a top into a zero.
Defenders can use the same logic. The opponents stopped in 2♠ and are wrapping up ten tricks while the rest of the room bid game? They're +170 against a field of +420. Don't invent a risky defense to hold them to nine. The board is yours already. Defend quietly and let them have their overtricks.
The hard part is reading the field
None of this works if your picture of the field is wrong, and most players are worse at it than they think.
The usual failure is imagining a disaster that isn't there. You're in 3NT with a 4-4 major fit, sure, but both hands are flat, 4-3-3-3 opposite 4-3-3-3, and half the room chose notrump for the same reason you did. Throw away the contract on a normal board and you've turned an average into a zero all by yourself.
Before you start swinging, ask yourself whether most pairs playing standard methods would reach your contract, and be honest about how strange your auction was. Then ask the question that catches people: does the risky line actually beat the field's number, or does it just lose by less? Going from minus 100 to minus 50 when everyone else is +140 changes nothing. The only gambles worth taking are the ones that climb over the field.
It helps to remember what field tendencies really means: you aren't playing against the two people across the table so much as every pair who will hold your cards. And some boards really are unwinnable. Don't chase those.
Where the read comes from
Nobody develops a feel for the field by guessing. You develop it by checking. Every time you finish a board and see where the other tables played it, you recalibrate a little: I was sure everyone would be in hearts, and a third of them stopped in 2♥. Fifty boards of that and your guesses at trick one get noticeably sharper.
That checking step is exactly what most home games lack. One table, no comparison, and the idea of a lost board doesn't even exist, because there's no field to lose it to. Bridge@Home fills that gap by having your table play pre-dealt boards and scoring each result against other players who held the same cards, so afterward you can see whether your +400 was a triumph or a quiet zero. It won't make the decision for you at trick one. It will keep telling you whether your read was right.
The takeaway
Before you plan the play, ask something different from the usual. Not "how do I make this?" but "what does the room score here, and can my normal line beat it?" If it can, protect it. If it can't, safety is just a slower way to collect a bottom.