Every bridge argument eventually lands in the same place. You held 15 points and a flat 4-3-3-3, partner invited, you passed, and 3NT would have made. Partner says you should have accepted. You say the hand was a minimum with no shape. Both of you are quoting rules you half remember, and neither of you is going to change your mind before the coffee gets cold.
There is a way to actually settle that argument. It just isn't the way most club players think of first.
What a bridge hand simulation does
A simulation takes the cards you can see and deals the rest at random, over and over, a few thousand times. Each fake deal has to fit what you know: if partner showed 8 or 9 points with no four-card major, every simulated partner hand matches that description. Then a solver plays each deal out double-dummy, meaning with every card visible, and counts tricks.
At the end you get a number. Say 3NT makes on 44% of the deals that fit the auction. Now the argument has something to stand on.
Programs like Thomas Andrews' Deal and Hans van Staveren's Dealer have been doing this for decades, and the double-dummy solvers behind them are fast enough that a few thousand deals take seconds.
Why the number matters more than it looks
That 44% means very different things depending on how you're scored.
At IMPs, vulnerable, a game that makes around 40% of the time is usually worth bidding, because the swing when it makes is far bigger than the swing when it fails. At matchpoints you want something closer to a coin flip, since a plus score in a partscore beats every pair who went minus in game. (The post on IMPs vs matchpoints walks through why the scoring flips the math.) So a simulation never hands you a verdict on its own. It hands you a percentage, and you have to bring the scoring to it.
This is the part I love. It turns "I had a feeling" into "the hand is a 44% game and we were playing pairs." One of those sentences ends an argument. The other one starts the next one.
Where simulations quietly mislead
Double-dummy play is perfect play. Nobody at your table plays that way, including you.
The solver always finds the winning opening lead. It always guesses the queen. It never forgets to unblock. Real defenders lead fourth best from their longest suit and hope, which means a contract that's 50% double-dummy might be closer to 60% at an actual table, because the defense gives back a trick on lead often enough to matter. Sometimes it cuts the other way. Declarers in real life misjudge which finesse to take, and the solver doesn't.
There's a longer explanation in what double-dummy analysis really tells you, but the short version is this: the simulation tells you what the cards can do. It doesn't tell you what people do with them.
The other trap is garbage in, garbage out. If you describe partner's hand loosely (any 8 to 10 points, any shape) the sim deals hands partner would never have bid that way. Your 44% could really be 35% or 55%, and you'd have no idea. Good simulations spend most of their effort on the constraints, not the dealing.
Do club players need this?
Honestly, most don't need to run one themselves. What they need is the habit the simulation teaches: a bidding decision is a bet on a distribution of hands, not on the one hand you were dealt.
When you passed that invitation, you weren't wrong because 3NT made. You'd only be wrong if 3NT makes often enough, across all the hands partner could hold, to beat the scoring threshold. One result can't tell you that. Fifty can't really either. Thousands can.
It changes how you talk about bad results, too. Was the line a 50% shot or a 75% shot that happened to fail? The game-or-partscore decision gets much calmer when you think in percentages.
The missing half: real tables
So simulations give you the theory. What they can't give you is the field: what actual players, with actual nerves and actual leads, did with the exact same cards.
That's the half you get from comparison. At a duplicate club, the traveler shows you that six pairs bid the game and four made it, which is a different and very human kind of evidence. Most home games never see that. You play the hand, write down a number, shuffle, and the information is gone. Tools like Bridge@Home close that gap for a home table by having you play pre-dealt hands and then comparing your result with other players who held the same cards, so you see what the field did instead of guessing what it would have done.
Put the two together and you get something pretty powerful. The simulation says the game is 44%. The comparison says most tables who bid it went down, because everybody led the same killing suit. Now you know something about the hand and about the people who play it.
The takeaway
Next time you're stuck in the "should we have bid it" argument, ask a different question. Not "did it make," but "how often would it make, and what does our scoring pay for that?" You don't need software to ask it. You need to stop treating one deal as proof of anything.
And if you ever do run a simulation, check the constraints twice. Partner's hand is the whole experiment.