What is the IMP scale?

The IMP scale is the fixed chart team bridge uses to turn the point difference between two tables on the same board into International Match Points. A 20 to 40 point gap is worth 1 IMP, a vulnerable game swing of around 450 is worth 10, and the chart tops out at 24 IMPs for differences of 4000 or more.

Your team finishes a match and someone announces you won by 14. Fourteen what? Not points, exactly. Not tricks either. IMPs, which stands for International Match Points, are the currency of team bridge, and the scale that produces them is one of the most useful things a bridge player can understand.

Most players learn IMPs as a vague rule of thumb: "bid your games, don't worry about overtricks." Fine as far as it goes. But once you see the actual table, those rules stop being folklore and start being arithmetic.

Where IMPs come from

IMPs only exist when two tables play the same board. In a team match, your pair sits North-South at one table while your teammates sit East-West at the other. Every board gets played twice, once in each direction.

After the match, you add the two raw scores together. Say you bid and make 4♠ vulnerable for +620. At the other table, the opponents holding your cards stop in 3♠ and make ten tricks, +170 for them, which means your teammates wrote down −170. Net result for your team: +450.

That 450 then gets converted into IMPs using a fixed table. It is 10 IMPs.

The IMP scale, in full

Here is the standard scale. The left number is the score difference; the right is the IMPs it's worth.

  • 0 to 10: 0 IMPs
  • 20 to 40: 1
  • 50 to 80: 2
  • 90 to 120: 3
  • 130 to 160: 4
  • 170 to 210: 5
  • 220 to 260: 6
  • 270 to 310: 7
  • 320 to 360: 8
  • 370 to 420: 9
  • 430 to 490: 10
  • 500 to 590: 11
  • 600 to 740: 12
  • 750 to 890: 13
  • 900 to 1090: 14
  • 1100 to 1290: 15
  • 1300 to 1490: 16
  • 1500 to 1740: 17
  • 1750 to 1990: 18
  • 2000 to 2240: 19
  • 2250 to 2490: 20
  • 2500 to 2990: 21
  • 3000 to 3490: 22
  • 3500 to 3990: 23
  • 4000 and up: 24

Don't memorize it. Look at its shape instead; the shape is the whole lesson.

What the shape of the table tells you

Notice how the steps get wider as you climb. At the bottom, every 40 or 50 points buys another IMP. Near the top, you need 500 extra points for one more. The scale compresses big swings and takes small ones seriously, but only up to a point.

Three practical consequences fall straight out of that.

Overtricks are nearly worthless. Making 4♠ with an overtrick when the other table made it exactly is a 30-point difference. One IMP. Compare that with matchpoints, where the same overtrick could be the gap between a top and an average. (If you want the fuller contrast, IMPs vs Matchpoints: How Strategy Changes Everything goes into it.) At IMPs, you should cheerfully give up an overtrick to protect the contract.

Vulnerable games are worth chasing. Run the numbers on the hand above. Bid the vulnerable game and make it: +10 IMPs. Bid it and go one down (−100) while the other table makes 3♠ for +140: you lose 240, which is 6 IMPs. Win 10, lose 6. That means you should bid a vulnerable game on anything better than roughly a 38% chance. Non-vulnerable, the math is tighter (win 6, lose 5), so you want something closer to 45%. That asymmetry is why good team players stretch for vulnerable games and don't apologize for it. Our piece on when to bid game and when to settle for a partscore works through more of these borderline hands.

Disasters are capped. A catastrophe of 1100 costs 15 IMPs. One horrible board hurts. It rarely ends a 24-board match. In a Swiss or round-robin event, that final IMP margin usually gets converted one more time, into victory points.

A slam example, because slams are where it bites

You hold a vulnerable 6♠ that makes: +1430. The other table stops in 4♠ and takes all twelve tricks, +680. Difference: 750. That's 13 IMPs, right on the boundary.

Now flip it. You bid the slam and it fails by one. You're −100; the other table is +680 in game. That's 780, also 13 IMPs, the other way. Win 13, lose 13, so a small slam at IMPs is roughly an even-money proposition. Bid it when you think it's better than a coin flip. Not before.

Doubled contracts climb the brackets fast; What going down doubled really costs has the figures.

IMPs without a team match

Here is where it gets interesting for home players. IMPs don't strictly need a team of four against another team of four. Any time your table's result can be set against someone else's result on the same cards, you can IMP the difference. Some pairs events do exactly that: each score is compared with a datum (an average of the field) or with every other table, and the differences are converted using the same scale.

What you can't do is IMP a board that was played once. A single table with shuffled cards has nothing to compare against, which is why most living-room bridge ends up scored as rubber or Chicago, where a lucky run of cards can drown out good play.

That missing second table is the real obstacle, and it's what tools like Bridge@Home are built around: your foursome plays pre-dealt boards, and your result lands next to the results of other players who held the exact same hands. Once that comparison exists, the whole logic of the IMP scale (small edges, protected contracts, vulnerable games bid with confidence) starts to matter at your own home game.

The short version

The IMP scale rewards big decisions and shrugs at small ones. Bid your vulnerable games. Stop fretting over overtricks. Treat a small slam as a 50% question. And when a hand goes badly, remember the scale was designed so that one board rarely beats you on its own; the next board is worth just as much.