It holds 50,000 random deals – a compromise number giving acceptable
search times and acceptably varied deals. For each deal, information is noted
about the cards dealt to each **seat**, the joint holding of each
**pair** and the **deal** overall. Sets of deals can be selected on
any or all of the three.

Information | Seat | Pair | Deal |
---|---|---|---|

High card points | Sum of A=4, K=3, Q=2, J=1 | Sum of hcp of the two seats | Difference between the hcp totals of the two pairs |

Length points | One point for the fifth and subsequent card(s) in each suit. | Sum of lep of the two seats | N/A |

Losing trick count | Count of missing AKQs for each suit, mitigated by suit length. | Sum of ltc of the two seats | Difference between the ltc totals of the two pairs |

Balanced | 0 = the hand is unbalanced 1 = the hand is balanced i.e. no singleton or void and no more than one doubleton. |
The number of balanced hands in the pair | The number of balanced hands in the deal |

Shape | The no. of cards in each suit, longest first e.g. 5-3-3-2. Actual suits are not indicated. | N/A | N/A |

Fit | N/A | 0 = the pair have no fit 1 = the pair hold 8 cards or more in one or more suits |
The number of pairs with a fit. |

Controls | N/A | f dash s where f is the number of suits with first round control and s the number with second round control. | N/A |

Deals in the database were generated randomly and the randomiser is pretty good. In high-volume tests, results conformed closely with predicted distributions. With our 50,000 deals, expect only minor anomolies.

From time to time, the database will be refreshed i.e. deleted and re-spun with new, random deals. This will prevent over familiarisation with its deals - particularly the low probability, extreme distributions so often re-visited.