WIN379: This Week’s Rate and the Overall Rate Can Both Be Right

Compare two sample periods in WIN379 statistics with a worked weekly and earlier record. Learn why totals need counts and why percentages cannot simply be averaged.

This week: eighty per cent. All recorded rounds: sixty per cent. Those figures can describe the same participant without contradiction. Their denominators refer to different stretches of the record. Anyone reading a WIN379 statistics discussion should ask which rounds belong to each figure before judging the arithmetic. A percentage is not a permanent attribute attached to a person. It is a summary of a defined collection of observations, and that collection needs a beginning and an end.

Twenty rounds in two periods

Our invented practice ledger contains twenty completed rounds. The earlier ten show four wins and six other results. The current week's ten show eight wins and two other results. We define the rate here as wins divided by completed rounds; no draws, abandoned rounds or unresolved results are included in this particular ledger. That explicit definition matters because another report might use a different denominator. All the calculations below follow the definition just given.

The current week has eight wins in ten completed rounds: 8 ÷ 10 × 100 = 80%. The earlier period has four in ten: 4 ÷ 10 × 100 = 40%. Combining the records gives twelve wins in twenty rounds: 12 ÷ 20 × 100 = 60%. The weekly and combined figures answer different questions. Neither needs to replace the other, and calling one “the real rate” without naming the period loses information.

The dates inside the rate

A date boundary makes “this week” less slippery. For a worksheet, define the interval as Monday 5 October 2026 at 00:00 IST up to, but not including, Monday 12 October at 00:00 IST. The included record is determined by the chosen completion timestamp. A round beginning before the boundary but completing inside it belongs according to that declared rule. Another report might use start time instead; the difference must be stated rather than silently assumed.

An all-time figure has a boundary too, even when its label sounds unlimited. It may mean everything since a record began, everything retained in a particular account or everything included by the report's current filter. In our example it means exactly the twenty listed completed rounds. A WIN379 comparison should not expand that claim into every round ever played. A record can only summarise observations that its defined scope actually contains.

Why the percentage average fails

The equal-size periods above make the average of eighty and forty happen to equal the combined sixty. That coincidence can teach a wrong general rule. Change the example: an earlier ninety rounds contain forty-five wins, while a current ten contain eight. The rates are fifty and eighty per cent. Their simple average is sixty-five, but the combined record is fifty-three wins in one hundred rounds, or fifty-three per cent. Period sizes make the difference.

The correct combination adds counts first: total wins divided by total completed rounds. It effectively gives the ninety-round period more weight than the ten-round period. A reader does not need a technical statistics vocabulary to verify it. Place forty-five marks in one box and eight in another; combine them over ninety plus ten opportunities. Fifty-three of one hundred is visible immediately. Averaging two headline percentages hides those different amounts of evidence.

We would present both the fraction and percentage whenever space permits: “8/10 this week, 80%; 12/20 in the complete worksheet, 60%.” The fraction lets readers reconstruct the calculation and see the sample size. Two participants can share an eighty per cent rate through four of five or forty of fifty; their records are not identical in length. This observation describes the data only. It does not establish what either participant's next result will be.

A changed figure may also reflect a changed window rather than a newly added round. When a week rolls over, older observations may leave the weekly summary while remaining in the total. Test this with our first ledger: remove two wins from the weekly set and the rate becomes six of eight, or seventy-five per cent, although no new loss was added. The membership of the sample changed. The explanation comes from which records moved, not the headline alone.

Try the closing question: has this week's eighty contradicted the ledger's sixty? No. Eight of ten belongs to the week; twelve of twenty belongs to the whole worksheet. The numbers can coexist because their scopes are explicit. In WIN379 statistics reading, the useful phrase to complete is “per cent of which rounds?” Once that answer is visible, a percentage becomes an understandable calculation rather than a label that appears to change its mind.