Unexpected Statistics in the House of Commons and the Media
This talk is an illustration of how statistics can take you into the public domain at lightening speed. It is based around a piece of consulting undertaken for a local police force and involved reviewing a government report which recommended the merging of the UK's 43 police forces into 12 'strategic' forces. This was a political hot potato. My report heavily criticized the statistical analysis in the government report which did not justify its main conclusion that the minimum size should be 4000 officers. It was passed to two local members of parliament who quoted from it in a House of Commons debate on the mergers. This lead to much press coverage and an interesting two days of radio and TV interviews, including Radio 4 Today and BBC2 Newsnight, and some lessons on how to deal with the press on statistical matters
A Statistically-based Opinion on the Issue ‘Does Size Matter?’ Sought by the West Mercia Constabulary in Connection with the HMIC Review ‘Closing the Gap’, Report for West Mercia Police Authority,19 December 2005
Fifteen minutes of statistical fame: unexpected statistics in the House of Commons and in the media. Significance,(2006), 2, 81-84
Sizing up the police evidence. Invited letter to Public Finance, March 2006, 3-9.
Merger report statistics ‘questionable’. Police Review, 13 January 2006., 8.
What Rate of Interest do You Pay ?
This talk will explore the concept of the internal rate of return, how it can be understood, and how it is required in new UK legislation which will harmonize with EU rules. The basic concept explored is concerned with how to present the true APR (Annual Percentage Rate) for financial transactions which have non-simple borrowing and payment plans, involving such aspects as introductory low rates, etc.
Interval interpretation for the internal rate of return. Mathematics Today, (2005), 41, 5, 153-156.
So You Play the National Lottery: Are You Crazy ?
This talk will describe the simple mathematics behind the National Lottery; explanations will be given of the published odds for winning prizes. More revealingly, the probability of not winning anything will be derived, e.g. that there is a 44% chance of no matches, a 41% chance of 1 match, and a 13% chance of 2 matches. The role of the bonus number will be explained. The probability distribution of the number of jackpot winners in a single week will be derived as the biggest binomial distribution in the world. All these results depend on the assumption that people choose their numbers at random, which they don't ? but Camelot won't tell us how they do ... in case we try to win.
Random Numbers on Your Computer
Random numbers are needed by computers for games and effects involving chance; also, more seriously, they are needed for real-world mathematical models which involve probability and statistics. Therefore, are computers secretly tossing coins or rolling dice ? No ! I shall explain why in this talk. Computer generated random numbers actually come from mathematical equations and hence are really deterministic - so how can they be random ? This is the enigma to be explained. I shall take the so-called congruential method, based on an equation of the form
xi=(axi-1 + b)Mod(m), i=1,2,?,N
with explanation and discussion. Warnings about the choice of a and b and m are needed, and a plug for theorems from number theory will be given. The requirement of an even lattice structure will be illustrated and emphasised. Illustrations of awful generators from computer manufacturers who did not know the mathematical theory will be given.