# Lottery Forecasting And Metastatistics

Introduction   How To Forecast    Overall Performance   Performance By Year    Latest News   Disclaimers    Good, Bad And Ugly   Contact

Lottery draws are physically independent, leading most mathematicians to believe that draw history has no bearing on the outcome of a lottery draw.

"It is impossible to predict which numbers will appear in the next draw unless there is a mechanical bias."

"Lottery balls have no memory."

"Lotteries are a tax on the stupid."

Quotes like these frequently appear in the media. However an important principle is established by the Walter Penney coin-flip game: in a pseudo-random environment, it is possible to 'get something for nothing' based on sequences of results. The author has a convincing body of evidence which indicates that the metastatistics of the draw history have an influence on future draws. The data on this site is only a small fraction of that evidence.

A statistic is a measure. Statistics are sets of measures, and derivatively, the mathematical tools used to collate and summarise them. The prefix meta is often used to denote analysis of a subject at a more abstract level. In this context, the term metastatistics is used to mean the analysis of the dynamics of a distribution, both internal and external, rather than its static formulation. Hot and cold number systems are trivial examples of this use of the term metastatistics.

How can you use metastatistics to forecast the lottery? You can simulate a metastatistical forecast without using sophisticated mathematics.
• First you need a statistically pure number population to work with. For the data on this site, the author uses complete historical data for the UK main draw back to the very first draw. The statistical purity is vital - incomplete or incorrect data will render the exercise sterile.
• Collate the data to form a number distribution.
• Formulate a hypothesis about how the distribution will be changed by the next draw(s).
• Work out which numbers must be drawn in order to bring about the distributional changes to satisfy your hypothesis.
• Bet on those numbers!
Statistically, 1.35 in 1000 such hypotheses (one-tailed test) will wrongly be statistically significant to 3 standard deviations, so it is vital you monitor the performance of your hypothesis for a protracted period. Don't expect to forecast the exact 6 numbers that will appear in the next draw - just think of all the midnight oil that has been burnt by others with that objective. You should set your sights on smaller effects that others might have missed, and aim to make a regular profit rather than become an instant millionnaire. For that reason it is stupid to buy lottery tickets per se - in the UK less than half the takings are returned as prizes, the rest going to "good causes" such as subsidising the pastimes of wealthy Londoners, or financing the efforts of dubious asylum seekers to bamboozle the legal system. Consider instead betting via a bookmaker - although the maximum you can win on a single bet is less, the percentage returns are much higher. In the UK you can bet on the Irish, Spanish and New York lotteries in betting shops, and on wider range at internet bookmakers.

"Minds are like parachutes. They only function when they are open." (James Dewar)

"Genius is one percent inspiration and ninety-nine percent perspiration" (Thomas Edison)

The performance of the author's original metastatistically-based forecasts for the UK main draw is summarised below. The forecasts are generated using Quattro Pro spreadsheets. The author's original Excel spreadsheet corrupted itself over a period of several weeks, and when the author regenerated the forecasts in Quattro Pro, the opportunity was taken to go right back to the very first draw for the sake of scientific honesty, even though there are reasons for the forecasts to have underperformed at times. (A subsequent attempt to implement the spreadsheets in a later version of Excel failed because Excel was unable to cope with their size and complexity.)

 A) Number of forecasts (excluding next draw) 8154 B) Number of hits 1024 C) Expected number of hits 998.449 = (A*6)/49 D) Percentage hits 12.558 = (B*100)/A E) Expected percentage hits 12.245% = (6*100)/49 F) Standard deviations better than expectation 0.863 = (B-C)/SQRT(A*(6/49)*(43/49)) G) Break-even point betting at a bookmaker with 6/49 odds of 6-1 14.286% = 1/7

And below are the forecasts for the next draw (subject to the disclaimers below):

 Draw number Date of draw Numbers forecast 1940 26/07/14 1, 7, 15, 18, 21, 30, 35, 39, 44, 46, 47

Top of page

Results of the metastatistical forecasts in 2014

Results of the metastatistical forecasts in 2013

Results of the metastatistical forecasts in 2012

Results of the metastatistical forecasts in 2011

Results of the metastatistical forecasts in 2010
(Incomplete and incorrect - being regenerated)

Results of the metastatistical forecasts in 2009

Results of the metastatistical forecasts in 2008

Results of the metastatistical forecasts in 2007

Results of the metastatistical forecasts in 2006

Results of the metastatistical forecasts in 2005

Results of the metastatistical forecasts in 2004

Results of the metastatistical forecasts in 2003

Results of the metastatistical forecasts in 2002

Results of the metastatistical forecasts in 2001

Results of the metastatistical forecasts in 2000

STOP PRESS

Top of page

DISCLAIMERS

Top of page

Thanks to Richard Lloyd and his
UK Lottery website

No thanks to Microsoft, whose virtual monoploy of PC software means they have no incentive to alleviate the limitations of their Excel Spreadsheet product.
No thanks to the British Broadcasting Corporation, who prefer to use UK National Lottery television programmes as free advertising space rather than show the draws taking place, and sometimes don't even allow time to note the numbers down.

(Counter courtesy of www.digits.com)