A 2011 article reposted at Australia's Quadrant Magazine, October 24, 2025:
Everybody makes forecasts. The high end engages writers, think-tanks, governments and business organisations making predictions on such weighty matters as the future of humanity, climate change, earthquakes, tomorrow’s weather, the economy, election outcomes, urban development, job prospects, and the costs and benefits of one or another grand project. At the low end perhaps as much effort is expended on fortune telling, gambling strategies and athletic events involving human, equine, canine or other competitors.
With this vast range, forecasts take many forms, use many different methods, and serve many agendas. In the form of dystopias they can provide powerful warnings, as in Orwell’s Nineteen Eighty-Four. In the form of economic predictions, they can influence the policies of whole nations. They can give hope, they can encourage, they can terrify. In the brief exposition of forecasting which follows I will draw largely upon personal experience, first as a professional forecaster, then as a consumer of others’ forecasts.
While working as a research scientist with the now defunct Forests Commission in Victoria, I developed computer models to simulate the development of even-aged stands of timber in order to facilitate forest management. A good range of quantitative studies on the growth and development of appropriate stands of trees was available. These studies equipped me to describe with some precision the rates at which the trees got bigger and fewer through time, and under various site conditions and cultural regimes. Converting the data into an interactive forecasting model required (I realised in retrospect) a surprisingly large number of processes:
• interpolation—putting smoothed trend lines through (but not beyond) data sets;
• induction—drawing a generally applicable conclusion from experience, in this case from observations of sample trees and plots of various age;
• deduction—the logical progression from premises to conclusions, here based on studies of sample plots thinned to various degrees;
• exclusion—the deliberate ignoring of certain possibilities such as bushfires (“no-surprise scenarios”);
• extrapolation—the extending of trends beyond the limits of the given data set.
In anticipation of what I say later about the hazards of extrapolation, in the above instance the trends were extended beyond any substantial data set for only one purpose, namely to ensure that the effects of great age were properly simulated, with all the trees eventually becoming moribund and dying, rather than continuing to grow more or less as in their prime. Few actual cases could be found, but limiting functions could be set in the light of general experience, and were unarguably sound.
Through such measures the forecasts ran as it were on sets of tramlines. In reality of course things don’t always go according to plan—forest stands are consumed by wildfire, or become diseased, or devastated by strong winds. But the prospective users only wanted to know the outcomes of stands which developed according to plan—that is, in the “no disaster” case. This is an important distinction, as will be shown in due course. By all reports the model has proved suitable for most of the intended purposes, though in the years since the model’s publication others from time to time have sought to improve its precision and to eliminate certain biases.
The law of induction figured prominently in my forecasting method. By way of example, because the sun is known to rise every morning without exception, it will do so tomorrow. In mathematics proof by induction is standard and unobjectionable. But in the real world forecasts based on induction occasionally go horribly wrong—as when our planet Earth finally stops revolving, or, as in the lesser disaster I cited for forestry, timber stands are badly damaged by wildfire. Such possible disasters were explicitly and openly excluded—my forest simulations were straightforward things intended solely to aid management in unperturbed conditions. There was no hidden agenda—unlike many forecasts I was to encounter later in my career.A few years later, when seconded to the Victorian Premier’s Department, I was given the task of forecasting the population growth of Victoria by regions. The work was required because of a situation arising out of the legitimate ambitions of certain major government agencies. These agencies—several now defunct—included the Melbourne and Metropolitan Board of Works, the Geelong Regional Commission, the Albury-Wodonga Development Corporation, the Country Roads Board and the State Electricity Commission. Most if not all these agencies had been making their own forecasts to facilitate planning, and, not incidentally, to bolster cases for a bigger slice of public money.
To assist me, an officer was seconded from the Australian Bureau of Statistics (ABS). The ABS secondee provided most of the initial material, starting from projections of Australia’s future population for a range of assumptions about net immigration levels, fertility and mortality. The ABS policy was to avoid the controversy often occasioned by definitive forecasts by drawing the line at simple projections—that is, sets of extrapolations showing the outcome if this trend or that was to continue unabated. It then fell to me to nominate the “most likely” scenarios, thereby transforming the ABS’s squeaky-clean projections into dinky-di forecasts, which bristled of course with political implications.
The first stage was to estimate Victoria’s projected share of the total Australian population growth using assumptions about future trends in net interstate migration. At that time Victoria was showing net losses, mainly to Queensland, but year-to-year data were highly variable. On little more than a gut feeling, I went for a smallish but positive net growth level for Victoria. Next came the most politically sensitive part—allocating Victoria’s forecast growth between regions. Melbourne region’s share had been growing steadily, and most other regions were more or less holding their own, save for the Mallee, which had been in decline for some years. From this basis, and taking a number of regional economic and social factors thought to be driving these trends, I made “no surprise” thirty-year forecasts for population growth in each of the regions. Here, too, as with forest modelling, the question of major disasters was quite beside the point.
When, via the State Co-ordination Council, we released the forecasts, all hell broke loose. With few exceptions, each major agency cried out that it had been hard done by. My defence was simply that, under even the most generous growth assumptions, there would be nowhere near enough additional Victorians to satisfy the agencies’ joint requirements. It was a bit like pointing out to a gathering of the world’s religious faiths that they can’t all be right.....
....MUCH MORE
Part of the outro from a May 2025 post.
And August 2014 - In Other Caliphate News: Boko Haram Says No, It Is The Real CaliphateBoth Abubakar Shekau and Abu Bakr al-Baghdadi are claiming to be the successor to Muhammad, one of them must be wrong.*
From the Daily Mail....
