Richard Jones of softmachines.org commented about how forecasting is unreliable because of bad forecasts He lists a prediction by Glenn Seaborg, then chair of the Atomic Energy Commission, predicting in 1971 that there would be 2100 billion Kwh of generating power for nuclear power in the USA in 2000. There was 780 billion kwh. It was a relatively linear prediction based upon the buildout rate in 1971 and the projection of known license applications out to 1976. This was part of a discussion where Richard is trying to show that futurism cannot be reliable and that superlative technology predictions should be abandoned. I completely disagree with him.
Of course in 1971 there was only about 50 billion kwh of nuclear power. So the right prediction was 1600% growth over 30 years instead of 4800%. So 10% per year growth rate instead of 13.75%.
Most forecasters are not very good. However, it is a relative thing like baseball. Hitting at a 0.400 rate or higher and you are a hall of famer. Hitting at 0.150 or less and you do not make the major leagues. Plus there is the quality of the swings.
The sport of being able to spot seemingly high profile bad predictions is a mostly useless endeavor. It is like a scout picking someone for the Yankees because they played well in the government civil service leagues and then people marveling at the inadequate performance. I don’t understand man Glenn Seaborg led the Atomic Energy Commission league. He was following in the grand tradition of expert forecasters like, hmm, no one from the Atomic Energy Commission has a good track record of forecasting. The census bureau does alright but those predictions are that most people alive stay alive and get older while they are alive and we will have a certain rate of births and immigration somewhat correlate to what happened in preceding years. Man I thought Glenn had the stuff to be a good predictor of energy markets. He had that long history of being a politician and a bureaucrat. What a disappointment. I am shocked. Shocked. that his prediction was not better.
The vast majority of the impact is from those who are very good.
Plus better predictions are from those who would stand to make or lose money based upon the accuracy of their prediction. What were the best commodities traders predicting ?
Billionaire Jim Rogers, legendary commodities trader, who picked the bottom of the commodities bull market in 1999. With George Soros, Jim Rogers co-founded the Quantum Fund in 1970. I had bought his 2004 book and knew that he had a good record on commodities. He identified the impact of the rise of China on commodities well in advance.
So the flaw is in looking to regulatory bodies where there are people with a government job putting together a forecast for accuracy. Sources that are consistently wrong should not be turned to again and again for another prediction.
Do you also look for Securities and exchange commission civil servant to give you stock picks ? Would you then cite a book on how most people, even “experts”, underperform the indexes in stock portfolio performance.
The better course of action is to look and find the consistent winners in picks and predictions and strategy.
Celebrating forecasting losers who have some kind of claim to authority but inaccurate predictions is bad strategy. All I respect is proven accuracy on predictions and the ability to select the correct high impact factors.
The list of losers is long. It is useful to know why they were losers and what the flaw was in thinking that they should have been right. Learn the lessons for identifying winners. Accept the unbiased feedback of the facts and results.
Forecasting is another area to seek out those who are Superlative. Superlative forecasters: they exist too.
A good forecaster also needs to be a scout of other forecasters. As a scout of forecasters one has to have the skills to identify what quality predictions look like. It is seeing the ability of the forecaster to spot the right big trends from the root cause. Being able to know an earthquake of a certain size will generate tsunamis and where they will hit. Saying that someone who tries to throw a dart from 29 paces (the nuclear prediction) is a bad forecaster when they hit the wall 4 feet above the board is not correct. It is knowing that throwing in the right direction and hitting the barn from 29 paces was actually a decent throw. It is also knowing that the forecaster was not very good based on his use of a linear projection prediction without having a more sophisticated model.
It was actually not a horrible prediction. But it was inferior because it was only a linear projection without identification of key factors with which the projection could be updated. It was also inferior for not identifying key factors such as the potential development of nuclear fusion, vastly superior wind and solar, lower natural gas prices etc…
Jim Rogers was good not just for picking the bottom but for getting the reasons right for why there was a bottom and why there would be a commodity boom. Plus he figured out how he and those who believed him could make a lot of money by his being right.
Superlative scouts and identifiers of superlative forecasters and forecasting methods: they exist too.
It is useful to know that there can be track record for technological forecasting. There is a track record for sports futurists. I would not turn to a university professor in some field related to sports or a government official in charge of an department related to overseeing sports. I would turn to sports handicappers. People who have a track record of picking winners and whose track record has remained good over recent years (not resting on past glory). Also, I would try to look at the specific record for the specific sport. Don’t ask the college football whiz about horse racing. There are plenty of publicly available forecasts on different aspects of technology. The more useful and profitable exercise is looking at who has a good record with technology forecasting. Also, industry types who predict Microsoft will maintain operating system market share are less useful than say Steve Jurvetson, Peter Thiel types who pick startup winners that become multi-billion dollar companies.