The Predictors’ Scorecard

Israel’s elections are but several weeks away – and the press is full of who will win, who will lose, and what does that mean for Israel. Not to mention all the other news topic prognostications. Overall, is this something you can trust?

As you’re presently reading The Times of Israel, especially this blog section, you probably have a favorite news analyst (or a few). Part of what commentators do is to try and predict the future course of the issue at hand. Indeed, if you compare newspapers (especially the print versions) from a few decades ago to present ones, the difference is striking. Back then, more than 90% of the front pages was taken up by “hard news”: relatively straightforward reports as to what happened yesterday. Today, by contrast, it’s almost all commentary, analysis, and prognostication: “what does today’s news mean, and where’s it going?”

But how accurate is such futurizing? More to the point, how (un)successful is your favorite analyst-predictor? After all, why bother reading someone whose success rate is lower than a university passing grade of 60? Perhaps there’s someone else who does a much better job than your favorite.

Now there’s a relatively easy way to find out: ask AI! Of course, its judgment on such a question cannot be perfectly accurate for the simple reason that “predictions” in the socio-political sphere are not like ball games with a clear winner and loser. But an AI analysis – especially if prompted to explain its assessment – can be highly informative, even eye-opening.

To show you what this entails, I asked Google’s Gemini the following about arguably the world’s most famous news commentator/predictor these days. My prompt went like this: Three-time Pulitzer Prize winner New York Times commentator Thomas Friedman makes many predictions in his column. Over the last ten years, what percent of his predictions: 1) came true; 2) proved to be wrong; 3) were partly true and partly wrong. Provide a few examples of each.

I have edited Gemini’s reply here for brevity:

Friedman is a “big picture” thinker, fond of metaphors like “the world is flat”. His forecasting record is a fascinating mix of prescient warnings and audacious optimism that failed to materialize. A rough estimation of his prediction track record over the last ten years:
• 30-35% Came True (The Prescient): Often on technology, climate, and general globalization trends.
• 35-40% Proved to be Wrong (The Follies): Often on geopolitics, specific foreign leader actions, or the immediate pace of change.
• 25-30% Partly True/Partly Wrong (The Mixed): These are his “Yes, but…” predictions.

Friedman’s strongest predictions often lie at the intersection of technology, climate change, and globalization: 1) The rise of artificial intelligence (AI) and digital connectivity; 2) The “4th Industrial Revolution” and climate change; 3) The need for “smart” infrastructure.

Friedman’s mistakes often stem from excessive optimism about the rapid adoption of democracy or a failure to grasp the resilience of local political cultures and authoritarian leaders: 1) Underestimating Netanyahu’s resilience (2015-2021); 2) Miscalculating the U.S. embassy move to Jerusalem (2017-2018) that he predicted would lead to “severe repercussions and a total collapse of regional stability”; 3) “Arab Spring” optimism about democratic shifts in the Arab world and his general “tilt” theory (nations were tilting towards freedom).

Mixed are his predictions that are often correct on the trend but wrong on the timing or the intensity: 1) China’s “crisis” and economic slowdown; 2) Russia as a gradually challenging power vis-à-vis Europe through indirect means; 3) Trump and the U.S.-China trade war, correctly asserting that his trade policies would not effectively bringing jobs back to America, but wrong in predicting an immediate loss of U.S. global competitiveness.

**********
Interestingly, the AI response also explained Friedman’s mixed record as a “big picture” person focusing on global trends, but who often overlooked small, annoying local details such as culture, history, and individual choice. In short, Friedman’s a good “systems observer” but a weak “political scientist” – quite good as a futurist regarding technology, but a shaky predictor of foreign leadership behavior.

Does all this make Friedman a great, middling, or poor analyst and prognosticator? My answer: compared to whom? How well do other commentators do? More germane for you personally, how well do(es) your favorite columnist(s) perform on this score? If you found your family GP sporadically rendering the wrong diagnosis, wouldn’t you switch to another doctor? Conversely, if s/he is almost always correct, wouldn’t you recommend your physician to other family members and friends?

So go ahead and try this with your preferred columnists through your favorite AI program. In fact, just as with doctors, ask for a second opinion – from a different AI you tend to use (Gemini, Perplexity, Claude, ChatGPT, et al). If the two pretty much agree, you’ve received the real picture regarding your favorite news analyst(s). But if these two AIs disagree, then… they need to be checked as well! You should then ask other AIs to provide analyses of your favorite AI’s successes and failures. And if all this is too confusing, try something revolutionary: read the news and make up your own mind as to what it all means and where it’s heading.

As for predicting the outcome of Israel’s election, wait until Sept. 8 when all the parties have had to officially register (or combine with others). Until then, today’s “results” won’t be much better than crystal ball foretelling.