We’ve written before about the precarious nature of forecasts, and an article in last week’s Wall Street Journal sings the same tune—that investors shouldn’t get “carried away” with predictions that the market is headed for one of the best years ever.

At the beginning of this year, it says, the average prediction was that the S&P 500 would “end at 2216, a figure hit—and passed—for the first time on Wednesday.” It argues, however, that “this year’s accuracy was a matter of luck, not a sudden improvement in the forecasting power of brokers.”

Among the naysayers is David Kostin, chief U.S. equity strategist at Goldman Sachs, who suggests there will be “overexcitement about Trump, then reality” adding that high hopes for the new president could push the S&P to 2400 by the end of March but that gains will be tempered once Congress enters the mix. The chart below outlines how predictions have generally stacked up as well as some pending market forecasts:

“The average strategist,” says WSJ, “hasn’t started a year predicting a stock-market drop in any of the surveys carried out by Bloomberg since 2000, while shares fell one year in three.” It quotes Jan Loeys, chief market strategist at J.P. Morgan, who says, “We do forecasts because our client base asks for forecasts, not because we have any great advantage in forecasting levels.”