Not really. Two cases. If one makes a prediction via some fair random means (flipping coins, tossing dice, etc) then there are consistently the same chances of being right vs wrong at every instance. And if one makes a prediction by actually studying relevant trends (like weather forecasting, etc), then over time it becomes increasingly easier to make right predictions.

The only route to being consistently wrong is through bias, whether via unfair random means or being deliberately blind to the trends. And those can't be called predictions, by definition.