Binary probabilistic questions are questions that only have two possible answers: Yes or No. You answer by giving a 0–100% probability of an event happening. A 0–49% probability is interpreted as a negative answer with various degrees of confidence, while a 51–100% probability is interpreted as an affirmative answer with various degrees of confidence.
A conditional binary question is a question that asks “If event A happens, will event B also happen?”
For example: “If Neo trades above $80 by 20 December, will NEO also trade above $120 by 10 January?“
Neo would have to reach $80 by December 20, or else all predictions would get a 0 score. Therefore, you have to predict the likelihood of the second event happening (Neo above $120) and it will only be counted if the first event happens as well.
Tips from our forecaster Jonathan Martin:
This type of question most generally asks you to identify a micro or macro trend in a given timeframe. Binary questions cover a period of approximately 3 days — 1 week, which is usually enough time for a trend to change.
The scoring for the binary questions is simple, scaled to the % given as input, with a maximum of +/- 50, by subtracting from the percentage given, IE “Will Bitcoin’s price rise to $6,500 by 30th July?” If the correct answer is “no” (= 0%), and the forecaster input a 65%, they will lose 15 points (50–65). A forecaster who answers 20% to this question will get 30 points (50–20) for the right answer.
My current process for the binary questions goes like this: read the question and find the price level it asks about, noting carefully the date which the binary closes, and whether 50 + is affirmative or negative in the context of the question. Then, I go look at the 1D charts provided, and trace back the same number of days as I will be predicting. Past behaviour does not necessarily predict future, but you can be reasonably sure that most of the time it will be along the same lines, or at least complete the current trend. If I need further clarity, I will consult the 4H, 30M, or 1H charts as well.
I identify what I most logically think will happen, and it amounts to one of five options:
- a steady rise
- a steady fall
- a rise with subsequent falling retracement
- a fall with subsequent retracement upwards, or a complete break in trend, which what would be referred to as a Black Swan Event. If I get an inkling of such a massive reversal coming, I wait for the market to settle down, and watch for some clarity to emerge from the patterns.