On such random patterns that got you in at the first time you can generally only make good on the loss by increasing your bet size, which often leads to a martingale approach. In the end that is why so many traders fail, they pick up pennies in front of the steam roller and every now and then easily gamble away 10% or so of their account because it only takes this one out of x times to not reverse.
I don't normally do this, but I'll bite here on a trade I made this morning.
First circle is the long signal based on a pattern, Event A. Expected movement did not occur, held into drawdown. Because A did not occur then, probability of B increased. What is B? The subsequent 2 circles. When A fails, it allows one to get out at breakeven; this is what Jesse Livermore says when you could have gotten out at breakeven, known as pattern B. We all have own names for them. The market broke above and allowed reentry on the final, tiny circle. However, the probability had been reduced at that point due to correlations among other vehicles. This is how one can trade with a high probability of a win or breakeven. The TSLA trade also did this, yet we can see it worked out AH because the trade occurred later; I expect a gap down on TSLA tomorrow or a red day. If it gaps above, it should test the aforementioned 668.43 level as an event type.
