Positive results on a quasi-scientific self-experiment is not a good way to gain knowledge. At best it is how you specifically react to the variables introduced; however, it seems that you only performed this self-experiment on two separate occasions or perhaps only when you planned to go to the gym. Although your attempt to limit confounding variables is commendable, it is not enough to reach a conclusion. To definitely find the truth on the matter, many more participants must be enrolled and the experiment should last for at least three weeks, but hopefully much longer so as to definitely reduce any confounding variables that will influence the results. Some of the better studies regarding physical performance were running for around six months just for this reason.
Specifically regarding your self-experiment, if it was indeed performed as one session vs. one other session, there could be any number of confounding variables that could have effected your performance other then masturbation. Did you keep a sleep journal during the experiment (recording night mood, morning mood, amount of sleep)? Did you consume alcoholic beverages in the seven days leading up to the session? What was your specific workout routine?
In my opinion you should continue this experiment, but record more data for longer amounts of time; say six months of abstention, and six months of regular masturbation. Of course, this produces it's own confounding variables because it could very well be that the result you get has nothing to do with whether or not you masturbated, but rather the gains you made during the first trial period. Which is why to be able to tell people what is truth and what is incorrect, more participants must be recruited to validate the results; otherwise, it will only ever be (again) at best (i.e. confounding variables properly accounted for to the best of your ability), how you specifically react to abstention.