If you buy BTCGOOG today and sell next year for a 0.00000001% more you have certainly made a gain - the risk has ended with your divestment and turns out to have been 0% (in that case).
No - just because something there was a risk of didn't happen does NOT equate to the risk having been zero.
If you bet on a coin-flip being tails, I flip the coin and it comes up tails that does NOT mean that the risk of heads coming was 0%. Same for any other risk - it failing to occur doesn't mean it never existed.
Similarly consider if you invest in 1000 absolutely identical (as far as all available information indicates) securities believing it likely there is a 10% risk of failure for each of them then 200 fail. That does NOT mean there was a 0% risk on 800 of them and a 100% risk on 200. It tends to suggest there was a 20% risk of failure on ALL of them.
Now it COULD be argued that everything is predetermined and so the ones that failed were always going to fail etc. But whilst an interesting philosophical point it's entirely useless for practical purposes. We have to deal with risk that we can quantify/measure/estimate in advance - and collapsing that to either 0 or 1 based on results is only useful if there's some measurable (in advance) factor that can be used in future such assessments. If the real risk is only definable at all in hind-sight then we have to group together securities/whatever with a similar visible profile anyway.
