I agree that the general answer is statistics but is also a definition problem. Some people certainly hit improbable if not seemingly impossible outcomes both positive and negative but that still doesn’t tell you everything.
For example, you have the story of Violet Jessop who survived the sinking of the Titantic as well as its sister ship HMHS Britannic. There is more. She also was on board the RMS Olympic when it collided with another ship. Those were the three mostly identical ships of the same fleet and she was on board all of them when they were destroyed or threatened. It is quite unusual but is that luck? One could argue that any person that was never aboard any of the three was luckier than she ever was. The same is true for Tsutomu Yamaguchi who happened to be on site and survived both the Hiroshima and Nagasaki atom bomb detonations over Japan. Is that good luck or really bad luck?
Other cases are more confusing. There have been several people that have won large lotteries twice. It is surprisingly inevitable when you look at the statistics behind it. However, none can beat Joan Ginther who has won the Texas lottery an astounding 4 times for $21 million so far. We don’t know if that is true luck or not. She is also a semi-hermit and also a college math professor. They know she isn’t truly cheating but she may have invented some lottery hack just like Marjorie and Gerald Selbee did for the Massachusetts Win-Fall scratch ticket game.
Luck can go the other way as well. There are people that lose all of their kids and their house within days of each other in separate accidents. You can use statistics to calculate the individual probabilities of different events. However, the universe of possible outcomes is infinite so it makes the overall improbability of real-world outcomes seem much higher than they are. In other words, there are an infinite number of events that are in the 1 in a septillion range or even higher but there are so many of those that a few of them are inevitable. Some of those will be very good and some will be very bad. Random clustering of major events that affect a single person is the best definition that I can think of for ‘good luck’ versus ‘bad luck’ and there is no way around that. Outcomes that are the result individual choices and decisions (like drug use or lack of work ethic) don’t fall into the luck category at all.