In such small quantities it would not be a very good predictor of what will happen regardless of what stat is being discussed. It's not necessarily a bad descriptor of what has happened, especially if we can support the conclusions with other statistics.
Given that it seems to be quite random that which takeaways, giveaways, hits and shot blocks get recorded and which do not, unfortunately we
can't use those stats even to show
what has happened.
For example Laine had like 4 hits on his first game, but only 2 were recorded. We cannot just what the stats and say "Laine had 2 hits on his first game", as it's not true. Only 2 hits got recorded.
Most of the time I guess you can use those stat as the minimum: Laine had
at least 2 hits on his first game and has had
at least 4 takeaways and 1 giveaway on his 3 first games, but just by looking at the stats we can't say if the real numbers should be 4 takeaways and 6 giveaways or 10 takeaways and 1 giveaway.
Due to the randomness, large sample sample size is must. Given that there is no systematic bias, no conspiracy etc., in large enough quantities those stats should be usable for comparing guys within one team, like random sample used in statistics. Not as absolute numbers, like "Laine has 100 giveaways", but as relative measure: "Laine has more giveaways than takeaways, and similar amount of takeaways than Wheeler".