"Can I trust [insert not especially good player]?"
The answer is always no. There are situations in which you cannot trust Mike Trout or Chris Sale. Granted, those are relatively uncommon. When the player in question is Daniel Palka or Felix Pena, trust should not be part of the equation.
In many ways, this is a failure of language. I understand the intent of the question. Something to the effect of: "Can I use this guy without worrying?" And the answer is always no. That is the nature of baseball. Things go wrong in spectacularly unpredictable ways.
Here's an example. Kyle Gibson had a strong season. In mid-August, he tossed a pair of seven inning gems against the Tigers. Then he visited the White Sox. They had a worse offense than Detroit. They were the most strikeout prone lineup. And Gibson coughed up seven runs in 4.2 innings.
In May, after a four start stretch featuring 45 strikeouts and seven runs in 28.2 innings, Chris Sale was hammered by the Braves for six runs in 4.1 innings (he still managed eight strikeouts).
Players aren't things to be trusted. All you can do is manage them such that they're in the best possible positions to succeed. In roto and points-based formats, it's easier to take the long view. Every player comes with lumps, but they usually even out over the course of 162 games. Sale had a couple rocky starts and missed time late in the season. He also posted a 2.11 ERA and 237 strikeouts in 158 innings. His owners got what they paid for.
Head-to-head formats offer an additional challenge. This is where you might find yourself benching a Sale or Trout late in the week to protect rate stats.
The moral of the story is simple. You're wasting time fretting about "trust." Trust nobody. Use them anyway - because you have to. Profit.