It is not automatically a fallacy. It is not something that can have a hard-and-fast answer. The estimate of the chance that the errors happened in the way they were observed is complicated in the extreme and subject to many errors, biases, and filters.
You would need to examine the types of errors and their expected range, the possible effects of errors in the expected range, survivor bias, and a bunch of other things I'm forgetting just now.
Just as an example, consider my comment on @Mark Andrews's answer. Perps that make very few mistakes tend not to get caught. So those who get caught will tend to have made lots of mistakes that increase their chance of getting caught.
A lovely example of survivor bias comes from data about damage to aircraft in WW2. But it is a long side-bar. Please read at the link if you are interested.
So data about errors needs to be extremely carefully examined. Maybe there is some "innocent" reason for the errors to be biased. Innocent in that the person is not doing them deliberately.
One trivial example: Consider a guy who has a lot of car crashes. When driving, any error steering is going to have a high chance of creating a crash. You would need to examine these quite carefully to conclude they were deliberate.
So, if the person had some neurological disorder that causes improper hand-eye coordination, the errors might not be deliberate. The subject might be unable to do any better. (Though, if he knows he has the condition, we might still hold him culpable for continuing to drive.)
If, however, he was constantly racing, driving much too fast, driving on the wrong side of the road, passing on the shoulder, following too close, etc., then these sorts of errors would be hard to interpret any other way than deliberate. That is, you must characterize the type of error and have some information about the origin of the error.
There are some ways that the data can be characterized fairly convincingly. Consider that, at some future point, the POTUS is from the Stupid Party. And the media is constantly making mistakes that make him look bad. Dozens of mistakes, day after day, of all different types. Mis quoting, mis attributing, getting dates wrong, reporting false facts, etc. etc. Then there is an election and the Stupid Party is out, and the the Evil Party is elected for their turn. Now the same media continues to make about the same number of mistakes per week. But this time, they are reporting things in a way that makes the POTUS look good.
Because it is the same exact people making similar mistakes but with the opposite polarity, it is difficult to believe that they could be doing it accidentally. When they report "the chocolate ration fell" during the Stupid Party administration (and it actually stayed the same), but report "the chocolate ration rose" during the Evil Party admin (and it still stayed the same) it is difficult not to turn to malice for an explanation. The only factor that changed was which party was in power. It was similar data in a similar situation about a similar subject. Bias seems inescapable.
There is one more caveat to keep on your dashboard. There is probability involved here. There is no hard threshold past which you conclude it is deliberate. You still cannot mind read. You must examine the data on the basis of the probability you conclude from the data and stay as open minded as the probability justifies. If it's 5 coin tosses that all came up "heads" it is still only 1 in 32. You begin to be mighty sus. But it is not logical proof. Not good enough for criminal court, but likely good enough for civil law suit.