In addition to this being a Vegas line rather than a prediction, it's also worth noting that few people try to project/predict outliers, at least not in a systematic way (i.e. projecting win totals for all 30 teams, rather than an isolated prediction like "the Padres will lose 105 games this year"). Very, very few teams will ever get projected or predicted to lose 103 games, but that doesn't mean projections/predictions are inherently bad/incorrect/worthless. I'd probably judge them more on general contention/non-contention buckets -- i.e. if the Twins are projected to finish with the 3rd worst record in the league, well out of contention, and they actually finish with the worst record in the league, well out of contention, I don't think that's a failure of the projection system, even if their projected win total was, say, 70 and their actual win total was 59.