You can only really do it consistently with the opposition and only in league play, though. According to the rules, you may only choose a player once in a weekend. So when you're playing the same team 2-3 times in a week, the assumed starting goaltender may only be chosen once. On top of that, as you are not allowed to change opposition mid-week, a team that splits shifts is a coin flip in and of itself, although some schools have a pattern associated with this (with CCT, it's play-'til-you-lose). With that aside, assuming the teams are on equal footing, the chances of a team winning is about 43% (as there is a 14% chance of a tie; this came from some research done a couple years ago, I think TBRW). We'll say it's about 45% chance of you scoring points, as a certain percentage of those ties will be 0-0, which counts under shutout rules. Adjust as you will for the aforementioned anomalies.
Now consider the selection of a skater. First, you must factor in eligibility. Then, you must factor the amount of ice time the player gets, and of course the goaltender anomalies still apply. When a goal is scored, even when you factor in that your line/pair is on the ice, you have essentially a 50/50 chance of scoring any points at all (adjusting for the number of assists awarded on goals, especially considering a number of goals in this league come off turnovers as opposed to sustained pressure, and point weighting). Do some lines score more than others? Yes, and you may factor that in accordingly, but being on the ice isn't enough. Obviously this is offset by point multiples, but given the number of goals scored in a typical game, it doesn't make up for the chance of a skater getting a point to lower the house edge.
Therefore, when you consider from a standpoint of house advantage, you will get more points in the long run from consistently choosing goaltenders, and doing so on both teams. Is this a guarantee that you will win PtC? Certainly not, as choosing skaters is improbable, but not without possibility. At this point in the season, choosing nothing but goaltenders as allowed, and maximizing for point totals, would net you a minimum of 45 points (assuming none of the goalies had assists) before week wins and other skaters chosen. Have many players scored more than this? Yes, but have done so in the short term.
