Wins tell you how the starting pitcher from team A pitched compared to the starting pitcher from team B during a specific game. That is why the stat was created, although when it was created, pitchers routinely pitched complete games, so it was much more unlikely that the bullpen could screw it up. With the exception of a couple outliers on both ends of the spectrum, I believe that most MLB offenses are within variance of each other. If the conditions on game day are rainy, cold, etc. then both SPs will likely do well and have low ERA's for that day. But if pitcher A gets the win, then it tells you that he pitched better than pitcher B who got the loss. If it is hot and the ball is jumping, both pitchers will likely give up runs and have worse stats, but the pitcher who does better will get the W.
In my opinion, a pitcher's W-L % was a more useful stat in the old days when starters regularly went the distance. Now it is so much more likely that the bullpen will blow it and give the pitcher a ND or even a loss. Over a long career, a pitcher with a good W-L% is generally regarded as a good pitcher, and a pitcher with a bad W-L% is regarded as a bad pitcher, regardless of the team(s) they played for. It is a way to separate good pitching careers from bad pitching careers, but on a season-season basis it's not very helpful.
What is Mind? -- Doesn't Matter!
What is Matter? -- Never Mind!