I like that PECOTA plays the sim through the schedule and goes with those results. You can see that the run differential isn't necessarily correlated to winning percentage. On the other hand, FanGraphs seems to take the run differential and apply the win loss record based on that. I didn't calculate any team just to check, but it looks like the forecasted record might be pythagorean.
The noticeable result is that FanGraphs seems to have a tighter range of win loss records. All teams are within 64 and 97 wins. PECOTA has a couple of teams above 100 wins and a couple of teams below 60 wins.
Its just kind of fun to look at. We can poke holes at the results and how they were arrived at. Playing time is a good variable to criticize. There are still a few free agents to be factored in, teams have spring training to get through, there will be injuries, trades, disappointments, surprises, etc, before teams break camp. And even then, these forecasting systems can have better inputs to use, but there will be more injuries, trades, disappointments, surprises, etc, throughout the season to skew the actuals from the projections.