Great work Durvasa, very interesting analysis. I think the team has nice momentum right now. Roles and rotation are more defined.
Great analyse and we are going on a pumping road from next year onwards... Deserve being repped Good luck, Rockets
I KNOW WE ALL REMEMBER 2 YEARS AGO WHEN BOSTON SNAPPED OUR 22 GAME WINNING STREAK. I WAS THERE FOR MY BIRTHDAY AND WASBUMMED NOW THE ROCKETS WILL HAVE THE CHANCE TO SNAP THEIRS (IF THE MAKE IT 5 MONRE GAMES W/O LOSING) WHICH WOULD MAKE UP FOR IT! if u ask me thats pretty ironic if we snap there streak when they snapped ours. :grin:
So the mechanism you use to measure the SOS is as follows: You use the opponents winning percentage at the time they played but with a memory of five games (moving average of 5). Hollinger computes his as the average winning percentage of the opposing team at the time the game was played. Every game gets compounded into the equation. Consequently, trends become important as it indicates softening and difficulty of the schedule. I use the traditional equation, which is much more difficult to compute, which is: 2/3*(Opponents' Average Winning Percentage at the time they played)+ 1/3*(Opponents' Opponents' Average Winning Percentage at the time they played). To me this equation is more robust than Hollinger's method. I really like your concept (although I can not really understand your justification for the moving average of 5 -I guess it would help if you explain why not three or eight?) but I don't understand how you compute the strength of the schedule from today moving forward. Do you assume that the rankings remain the same through out the rest of the season for this exercise or did you compute the standard deviation and simulate the rest of the season with a Gaussian (Normal) distribution?
What I did is much simpler. I am assuming that the strength of the last 5 opponents at every point in the season is, more or less, captured by their average current record (actually, point differential). I figure that after 2 months, the point-differential statistics is a fair gauge of how good teams are. It could be interesting to try another method -- like simulation -- and see if the "SOS curve" looks much different.
I though we may want to revisit this, since it appears the Rockets might be in position to make a run at the playoffs. The SOS above was projected based on statistics back in December. Still, I think it more or less reflects what we have to deal with down the stretch. You can see that we are currently in a softer portion of our schedule, and our recent surge pretty much coincides with the schedule easing up since early February. Late March/Early April things can get tough again. We have a home stand coming up (against some pretty decent teams). We need to take advantage while we're at home, or we're toast.
Strength of Schedule Would it be too much to ask if you could implement the recent records of all teams to see the difference on how the graph would look now?
