The computer looks at everyone's entire body of work. It is impossible for any system to rank every team over every team they beat. Simply because of A beat B, B beat C, C beat A scenarios. My computer gives team's bonuses for beating teams, but more or less, if you play a team ranked well ahead of you then you need to beat them soundly to convince the computer it was most likely because you were the better team and not because of statistical noise (you got the breaks in that game and weren't really the better team).
So here is how Pulaski County got to where they are now:
Pulaski County started the season with a 58.09 rating. They played Northside who was rated 80.06. Pulaski County pulled off an upset 22-20. This kind of victory (2 points) is not the kind of victory that convinces the computer that the better team clearly won this game (2 point victories usually only go to the better team something like 54% of the time). Nevertheless it did convince the computer to move Pulaski County up to 66.13 and Northside down to 72.02, figuring if they played again Northside would only be a 6 point favorite. While this might distress some people, it's fine with me. If Pulaski is really better than Northside, I would say, the ratings would sort that out soon enough.
Pulaski(66.13) then faced Franklin (34.01). The computer expected a 32 point win out of Pulaski, but Pulaski only won by 23. This told the computer that maybe it overshot with Pulaski the week before. Time to adjust...Pulaski down to 63.49.
Next up on the schedule for PCHS (63.49) was Bassett who came in at 24.64. This was an obvious running clock blowout if the computer was right. PCHS by 39 points. Pulaski only took this on by 33. Not a bad win, by any means, but not enough against such a hapless foe to keep Pulaski's rating in tact. A slight drop to 61.94 was more in order.
Next up was Cave Spring (45.18). PCHS would be nearly a 17 point favorite in this one. Bingo! They win by 18, it's week 4 the computer starts settling in on ratings that work. It was a defensive battle, the computer gives a little bonus for that and treats this win more like a 21 point win. This means Pulaski moves up a little bit. They move back to 63.09.
Finally, they faced the dreaded PH. PH had a 74.20 rating. That meant PH was an 11 point favorite. Pulaski pulls off the nice upset for the second time this year by 7 points (the defensive bonus means it gets treated more like an 8 point win). Once again, it wasn't a convincing enough win for the computer to decide all previous data on the two teams was overwhelmed so it didn't throw Pulaski past PH, but it did move Pulaski up to 67.80 and pull PH down to 69.88. So it has decided the teams are about even.
As for Northside, they have played pretty consistently where the computer has predicted them since that first game so their rating hasn't moved much from the 72.02. If you watch the ratings closely, once we get to week five or so team's ratings tend not to move too much. It's just hard to do what it takes to move them. However, you can play with the
ratings calculator and see what you'd have to do to get to a certain level.
Next up for Pulaski, for instance, is Christiansburg. Christiansburg has a 50.96 rating. Pulaski has 67.80. To make a long story short, Pulaski is favored by 17 and needs to win by at least that much to keep their rating.
If Pulaski would like to get that rating to 70 they would need to win by 27 points (a little less if it were a low scoring game). To get to 80 the work gets kind of crazy, the Cougars would need an 82 point win (you don't usually move up as much by winning one game big as you do by consistently "beating the spread").
Of course pounding a team ranked considerably higher than you is the fastest way to shoot up in the ratings. It's harder to move up against a bad team because the expectation for your team is so high. L.C. Bird can't move up against George Wythe in their next game unless the Skyhawks win by more than 104 points.
So, does a coach refusing to run up the score end up hurting his rating? Yes, a little bit, but if a team finds themselves underrated, the next time they play they should make that up because the bar will be set too low for them.
The short answer to your question is this....The computer is saying that if Pulaski were to play Northside or PH again it still thinks Pulaski is more likely to lose either game than win. However, the margins in either case are very small and the computer would not make those calls with a high degree of confidence at all.
Please be aware that the algorithm I use is designed to make predictions about future games, not to reward teams for past games. Margin of victory and strength of opponents are very important. I have algorithms that I would use to determine who deserves better seeds in the playoffs and things like that and in those algorithms I'm sure Pulaski is far ahead of either Northside or PH.
I hope you'll keep an eye on the ratings to see if things end up about where you think they should be (they're never going to satisfy everyone). Just bear in mind this fairly simple rule. Outperform the spread and you'll move up, underperform and you'll move down (There are some tiny exceptions to this--defensive battles end up adding a bit to the margin of victory before making the adjustments, and if you're favored by more than 40, then the computer will let you hold your ranking if you win by enough, so that a team like Bird doesn't have to win by 102 to keep the computer satisfied -- just by 71).
So what you get from the computer is a compromise. Pulaski tends to play better against good teams and worse against bad teams (against the spread anyway), so when the computer is trying to average out how Pulaski does overall when it assigned the team one number, that number probably under represents how they play against the powers and over represents how they play agains the poor teams. It'll be interesting to watch them the rest of the year and see if this is real pattern or just a bit of an early season statistical quirk.
P.S. I hope you win them all and that my computer has them way under ranked right now. It probably seems weird to most people, but the season is a lot more fun when the computer is wrong, and I don't really have any pride about being right or wrong on a particular game. I pretty much know what percent I'll get right in each week and what my standard amount I'll miss a prediction by will be. Some years there's just a team or two that always seems to make me look bad. Right now Pulaski County is one of those schools.
