The complicated world of Strength Index and OSI: Why it’s important scores are reported correctly

For years, every team had a power points guru.

No, not someone who could do a slideshow presentation. Power points were – and still are – how the NJSIAA seeds the playoffs in all sports.

But in 2018, after stakeholders from all over the state worked to refine the playoff qualification system, the state added a second metric for football: the Born Power Index. Every team had a rating based on how well they did score-wise – not necessarily win or lose – when compared with the rankings of the two teams.

If one school was rated 14 points higher than another, they would be expected to win that game by a couple of touchdowns. Win by more, you’re considered better and your stock goes up. Don’t meet that expectation? Your rating goes down.

It was good, and probably is the best true determinant of strength of a team, without looking at the players and coaches and game plans themselves. But it rewarded teams for winning by larger margins, and was scrapped after a year.

It’s replacement has stuck around since 2019 – the Strength Index – a similar formula, but now the second metric for playoff seeding is OSI, the Opponent Strength Index. Essentially, the Strength Index values of your opponents are averaged, with teams getting the full value for an opponent they beat, half the value for a loss and 3/4 value for a tie.

But what happens when an incorrect score is reported to Gridiron New Jersey – which makes all the official calculations for the NJSIAA – as happens with surprising, and alarming frequency? Anecdotally, there’s at least one a week, sometimes more.

We saw that example two weeks ago, when Gridiron’s Jon Fass tweeted on October 8th: “In addition to accounting for @GoTigersTHS victory today, UPR standings have been updated to include a couple of scoring corrections as well.”

There have been times I’ve seen three different scores at three different media outlets.

So, one might assume things would be better this week. But no. Again, there was a score correction, but this one back on September ninth.

Gridiron didn’t mention the game, but “several teams” were affected. By our count, there were almost two dozen teams that changed by more than 0.01 in Strength Index.

(We use that threshold because there are often differences in how numbers display, or are truncated or rounded when published. For example, a 60.146 would round to 60.15 but be truncated at 60.14, and it’s unclear which Gridiron uses in the many steps of the mathematical process.)

How are so many teams affected?

Well, let’s say Team A plays Team B on September 9th. If those two SI values are wrong, then when Team A plays Team C on September 16th, and Team B plays Team D that same weekend, all four teams’ numbers will be wrong. It’s the NJSIAA math version of “going viral.”

And the mistake can spread to other conferences two if there’s a non-conference game in there, much like, oh, say, a global pandemic.

The game was in South Jersey, and affected the South Group 2 and South Group 2 supersection standings, mostly. Any other effects were minimal and didn’t affect standings.

Here’s a sampling of the more notable changes:

• Allentown went from a 70.69 down to a 70.43.
• Audubon dropped from a 52.18 to a 52.15.
• Glassboro went from 71.15 to 71.06.
• Haddon Heights went from a 721.31 to 72.21.
• Haddonfield rose from 88.87 to 88.99.
• Paulsboro went from 52.89 to 52.87.
• Penns Grove dropped from 64.96 to 64.60.
• Pennsville fell from 50.74 to 50.64.
• Salem went from a 62.92 to a 62.83.
• Sterling went up from 27.75 to 27.79.
• Woodbury dropped from a 78.83 to a 79.31.
• Woodstown went from a 74.75 to a 74.84.

St. Joseph-Hammonton also moved from a 40.23 to 39.83, in a change that could affect Carteret, which beat the Wildcats 42-12 in Week 1. They now will get 39.83 factored into their average when the playoffs are seeded, instead of 40.23. (SI values are locked this weekend. They’re still tracked throughout the rest of the year, but values won’t change for playoff calculation purposes after this weekend’s games, which makes the process much simpler for all involved.)

Carteret is still in the same spot in the standings, but there’s no guarantee that wouldn’t affect things this weekend.

And Lyndhurst, which may have been affected after playing Cresskill in Week 3, went from a 60.95 to a 60.88. That, of course, affected Park Ridge, which beat Lyndhurst by 19 this weekend, but also had its SI adjusted from 77.58 to 77.39.

All of this could affect countless other schools who may have played them even before the error, since SI values change with every game.

Yes, it’s a wacky world when it comes to Strength Index and OSI.

All the more reason coaches, assistants, managers – whoever reports scores to Gridiron New Jersey, or any other media outlet, for that matter – get them right the first time.