Sidestepping risky prediction area

Tuesday, October 16, 2018

Some of my sports journalism colleagues, including ones at the Hendricks County Flyer and the Indianapolis Star, choose to make and publish weekly predictions on the winners of high school football games.

I’m sure part of the motivation is to build interest (and readership) with their sometimes-controversial selections. I will universally pass on that opportunity. Not that I can’t take criticism, when it’s warranted, but such a practice just begs for confrontation in my eyes.

I will admit that before leaving the office on game nights I will often predict the winners in conversations with my office colleagues. The first couple of weeks always don’t go really well, but after that it’s generally not too difficult (especially in football) to come up with the correct winners.

At least my newspaper colleagues and I have actually seen the teams whose games we predict, whether publicly or privately.

In the case of John Harrell and Jeff Sagarin, however, they merely let the numbers do their picking through their various computer algorithms – and those picks have become a pretty hot topic of conversation among at least one Putnam County team.

Last year, I started including Harrell’s predictions for each game in the capsule preview for football and basketball games that we run in the sports section along with a great deal of other information taken from Harrell’s site (www.johnharrell.net).

What a deal. There are predictions, but not by me, and anyone disagreeing can get in touch with Mr. Harrell.

Each year, Harrell publishes predictions on sectional winners based upon a complex formula including ratings, tournament draw and home-field advantage. He will publish his percentages after each level of the tournament, and generally does pretty well with his picks.

One team that has definitely skewed his success rate this year has been Greencastle football.

The Tiger Cubs have compiled an 8-1 record this year, and Harrell’s numbers have predicted them to lose four games this year that they eventually won. Those numbers are far less than Harrell’s normal success rate.

So why the discrepancy? From talking to Harrell extensively a couple of years ago for a column that ran last year, I learned that two of the biggest factors involved in computing the predictions are strength of schedule and winning margin.

In these two areas, Greencastle definitely does not provide glowing results. Prior to last week’s game against Edgewood, the Tiger Cubs’ schedule was ranked 216th out of the 320 teams that play high school football. Following its convincing 34-6 win over the Mustangs (who were 6-2 at the time), the Cubs’ schedule ranking improved minimally up to 190th.

My dad would always say that “you only have to win by one” run or point, and it didn’t matter if you won by more than that minimum amount or not. In the world of computer projections, however, how much you win by appears to matter a lot. At least in some cases.

The Class 3A Cubs started the season with a nine-point win over a Crawfordsville team that lost its first eight games of the season before winning last week, then won by 10 points over a Class 2A Cloverdale team that has only won two games. In week four, the Cubs won by 12 points over a Class A South Putnam which has won three games.

In the real world, a win is a win is a win, but the computers aren’t smart enough (yet) to factor in other elements that also lead to outcomes. Did the computer know that the Greencastle-South Putnam game was played in a torrential downpour, like most games that week, making scoring incredibly difficult? Probably not.

The only Tiger Cub loss has come to Northview, a team ranked in the top 10 in the state in Class 4A. That loss didn’t hurt Greencastle at all in the two high school ranking systems operated by human input.

The week after losing 42-9 to Northview, Greencastle’s spots at the top of the “also getting votes” category were basically unchanged – as they should have been.

The Cubs beat three teams who have recorded six wins (Edgewood, North Putnam and Sullivan), yet still are not getting a lot of love from the human element.

When the pairings were drawn, the Indianapolis Star ran an article briefly detailing the matchups in the sectionals involving teams from its coverage area.

While at the time Greencastle was 7-1, the article indicated that “Monrovia (4-3) and Sullivan (6-2), probably the top teams in the sectional, would meet in the semifinal.” No mention of the team with the best record in the field, which had won at Sullivan a few weeks before.

The Greencastle football parents club has had fun with the wrong predictions on social media, and several players and other students have also referred to the errors in a lighthearted fashion.

I’m positive that coach Mike Meyer could not care less about whether his team is picked to win a game or not, and he looks forward to each Friday night and the opportunity to put his team on the field to see how well it can play. If such predictions serve as “bulletin board material” and inspire his players to perform even better, so be it.

If I were the Cubs, I would hope Harrell’s predictions would go against them every week considering how many times they have risen above expectations.

Harrell’s first-round sectional pick for Friday night? Danville 28, Greencastle 24.

Harrell also rates the percentage for each team’s chances in each sectional, and if the Cubs draw inspiration from single-game picks then they should really love this data.

Monrovia, with a 4-4 record, is given a 32.6 percent chance to win Sectional 30. Sullivan (7-2) is second at 25.84 percent, followed by Tri-West (5-4) at 20.69 percent, Danville (5-4) at 10.44 percent and Greencastle (8-1) at 9.51 percent.

Monrovia has played clearly the best schedule among the sectional field, and since that aspect is so important it’s not surprising the Bulldogs are the choice.

They have lost to Triton Central (8-1), Beech Grove (6-3), Indianapolis Lutheran (6-3) and Indianapolis Scecina (8-1).

Such predictions are harmless, intended for fun and to promote fan interest. No one could possibly watch all 320 teams play and make such judgments on visual evidence, and even following the teams closely enough to know when injuries or bad weather happen that could affect outcomes would be nearly impossible.

So, computer metrics are the best we are going to do to try to make such projections.

However, finding teams whose ratings and predictions oppose the logic of the criteria mentioned above doesn’t take long.

Let’s look at South Putnam’s sectional.

The Eagles are the only Class A team in the state besides Lafayette Central Catholic who don’t play another team in their enrollment classification during the nine-week regular season. Being in a conference division with Northview, Greencastle, Sullivan and North Putnam (teams who are a combined 30-6) should help South Putnam in the prediction area, and it does.

The Eagles are the second-highest rated team to win Sectional 45 with a 12.54 percent chance. At the top? Undefeated Covenant Christian, a team which averages 34.9 points per game while allowing 5.6 points each week.

Based upon those numbers, the Warriors seem a logical pick to win the sectional (70.32 percent).

But, then there’s that selectively-important schedule thing.

Covenant Christian’s schedule is horrific, to be as kind as possible. It is ranked 302nd out of 320 teams, and for good reason. South Putnam’s is 187th, undoubtedly hampered by two games against winless Owen Valley – ranked by Sagarin as the eighth-worst team in the state.

The Warriors have played more than half of their games against fledgling, small private schools such as Oldenburg Academy, Heritage Christian, Traders Point Christian and Bowman Academy. Their opener was against Jasonville Community, a non-IHSAA team that is basically a club team.

None of this conjecture actually matters when the first kickoffs take place at 7 p.m. on Friday. Teams still need to go out and avoid making mistakes, control the line of scrimmage, win the turnover battle and avoid injuries.

The best team will hopefully win each game, whichever team that is.

But it will be fun to look back in a few weeks and see how accurate (or not) the predictions were.