Twitter is what you make it. I’ve filled my Twitter experience with people like James Hilts, and it’s made me a better teacher, coach, and person. This is amazing 👇

No idea who made this game, but it's extremely fun and addicting. Thank me later. 82-0.com

A master of his craft, @JinkiTomita11 left his mark.

A few clips from my last high school baseball game. Clip 1: my favorite 3 pitch sequence; split, curve, 4-seam Clip 2: backdoor two-seam Clip 3: 4 fastballs + slider

Your 2026 @PILathletics Pitcher of the Year, @JinkiTomita11 55.1 IP, 73 K, 11 BB, 1.52 ERA, 0.92 WHIP 3 out of the last 4 POTY are Generals ⭐️

Here’s where Generals finished compared to all other high schools in the city:

This guys pretty good. Took down a team coming in with the highest BA in the league at .314 and did it in dominant fashion. @JinkiTomita11

That's because RPI is a flawed ranking method! Full research findings brought to the NAIA coming soon... Here is my snippet on RPI, lacking some charts and graphics due to X-posting rules. Primary Criterion 1: The RPI Of the official metrics, the Ratings Percentage Index is the most familiar and the one with the most damaging mathematical structure. Here is the formula: RPI = WP × 0.25 + OWP × 0.50 + OOWP × 0.25 WP is the team’s winning percentage. OWP is the average winning percentage of the team’s opponents. OOWP is the average winning percentage of the team’s opponents’ opponents. Strength of Schedule is reported separately as OWP × 0.67 + OOWP × 0.33. Flaw 1: 75 Percent of the Rating Is About Other People A quarter of a team’s RPI captures what that team did. The remaining 75 percent captures what the teams it played did, and what the teams those teams played did. The intent was to encode the strength of schedule. What actually happens is that any structural distortion in opponent winning percentages gets a three-to-one amplifier on the final number. As the Mechanism section will show, the amplifier hits the West region directly. Flaw 2: Two Degrees of Separation Are Insufficient RPI’s formula stops at OOWP, two steps removed from the team being rated. With 188 NAIA baseball programs, five regions, and 18 conferences, two degrees is mathematically insufficient to connect the league’s full schedule network. Boyd Nation, who has maintained the Iterative Strength Rating since the 1990s, summarized the problem in his published FAQ: “in sections of the country with fewer Division I baseball schools, such as the West, the pool of available opponents tends to be smaller, which tends to pull winning percentages towards .500. As a result of the RPI only considering two levels of interconnectedness, teams from these regions tend to be underranked by the RPI’s.” Iterative systems like ISR and Massey solve this by computing every team’s rating at once and propagating connections through the entire schedule network until the values stabilize. RPI does not iterate. Whatever distortion exists at two steps stays at two steps. Flaw 3: No Margin of Victory A 1-0 extra-inning win and a 14-2 blowout count are identical in RPI. Every published statistical study of rating systems has reached the same conclusion: systems incorporating margin of victory predict future outcomes better than systems that do not. The NCAA Division I men’s basketball committee retired RPI in 2018 in favor of the NCAA Evaluation Tool, which incorporates scoring margin, offensive and defensive efficiency, and quality of wins. The NCAA’s published rationale stated explicitly that NET considers “game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and quality of wins and losses.” Four of those six inputs are absent from RPI. The NAIA baseball committee has not made this change. NCAA Division I baseball, after extensive discussion at the 2024 ABCA Convention in Dallas, retained the RPI formula but added the Kevin Pauga Index as a supplementary, non-selection tool plus a quadrant view. NAIA baseball has done none of this. Flaw 4: Miscalibrated Home and Away Weighting The post-2013 RPI weights road wins at 1.3, home wins at 0.7, road losses at 0.7, and home losses at 1.3. Those weights assume home teams win approximately 70 percent of the time. That figure is roughly correct in basketball. It is not the home win rate in college baseball. A 2021 study from the Samford University Center for Sports Analytics, which examined more than 4,000 in-conference college baseball games from 2015 through 2019, found that the actual in-conference home winning percentage runs between 55 and 60 percent. The same study concluded: “we would argue that the RPI model for college baseball should be adjusted to give less weight for a road win, possibly 1.2 / 0.8 or 1.1 / 0.9.” The implication is direct: RPI overvalues road wins relative to actual home-field advantage in college baseball. The miscalibration is likely larger in NAIA, where travel distances and facility differences depart further still from the basketball assumption. Flaw 5: Perverse Incentives at the Margin Because 75 percent of RPI is opponent records, beating a weak team can lower a team’s RPI. The win adds to the team’s WP, but it also adds a loss to that opponent’s record, which lowers the opponent’s winning percentage and pulls down the team’s OWP. The same arithmetic works in reverse: losing to a strong team can raise a team’s RPI by raising the opponent’s record. Rational scheduling strategy, therefore, is not “win as many games as possible” but “play opponents who are good enough to elevate your OWP but beatable enough that you can still win.” A well-designed ranking system should make schedule optimization impossible. RPI is so optimizable that scheduling has become part of how programs build resumes. Flaw 6: The Documented Quantitative Bias To test the RPI’s regional behavior, we computed an RPI bias score for each team in the top-100 RPI cohort. The bias score is RPI rank minus the average rank assigned by four independent rating systems: Boyd’s World ISR, College Baseball Report, Pear Ratings, and Massey Ratings. Positive values mean RPI ranks the team worse than the predictive consensus does. Negative values mean RPI ranks the team better. My study shows West teams have an RPI bias of +23.2 spots, compared to +1.3 in the North, East, and Midwest, and -1.3 in the South. Statistical tests on these results: Mann-Whitney U test, West versus non-West: U = 263, p = 0.0002. Under the null hypothesis of no regional difference, the probability of observing a gap this large or larger is approximately 1 in 5,000. Mann-Whitney U test, West versus South only: U = 151, p = 0.0006. Bootstrap 95 percent confidence interval for the West mean bias: [14.8, 35.2]. Even at the lower bound, West teams are ranked roughly 15 spots worse than their consensus position. The visualization below makes the team-by-team disagreement visible. Teams above the diagonal are ranked worse by RPI than by predictive consensus; teams below the diagonal are ranked better. The West region’s average team in the top-100 RPI has a true ranking 23 spots higher than where RPI places it. The South region, by contrast, sits within 1.3 spots of its consensus position. No other region shows a comparable distortion in either direction. This is signal, not noise. Flaw 7: Predictive Systems Agree With Each Other More Than RPI Agrees With Any of Them Spearman rank correlations among the four independent rating systems run high, with a median pairwise correlation of approximately 0.86. RPI’s correlations with the same systems are lower in median terms: The four predictive systems correlate with each other at values between 0.78 and 0.96 (median 0.86). RPI’s correlations with the same four systems fall between 0.68 and 0.90 (median 0.71). The single highest RPI correlation (ρ = 0.90 with ISR) is itself an outlier within the RPI row: RPI agrees most strongly with ISR specifically because ISR is the only one of the four predictive systems that does not use margin of victory. Strip ISR out of the comparison, and RPI’s correlation with the other three predictive systems falls to a 0.69-to-0.73 range, well below the 0.87-to-0.96 range those three systems show with each other. RPI’s lack of margin-of-victory data accounts for most of its disagreement with predictive systems that do use it.

0-0 curveballs have the highest run value out of all pitches, when thrown in zone. Why? Because it is a FREE strike. Curveballs are swung at just 23% of the time in 0-0 counts when it is IN ZONE. All 0-0 curveballs? Only 15% swings. Knowing this, if you can't land a curveball, it probably isn't a good 0-0 option for you. Understand what pitches you can land and can't. Understand what pitches you are trying to get swings on and which ones you are trying to get takes. Knowing who you are and the purpose behind each pitch is vital to your success.