With portal season heating up, I wanted to take a look back at our past transfer hauls, where each player ranked coming in, and how they actually performed. My plan is to use this post to break down our past portal classes, and then come back once we have a roster in place to analyze the current class and what it means for the upcoming season. All data displayed is from EvanMiya.com.

Why is EvanMiya the gold standard for college basketball analytics? Historically, sites like KenPom and Torvik leveraged advanced metrics to analyze teams on a per-possession basis. This helps normalize analysis for pace and puts fast and slow teams on an even playing field. EvanMiya builds on those tools by starting with line-by-line, play-by-play data. Because of this, his model generates metrics that account for pace and for who is actually on the floor. If a game is in junk time and a bench player scores a ton of points on the other team's benchwarmers, this system accounts for that directly and won't artificially prop the player up.

What are the metrics?

- OBPR: The impact a player has on the offensive side of the floor. An OBPR of 0 is the average across NCAA D1. A player like Jason Preston with a 4.81 is well above average, and someone like Jalen Breath with a -1.93 is below average.
- DBPR: The impact a player has on the defensive side of the floor. Similar to OBPR, the D1 average is 0. Someone like Jason Carter with a 2.35 is an above-average defender, and someone like Dior Conners with a -1.78 is a below-average defender.
- BPR: OBPR + DBPR. This is the player's overall impact on the game. Unsurprisingly, the highest of the Boals era was Jason Preston with a 6.37 (4.81 OBPR + 1.57 DBPR). Others that stick out include Mark Sears with a 5.44 (4.67 OBPR + 0.77 DBPR) and Jaylin Hunter with a 5.39 (4.37 OBPR + 1.02 DBPR).
- Projected BPR: A projection of how good a transfer portal player is going to be the following season. This attempts to account for situations where a player may have been elevated or held back by their previous team being exceptionally good or bad.

Let's walk through past portal classes, starting with 2022 and going through last year. (Note: Players who transferred up from non-D1 schools, like Tommy Schmock or Vic Searls, do not have data for projections).

Past Portal Prospects (Year - Player: Previous Season BPR | Projected BPR | 1st Season Actual BPR)

- 2022 - Jaylin Hunter: 2.36 | 2.56 | 5.49
- 2022 - Gabe Wiznitzer: -0.77 | 0.04 | -3.76
- 2022 - DeVon Baker: -0.06 | -0.59 | 0.37
- 2023 - Shereef Mitchell: -0.86 | -0.08 | 2.47
- 2023 - Ike Cornish: -2.15 | -0.65 | -2.29
- 2024 - Jackson Paveletzke: 0.89 | 0.58 | 1.15
- 2025 - Javan Simmons: 1.94 | 2.70 | 2.20
- 2025 - Jalen Breath: 0.16 | 0.25 | -1.39
- 2025 - Dior Conners: -0.90 | -0.51 | -2.82

Who were the biggest wins relative to their projections?

- Jaylin Hunter: Came in as a very good portal prospect and emerged as one of the best players in the conference for two years. His defensive impact metrics really stick out - he was a good offensive player, but his defensive stats made him an elite MAC player.
- Shereef Mitchell: Outperformed expectations. Worth noting that his first two seasons at Creighton saw numbers similar to what he did at Ohio, so his projections were likely skewed down by a single bad season.

Who were the biggest losses relative to their projections?

- The Wiz: Projected to be a solid rotational piece but netted out his first season as a -3.76, one of the worst rotational seasons in the Boals era. Fun fact: The Wiz was the first player that honed me in on the EvanMiya data. I had high hopes for a former 4-star, 6-11 center, but noticed early on that the other team tended to go on runs when he was on the court. The data very much supported that theory.
- Conners and Breath: Both ranked as roughly average portal prospects, and both saw significant regressions from where they were in prior seasons.

Moral of the story: players projected to be high-value prospects (Hunter/Simmons) tend to hold up and, in the case of Hunter, outperform projections. Players projected to be negatives tend to remain negatives, and average players tend to remain exactly that. Are the projections perfect? Absolutely not. The model needs enough historical data on a player to be accurate, and it really struggles to predict outcomes for guys who transfer down without much playing time. The model forces those players toward average, which makes them incredibly volatile - a Daniel Freitag at Buffalo can project as average and emerge as a top player in the conference, while an Ike Cornish can project as average and turn out to be a negative. Even with players who have a lot of minutes on tape, it is hard to account for system fit from one spot to another.

I plan to follow up in this thread when the portal haul is complete to compare where the team projects out relative to past teams. As a sneak preview, Kyler D'Augustino projects as an even better portal prospect than Simmons. EvanMiya predicts him to be a 3.13 for next season, making him the highest-rated portal prospect we have landed in the portal era. It is also a huge plus that he is a local kid and will likely bring in some extra attendance.

Great info, Quant. It's always been apparent one of our staff's strengths has been adding PGs via the portal. This shows that and I expect continuity with that in KDA.

Simmons aside, it's everywhere else where seem to be lacking which has shown in the roster imbalance and on court results.

From QuantCat's post: "Note: Players who transferred up from non-D1 schools, like Tommy Schmock or Vic Searls, do not have data for projections."

As for Vest and Dwight, EvanMiya is pretty new, so he doesn't have data going back that far. But if I had to guess, I'd wager Dwight was a strong positive, beyond what was expected, and Vest was in the 0.5-1 range.

Part 1 - The How

Now that we're closer to a full roster, I wanted to follow up on the earlier post, and am breaking this up into 4 parts:

Part 1: A quick overview of what I did and how
Part 2: An overview of OU
Part 3: An overview of MAC
Part 4: A final wrap-up

My original plan was simple: look at who each MAC team brought in through the portal, compare it to who they kept, and figure out what it all meant for the Bobcats. I figured it would come together quickly like it has in years past. Then the staff went international, things slowed down, and I suddenly had time on my hands. So, with that time and a little boredom, I started pulling in a ton of data:

- All EvanMiya (EM - https://evanmiya.com /) player and team data back to 2016-17
- All 247 and On3 recruiting data back to 2018
- RealGM profiles and stats for all 16,000 players who touched NCAA Division 1 basketball from 2019 to today. The nice thing about RealGM is that it carries a player's entire basketball history: international play before and after college, D2, D3, JUCO, NAIA, and national team play (think U19 World Cup).

One quick note before going further. The number I keep referencing is EM's BPR. The simplest way to think about it is a single number for how much a player (or team) helps win games. A 0 is roughly a replacement-level Division 1 player, positive is good, negative is below that line, and the further from 0, the more extreme.

I used all of that data to build models that predict a BPR for every player and every team, every season. Players get sorted into categories, and each category has its own model built specifically for it:

- Freshmen: uses 247/On3 recruiting data plus a 3-year rolling average of the team's strength. Unranked guys are graded off the strength of the team they're going to (an unranked commit to Akron is probably better than one to CMU). Elite prospects like Cooper Flagg get an extra bump if they're projected as lottery picks before the season.
- Returners: looks at a player's trajectory (steady climbers vs. guys who hit the same number every year) and how many years they've played. Returning redshirts are modeled off prior experience if it exists, or dropped to a lower team baseline if they're a Carter Reese type.
- D1-to-D1 transfers: starts with the EM projection, then moves up or down based on the destination team and other factors.
- Non-D1 transfers: measures a player's strength relative to other non-D1 players, factors in the team they're heading to, and projects them off historical data.
- International recruits: there have been fewer than 250 international players (with no U.S. high school experience) in the NCAA over the last 7-plus years. There have, however, been a huge number of players go the other way, from the NCAA to international leagues. So my approach was to build a "dominance score" for a player within his league, find former D1 players who dominated that same league at a similar level, and comp his projection against how those guys did in college.

Every player and team then gets three numbers: a Floor, a Base, and a Ceiling. By design, that range only tries to capture the middle 50% of outcomes. Across the last 7 years of data, the actual result across NCAA D1 teams land inside the predicted range about 50% of the time, with 25% overperforming and 25% underperforming. That is on purpose. It's impossible to predict things like injuries, or a team that suddenly gels (or one that never does).

Part 3 - The MAC

If we have a solid group of returners, D1 transfers, and international guys, where does that leave us against the rest of the MAC when the same models are applied across the board? To be clear, this isn't me manually typing in numbers I like. These are automated models grounded in 16,000-plus players across many seasons, and each player type (international vs. returner vs. the rest) has its own model fit specifically to it.

That said, I did use my own knowledge of the MAC and Ohio to sanity-check each model and tweak where I saw problems. For example, the early returner model badly underrated last year's Miami team, projecting guys like Eian Elmer to get worse year-over-year. There was no world where a player with Elmer's clear track record of improvement should be projected to decline, so that became a validation point. Every model went through multiple rounds of checks and iterations to make sure it landed somewhere sensible.

Now, the state of the MAC. Using the same models on a fully automated basis, here is how it ranks out. Teams with an asterisk still have roster spots to fill.

1. Miami (3.9 - 9 players above 0): Miami lost a ton of talent (Byers, Suder, Woolfolk, Elmer) but kept key pieces and added some strong transfers. Their biggest strength is far and away the best guard trio in the league: Luke Skaljak (projected MAC Player of the Year), transfer Stevie Elam (4th in the MAC), and Evan Ipsaro (7th in the MAC). The question mark is what Steele gets out of his bigs and wings. Almar Atlason, Rich Rolf, and Preston Copeland project at 1.1, 0.9, and 0.6. Can Almar break out? Can Rolf build on his outlier 1.4 season at Youngstown? Can Copeland take a Year 2 jump on a better team? Miami is the clubhouse leader heading into the season, and you can probably bank on that big trio outperforming.

2. *Ohio (3.3 - 5 players above 0): The individual projections have Ohio holding 3 of the top 11 players in the league entering the season. This is a roster with very good talent at the top. The potential weaknesses are depth relative to WMU/Akron/Miami/Kent (there's a hard drop-off past the top 5) and the volatility that comes with international recruits. We'll lean heavily on those international guys hitting, and the data shows that while the model is serviceable there, those players can be very hit or miss.

3. Kent State (1.6 - 8 players above 0): Kent brings back Quinn Woidke (projected 2nd in the MAC), who in almost any other year, without Leroy Blyden, would have been the league's Freshman of the Year. They also get Jamal Sumlin, a 2.3 player in 2024-25, back from a season lost to injury. Around them is a strong group of transfers plus a highly touted freshman in Dhani Miller, who has as good a shot as any MAC freshman at a Blyden/Sears-type breakout. Like Miami, Kent will have excellent guard depth and will likely play three guards at all times. What makes or breaks their season, and could make them a title winner, is their less-proven group of bigs.

4. *Akron (0.6 - 10 players above 0): Akron lost both a coach who always seemed to have his teams overperforming, and a significant amount of talent. The model nailed their dominant season last year, projecting them as the 75th-best team in the country (9.0); they finished 76th (9.6). They brought in a mix of proven mid-major players (Staveskie, Pence, Hammer, Burton) plus two high-upside guys in Colin White (Ohio State) and Exavier Wilson (Kansas State). In two seasons with real minutes at Ohio State, White posted a 1.5 and a 0.0. For comparison, Evan Mahaffey went 0.6, 1.5, and 3.9 over three years at Ohio State before jumping to a 5.4 at Akron. White has real breakout potential in a bigger role. This team's ceiling won't be what it was in 2025-26, but Akron will be a contender, like they always are.

5. UMass (-0.8 - 6 players above 0): UMass pairs two key returners, Jayden Ndjigue (2.7 in 2025-26) and Danny Carbuccia (0.7 as a freshman and a prime breakout candidate), with three key D1 transfers in Jordan Clayton (Northwestern), Junjie Wang (San Francisco), and Abdullah Ahmed (BYU), plus a key international addition in Noa Zemljic. They are massive, with six players 6'9" or taller alongside a good guard tandem and a wing in Jayden. If the new group gels and they get some freshman contributions, they could be a threat. The downside is a lack of depth, and top-end players that project a bit lower than the other teams at the top of the MAC.

6. Western Michigan (-1.1 - 8 players above 0): This is the breakout candidate from the bottom of the league. Over the last six years, WMU has always projected somewhere between -7.7 and -12.8, and has never actually finished better than the -9.3 they posted last season. This year they absolutely crushed the transfer portal at every position and added a solid international center in Janis Nemann. Here's the big caveat: several of the models, particularly the transfer and freshman models, use each team's 3-year rolling average as a predictive input. If you replace WMU's rolling average of around -10 with a 0, this team jumps all the way up to right behind Ohio in the rankings. This is not the WMU of the past. They're going to be a real threat.

7. Toledo (-2.2 - 6 players above 0): The Rockets bring back three core returners, led by Austin Parks (1.5 in 2025-26, projected for 2.3 in 2026-27) along with Will James (1.7 projected) and Mynor Strong (1.5 projected). This is a team that lost its three best players (Blyden, Craig, Wilson) and replaced them with transfers in Sean Jones, Rich Barron, and Angel Montas. Those transfers all have 3-plus years of EM data suggesting they'll land as average MAC starters and not much more. The wild card is their four freshmen. Todd K has a real track record of finding talent other teams overlook, and while it's unlikely, another Blyden-type could once again lift this team into contention.

8. Bowling Green (-4.2 - 7 players above 0): BG has a solid group of about eight players projected between -0.3 and 1.7, made up of two returners, four D1 transfers, and two non-D1 transfers (Todd Simon has a strong track record of finding lower-division talent). This is a team with real depth that should protect them from bottoming out, but becoming a contender would require multiple breakouts or the team playing well above the sum of its parts. They've been feisty the last few years and should remain a tough out.

9. Ball State (-5.2 - 6 players above 0): This is a team with only three returners and a new staff that completely rewired the roster. Their seven D1 transfers include two high-major "transfer downs" with very wide ranges of outcomes, plus several proven, solid mid-major players. Looking through the data, high-major transfer downs (Aaron Fine and Jazz Henderson) are very unpredictable: most land around their new team's average, a lot bust (like Cornish), and a select few have Daniel Freitag-type seasons. This team will be competitive, but in my mind not a true contender at the top.

10. Buffalo (-6.8 - 3 players above 0): Another team with very few returners, only one of whom played real minutes last year (Kyle Jones, 280 total possessions). They have seven D1 transfers, two non-D1 transfers, one freshman, and four "returners," only one of whom has touched the floor. Their portal haul is interesting: they landed a true stud and All-MAC contender in Drew Steffe out of Southern Illinois. The other six transfers have real playing experience, but only one of them has ever rated above a 0 by EM. This could look a lot like last year's Buffalo team, one dominant player who can win games with an unproven supporting cast.

11. *Eastern Michigan (-7.6 - 3 players above 0): I'll start by saying Billy Donlon was a phenomenal hire for EMU. This is a coach who was fired after a 22-win season at Wright State (where he had 19-plus wins in four of six seasons, three of them 21-plus) and again after 19 wins at UMKC in 2021-22, before spending the last few years on Brad Brownell's staff at Clemson (the same guy who hired him as an assistant at WSU). He'll make the most of limited resources at EMU. He's already retained MAC All-Freshman honoree Mohammad Habhab along with Greg Lawson and Braelon Green. Beyond the returners, they have three slightly unproven D1 transfers and six non-D1 transfers. With two roster spots still to fill, they project as a team that can win some games under a good coach, but whose ceiling is probably sneaking into the MAC tournament.

12. Central Michigan (-9.1 - 3 players above 0): This team will be led by returners Phat Phat Brooks and Keenan Garner. Beyond those two, it's a whole host of new faces: one D1 transfer, three non-D1 transfers, and four unranked freshmen. That comes after last season, when their coach (who came from a D1 program) brought in nine transfers from the lower ranks, meaning in two years they've added 12 non-D1 transfers (look away, Fishbates). It's tough to see a path where this team contends, let alone makes the MAC tournament.

Nicely done, QuantCat. Great stuff here.

Any analysis on Chansey Willis at Kent? To me, he's the biggest incoming transfer given his already body of work in the MAC and now he's with Senderoff.

After digesting this more, it seems pretty important that our staff finds use of at least 1/2 remaining schollies. Especially on the guard front. I echo the thoughts that Che, Kuany, and Moek end up being the more impactful bench players this season. How impactful will be the story. I think everyone on this board is excited about the potential of starting five including Javan, Dusan, Siebe, Kolten, and KDA.

Western Michigan -- yea, this team will be a problem. Excluding Junemann and other additions (#189 overall player and #23 PG Jordan Sigmon), WMU's D1 transfers alone combined for 15.8 win shares this past season. Only 2 out of the 8 players started at least 30 games.

6'8 Augustinas Kiudulas might be due for a nice bounce back after going from a starter at VMI (15ppg, 3.2 WS) to a role player at Colorado State. 3 others in this portal class are 6'6 and including Kiudulas, all hit the 3 at least 39%. This is going to be a very tough team.

Willis is an interesting one, as he is very likely under predicted.

His first 2 seasons were 1.8 (2024-25 WMU) and 0.9 (2025-26 Minnesota). The EM projection for him in the transfer portal data was a 1.1 which got weighted down by his regression at Minnesota. My projection doe transfers uses the EM portal number as a starting point and adjusts up or down. Mine saw the peak of 1.8 and a team destination of Kent and bumped him up to 1.3.

This Kent team has even more depth at guard than Miami. Miami has the 3 headed monster, but Kent may very well go 5 deep with All-MAC potential quality guards. Kent always plays more of a defensive minded game so interested to see if Senderoff goes full small ball.

Great data and analysis Quant. I live for this stuff.

7/16 Update based on 2 core changes:
- Ohio adding Marin Petkov
- Akron getting Evan Mahaffey back for a 5th year as a result of an ongoing lawsuit related to 2022 HS graduates getting only 4 years of eligibility.

Projections for both of the 2 new players:
- Marin Petkov - Ohio (-0.6 - International): Petkov is coming from the MA Superleague, which ranks 43rd/46 in league strength, almost even with the BiH Liga that Dusan was just the MVP of and was 44th/46. His archetype is like a taller Dior Conners, and is a good shooter and facilitator but the defensive metrics were a knock to his overall impact. His projection is just behind Brogan, and I think Brogan is more likely to see bigger minutes because of his defense. Brogans defensive and playmaking ability may force him into lineups despite some of the shooting/finishing. Brogan is the guy who might be able to guard the best guards in the MAC.

- Evan Mahaffey* - Akron (5.8 - returner): Mahaffey was the best player in the MAC last year according to the EM data at a 5.4 and was criminally underrated in the all-MAC teams. There is nearly a 2 point gap in the 26/27 projection between him and the next best guy in the MAC in Skaljac out of Miami. The asterisks is because Mahaffey may not even be able to play, and is only on the roster because an Ohio Judge granted a set of Ohio college athletes an injunction against the NCAA. This set of players are all 2022 HS grads suing the NCAA for an extra year of eligibility, because they just graduated without the option to get the 5th year that the new NCAA policy gives to everyone after them. His presence swings Akrons projections as a team massively.

New MAC Team Projections:
1. Akron - 6.2 (Up from 0.6 solely because of Mahaffey)
2. Miami - 3.9
3. Ohio - 3.3 (Petkov as a -0.6 does not move us up at all but does add depth)
4. Kent - 1.6
5. UMass - -0.8
6. WMU - -1.1
7. Toledo - -2.2
8. BG - -4.2
9. Ball St. - -5.2
10. Buffalo - -6.8
11. EMU - -7.6
12. CMU - -9.1

Petkov is still a decent add and a needed 4th guard, but the gap between the Ohio starting 5 and the bench is still a lot bigger than the other highly rated MAC teams. The starting 5 will be one of the, if not the best in the MAC, and the success of the team will come down to who from outside the top 5 can outperform their projection and potentially some luck in the injury department.

Given where we were at a few months ago, I still think this offseason has to rate as a solid A and the team should be a real contender.