Making sense of mid-major WNBA draft prospects' statistics: How can we project success?

So, our list of consensus 2013 WNBA Draft prospects should've been posted months ago, probably in October some time.

But, among other things in life, I got obsessed with sidetracked by the challenge of figuring out what to do with mid-major statistics, due in part to two players that our statistical indicators might have overvalued last season, based on the outcomes: VCU's Courtney Hurt and Wisconsin – Green Bay's Julie Wojta.

Hurt's challenge to make the WNBA as an undersized forward have already been documented here, but to summarize offensive rebounding percentage transfers from NCAA D-I basketball better than any statistic and she was so dominant at VCU that it seemed as though she could find a way to contribute to a WNBA roster. Alas, she was drafted by the Indiana Fever in the third round of the 2012 draft and failed to last past the first week of training camp.

Wojta is a slightly different case: as described the other day, it could be argued that she was drafted into a bad situation for any second round pick – the Lynx really didn't have a roster spot available to begin with, so her training camp was more of an extended try out for the league. Yet even though the jury is still out on Wojta as she already has her second chance, the bottom line is that there aren't a whole lot of "successful" - in terms of just making an 11-player roster - mid-major draft prospects in the year they were drafted.

We definitely overestimated the value of mid-major prospects' statistics last year. While that might not be surprising to people who have watched those players closely - and will instinctively dismiss all basketball analytics as abstract folly - it does bring to mind whether there is a way to use statistics to more effectively project the likelihood of success for mid-major prospects.

There are a number of mid-major draft prospects out there this year putting up impressive statistics that probably deserve consideration for the 2013 WNBA draft, making a better understanding of what leads to success a useful exercise.* So if the statistics misled us about those two prospects last season, is there any hope that we can better sort out potentially successful mid-major prospects from the rest using statistics as a starting point?

The obvious challenge with mid-major prospects is that their averages tend to be "inflated" compared to other prospects because they play against a worse strength of schedule than the average prospect in a major prospect - it's easier to put up great per game numbers when the majority of your competition consists players whose hopes of making the WNBA are nearly non-existent. Rates can sometimes be helpful to allay some of that inflation, but it still exists.

So the most common response is to adjust for strength of schedule.

One way to do that would be find a formula that rated every prospect and included some sort of strength of schedule variable to adjust for that. My only problem with that - and why I tend to avoid evaluating draft prospects by single metrics - is that by simply adding a strength of schedule metric you're no longer measuring what the player actually did but who the player did it against. When you look at the most successful mid-major prospects for the NBA, all of them were dinged for their poor strength of schedule by analysts and all have proven that it doesn't consistently tell you much about how productive a player will be.

But that doesn't mean that there isn't something to measuring how a player performed against opposing mid-major talent: you would expect that a WNBA player facing mid-major competition would not only dominate the action efficiently in a variety of ways, but also do so consistently enough that their team actually wins games and preferably by a wide margin.

So rather than go through the effort of constructing a whole SOS model for mid-major WNBA draft prospects, I wanted to see first if I could find any sort of trend using numbers that I already had available to me. And there is one metric that I've already had sitting around in spreadsheets that actually does adjust for strength of competition and then some: Marginal Victories Produced.

I've used MVP most often for WNBA Most Valuable Player awards. It's a metric that measures a player's weighted contribution to the outcome of a game relative to both their team's production and the opponent's (you can read more about it here). Bearing in mind that this metric was designed for NBA players, it tells us how much a player influenced a game. Therefore, against weak competition - or less productive competition - a great player would make their team substantially more productive than their opponents, be responsible for a large percentage of their team's production, and have a larger statistical influence on the outcome than anyone else on the floor; it's not just about big time averages, but efficiency and versatility as well.

The reason to believe that MVP might be useful is that it gives credit to players who contribute a lot to outcomes in which there was a large differential between their team and the opponent. Players who make consistently large contributions to a team that consistently dominates their opponents will thus end up with higher MVP ratings than players on poor teams or players who make an average contribution to a good team. Rather than just adjusting a player's statistics for strength of competition, MVP is actually telling us a bit about the extent to which a dominant player carried their team to whatever success it achieved over the course of a season.

What might MVP tell us about successful mid-major WNBA draft prospects?

I limited the inquiry to the past five drafts, not only because that's when I've been paying most attention to the WNBA but also because that covers the 11-player roster era and the 2008 draft from which a number of the most recent successful mid-major prospects were drafted.

The MVP data for all 37 mid-major prospects from 2008-2012 - some drafted, some undrafted - that I looked at can be found here. The following is a list of players who actually made a regular season roster in the year that they were drafted.

Player	School	Year	MVP
Amber Holt	Midd. Tennessee	2008	22.95
Crystal Kelly	W. Kentucky	2008	21.97
Gabriela Marginean	Drexel	2010	18.65
Courtney Vandersloot	Gonzaga	2011	18.29
Leilani Mitchell	Utah	2008	17.18
Ta'shia Phillips	Xavier	2011	16.78
Amber Harris	Xavier	2011	16.04
Megan Frazee	Liberty	2009	14.88
Tamera Young	JMU	2008	14.64
Chastity Reed	UALR	2011	14.59
Quanitra Hollingsworth	VCU	2009	14.22
Kimberly Beck	George Wash.	2008	13.04
Krystal Vaughn	VCU	2008	11.62
Jene Morris	San Diego St.	2010	11.61
Valeriya Berezhynska	Rice	2008	9.99
Katelan Redmon	Gonzaga	2012	9.16

MVP ratings for "successful" mid-major prospects from 2008-2012. Click here for the full list of prospects.

The hierarchy of prospects might not be entirely clear from eyeballing this list or the full list, but when you start to apply the statistical red flags described previously it becomes much clearer.

Here's what we might be able to take from these numbers, at a very surface level:

2008: It was a very good year for mid-major women

The bulk of those players were drafted in 2008 and ended up being casualties of the 11-player roster limit after that. And really, when you go back further the pattern holds: with more draft picks and more teams in the league, mid-major prospects had a much better success rate just in terms of making a roster at all.

MVP over 14 to make a roster; 17 to contribute

Since 2009, there has only been one mid-major player with a MVP under 14 to spend more than one season in the league after being signed by the team that drafted her: San Diego State's Jene Morris. Gonzaga's Katelan Redmon is technically the other player with a chance to last longer than a season after playing for the Liberty after they drafted her last year, but that is obviously yet to be determined.

Really, were it not for the exceptions of Tamera Young and Quanitra Hollingsworth, a MVP of 16 would be an even better indicator and 17 would be a near lock. Yet we still have to take into account the same statistical red flags that we'd apply to players from stronger programs.

Undersized power forwards

There were a total of 21 mid-major players drafted with a MVP over 14 in the last 5 years. Of them, 10 did not make a roster in the year they were drafted. 6 of the 10 players who didn't make a roster were 6'1" and under; 5 of the 10 with a MVP over 17 were 6'1" and under and didn't make a roster on their first attempt.

And, for whatever it's worth, 2 of the other 5 who didn't make it were 6'2" power forwards so maybe we could increase the height threshold for mid-major interior players. As it turns out, there hasn't really been a stronger pattern among unsuccessful mid-major prospects than the struggles of the undersized interior player, which is not unique to mid-major players.

Inefficient ball handling

Mid-major players with pure point ratings less than -5 have almost uniformly failed to make rosters in the last five years, with one exception: VCU's Quanitra Hollingsworth.

This should also be a familiar red flag for all draft prospects with the difference being a shift in the statistical threshold and a seeming lack of exceptions outside of Hollingsworth.

- - -

As it turned out, MVP really helped to distinguish between successful and unsuccessful mid-major prospects, if we define success by making a roster in the year a player is drafted and apply all the relevant red flags. But I'd be remiss if I didn't highlight one of the major exceptions here: Tamera Young, who has spent her career with the Atlanta Dream and Chicago Sky.

2011 Sky TV feature on Tamera Young.

In short, looking solely at Young's senior year numbers at JMU, I would never have expected her to become a starter for a playoff contender at any point in her career. Remember all those statistical red flags that I described before? Young had nearly every single one of them that applied to a wing, fitting quite firmly into the type of player that I would've considered a questionable prospect based on the past few years on the whole mid-major or otherwise.

Young did have at least one quantifiable positive that was as good as any for a prospect of her size and position: she was an outstanding offensive rebounder. Her 10.43% offensive rebounding rate is better than some highly-touted post players that have entered the league during the same period of time. And beyond that were the intangibles: although she hasn't become a dominant rebounder in the pros, her college rebounding is indicative of the athleticism that has made her a strong defender at the pro level. Combine that with her 6-foot-2 stature and it's easier to see how she's found a niche in the pros.

Of course, perhaps there were some additional variables that helped Young along: She was drafted in a year in which there were simply more spots in the league - the presence of 14 teams with 13 roster spots each certainly gave her a chance to grow into her WNBA role in ways that today's prospects don't have. On top of that, she was drafted by the Atlanta Dream, an expansion team that won four games and started 13 different players in the process; it's fair to say, that the Dream offered any first round draft pick ample opportunity to play.

Does Young's profile render everything we think we've found out about draft prospect statistics is worthless? Hardly. But at the very least it means that statistics are only part of the story - fit with a team, hard work and a host of intangibles, are just a broad set of the unquantifiable factors that contribute to a prospect becoming a successful WNBA player.

We can only look forward to seeing which of the players from recent drafts will get another shot at making rosters in 2013.

For more of our coverage of the draft, visit our 2013 WNBA Draft section.

Notes:

* To the extent that drafting is hardly a science and there is so little data available about the WNBA draft in particular, it's hard to come to any conclusions about mid-major players - since the implementation of 11-player rosters, no more than four mid-major draftees in a year have even made a roster much less contributed much. So with each new prospect drafted, we'll have more data to work with. Nevertheless, I did want to try taking a stab at figuring out a way to systematically understand mid-major prospects better.