The likely reason the Cavaliers took Tristan Thompson over Jonas Valanciunas

English: Jonas Valanciunas, member of the Lithuanian Under-19 basketball team, which competed in the 2011 FIBA Under-19 World Championship in Latvia Lietuvių (Photo credit: Wikipedia)

The likely reason the Cavaliers took Tristan Thompson over Jonas Valanciunas

A major surprise of the 2011 draft was Tristan Thompson jumping to 4th overall by the Cavaliers. Jonas Valanciunas had been mocked higher throughout the year and conventional wisdsom says filling the C position is more difficult than PF.

Because Jonas had a European team buyout expected to mean a 1 or 2 year wait until he came to the NBA, many speculate this – or either a claim Jonas or his agent Leon Rose speaking out against playing in Cleveland – caused the Cavaliers to waffle at the last moment and pass on Valanciunas for Thompson even if they preferred the former.

I don’t see this a logical explanation. In regards to the buyout, the Cavaliers were years away from expecting to compete and thus immediate production for Jonas wasn’t any more needed than in Toronto. Hard to see why evidence why Jonas wouldn’t want to play in Cleveland either.

There’s a more simple explanation for why they made the pick. It’s no secret that the Cavaliers draft methodology under Chris Grant is advanced metrics heavy. The recent surge of advanced metrics statheads focusing on the draft, is based on this concept – Take all the careers of NBA players, then perform an advanced regression analysis to find which college statistics correlate to NBA success. The more adjustment for NCAA conference and competition, age, or teammates, the better of course. With millions to spend at their disposal, the complexity of the Cavaliers’ regression analysis is likely beyond anything we are able to see publicly on the internet. But if you’re a Cavaliers fan looking for a hint of who they’ll pick, check out http://hoopsanalyst.com, a blog run by a draft stathead who gained attention for predicting Jeremy Lin’s college success. The site is relevant for Cavaliers fans because he ranked Tristan Thompson as the 2nd best prospect in 2011 and Dion Waiters as 2nd best in 2012, behind Kyrie Irving and Anthony Davis respectively. The methodology he uses is thus likely to match up well with the Cavaliers’. Because of that site I correctly predicted before the 2012 draft that they Cavaliers would take Waiters 4th over Barnes who had dominated the mocks at that pick, but whom statguys universally hated as a top 10 pick. Another draft site getting attention this year is http://shutupandjam.net/draft-rankings/ , albeit he didn’t have his rankings up last year I believe. John Hollinger for ESPN also posted draft regression studies for years, while his replacement Kevin Pelton recently posted an article comparing his stats to scouting opinions, here http://insider.espn.go.com/nba/story/_/page/PerDiem-130329/nba-shabazz-muhammad-overrated

Tristan Thompson’s high rebounding and block rate and Dion Waiters high assist and steal rate, appears to have played very well with draft studies. However my real reason for posting this article, is why the Cavaliers would be unlikely to trust any of Jonas Valanciunas’ statistics before the NBA.

Yes, Jonas’ combination of rebounding, shotblocking and hyper-efficient finishing is traditionally what regression studies like the above would favor. If he had done it in the NCAA.

In 2010-2011, Jonas played 33 games in the Lithuanian Basketball League (LKL) for Lietuvos Rytas Vilnius, averaging 21:04 minutes averaging 11.5 points, 7.3 rebounds, 1.8 blocks and 66.8% FG, which is elite per minute numbers. But from a regression analysis perspective, these minutes and production are unusable due to all but no players crossing over from the LKL to NBA before Jonas, let alone young center prospects. The LKL is not a strong league and without comparable players in his situation and then gone onto the NBA, there’d be no way to knowing whether to trust his production beating up inferior competition.

Lietuvos Rytas played 15 Gs in the Euroleague in Jonas’ draft year, where he averaged 15:25 per game averaging 7.7 points, 5.8 rebounds, 0.7 blks on 70.8% FG, once again elite per minute numbers. While against more trustworthy competition, the problem here is sample size. He plays a total of 231 minutes total, which is just too small of a number to trust as indicative of real quality. Not to mention Jonas’ role playing as an “energy guy” reserve off the bench against opponents unlikely to have gameplanned against him, makes it more difficult to extrapolate his minutes to a full game as star caliber. In comparison, Tristan Thompson played over 1100 minutes his freshman year at Texas and as a featured player which is a far more trustworthy statistical situation.

A similar story is true of international competitions before the 2011 draft, such as the U18 tournament in 2010 where Valanciunas averaged a dominant 19.4 points, 13.4 rebounds, 2.7 blks 70% shooting split. However playing approximately 270 minutes in the tournament as well as the relatively low level of competition, makes the information all but useless.

While they may have more future NBA players than the LKL, the sample size of NBA players in the Euroleague and international competitions available to cross-compare in a regression analysis, is still fractional compared to a major NCAA conference’s.

It’s likely the Cavaliers were fans of Jonas’ efficiency and rebounding. But if truly all-in on using advanced metrics and regression analysis to predict draft success, it’s no brainer who’s statistics would be more trusted between Valanciunas and Thompson’s. From a regression analysis perspective, there wouldn’t be enough if any trustworthy statistics from Jonas’ pre-draft career, whether from level of competition or small sample size in leagues and competitions. Jonas’ path to the NBA is literally one of a kind in regards to competition. On the other hand Tristan playing over 1100 minutes in a role as well-treaded as Big 12 freshman power forward, would give the Cavaliers a much larger sample size of comparable players and ability to adjust for context that they needed to make a reliable regression analysis.