SomeGuy
Professor
Posts: 6412
Reg: 11-22-04
|
01-24-19 10:46 AM - Post#273452
Since we had a mrjames sighting elsewhere on the boards today, I figured it might be a good time to try to drum up some discussion of the incoming classes for next year. So I’ll take a crack at ranking the 2019 Ivy classes.
It seems pretty obvious that Harvard is first with five strong to solid recruits led by Ledlum. Cornell appears pretty clearly to be last for now, with only two recruits who both appear fairly under the radar. I’d put Dartmouth 7th right now (Hudson looks best of their recruits at first blush). That’s the easy part.
Tougher is distinguishing two through six. I like Penn as #2, though that may result more from my confidence in the staff than anything else. Lorca-Lloyd is the only recruit on the level of the Harvard and Yale recruits in terms of ratings and offers. However, I suspect that Martz and Dingle will be good fits for Penn and might play right away. Princeton also has a recruit at the Harvard/Lorca-Lloyd level in Langborg heading a four man class. Brown seems to have two mid-major recruits (Cowen and Friday) in a four man class. Columbia has one mid major recruit (Turner) and a good Ivy recruit (Forrest) topping a five man class. Yale has two of top 10 incoming players in the league, but nothing else. So by average recruit, Yale might be #2, but given the depth in all of the other classes, they might be 6th.
Any thoughts as to what I’ve got wrong are appreciated, as are any comments that trying to rank recruits is a generally useless endeavor.
|
mrjames
Professor
Posts: 6062
Loc: Montclair, NJ
Reg: 11-21-04
|
Re: 2019 Recruiting Classes
01-25-19 09:36 AM - Post#273502
In response to SomeGuy
I don't have a lot of specifics to add on this class - I've been paying far less close attention than I normally do.
What the research generally shows is that there are two things to look for in a recruiting class:
1) Depth - More players equals more ultimate win shares earned for the class. There are a lot of random factors in recruiting which can cause players to be overlooked or to develop in ways not thought possible. More draws from the urn with a given class will provide more chances that one or two of those fortunate overachievers will be in a given class.
2) True star players - While it's hard to see much of a discernible pattern between players ranging from "merely recognized as recruits by the agencies, even with 1 or no stars" to "somewhere in that two-to-three star zone," players that truly gain consensus support as three-plus star players (usually in that Top 150 or better range) tend to have a strong hit rate in our league - one that only gets stronger the higher on that scale you get.
Offers (real offers, that is) are important too, though I haven't accounted for those in my model. Anecdotally, though, true offers might be a stronger predictor than anything else.
Harvard might have the only player in this class that ends up in that consensus true star category in Ledlum, meaning that it would likely come down to number of recruits and their true offer set to divide up the rest. I haven't stayed too close to the offers to register an opinion, but that's how I'd frame my thoughts. I just don't have a lot of context beyond that to provide this year, sadly.
|
mrjames
Professor
Posts: 6062
Loc: Montclair, NJ
Reg: 11-21-04
|
Re: 2019 Recruiting Classes
01-29-19 03:56 PM - Post#274039
In response to mrjames
As suspected, the new rankings refreshes are happening, and Chris Ledlum has found himself both in the ESPN and Rivals Top 100. That has driven Harvard's 247Sports class ranking to 28th.
Ledlum also debuted at 8th in the NERR rankings, part of a Top 10 that got a huge makeover due to all of the other transfers this fall that were held out of the preseason rankings.
Idan Tretout checked in at No. 21, which is pretty good considering that the Top 19 players all went to power conferences aside from Ledlum and No. 14 Tyson Walker (Northeastern).
Max Lorca-Lloyd fell to No. 24, but again, New England is strong enough that players in the 21-30 range went to URI, UMass (x2), Richmond and Wake Forest.
Other Ivy Commits:
Tommy O'Neil (HAR) - 31
Jordan Jones (COR) - 46
Cameron Shockley-Okeke (COL) - 64
Marcus Filien (COR) - 100
|
Mike Porter
Postdoc
Posts: 3618
Loc: Los Angeles, CA
Reg: 11-21-04
|
2019 Recruiting Classes
01-29-19 05:32 PM - Post#274050
In response to mrjames
Assumed that Ledlum would really jump up there considering the monster numbers he is putting up.
Also assumed that MLL would fall down the rankings based on #'s to date (which are very similar to last year, but still not scoring a ton which also isn't surprising considering he is playing with Ledlum).
What I find interesting are the national rankings and how they can still continue to completely miss Jordan Dingle. Makes no sense to me... he is playing with a nationally ranked 2020 kid so folks must see him. He is a big, very athletic guard, scoring a lot of points, hitting 3 point shots, played well and put up #'s in AAU. Yet still, when they've updated other kids, they haven't bothered to assign him any ranking at all, not even 2 stars?
I'm excited either way for both Dingle and MLL, but always interesting to see how these rankings develop.
|
mrjames
Professor
Posts: 6062
Loc: Montclair, NJ
Reg: 11-21-04
|
Re: 2019 Recruiting Classes
01-29-19 07:26 PM - Post#274076
In response to Mike Porter
Yeah, even these days, the database never really gets a full scrub. Most of the “adjustment†attention gets placed on the major outlets getting their Top 100s or Top 150s and position rankings right.
The best way around that is to focus on offers, but obviously for the normal case of a kid committing before his senior season, there won’t be any new offers on which to gauge improvement.
It’s tough - which is why my recruiting model doesn’t tend to see a lot of difference in outcomes from merely being mentioned all the way up to 2+ stars. It’s also why I like to focus on reputable regional sites where available, because they can catch an improving or under the radar player and will scrub their rankings at that 2 star level.
|
bradley
PhD Student
Posts: 1842
Age: 74
Reg: 01-15-16
|
Re: 2019 Recruiting Classes
01-30-19 10:55 AM - Post#274136
In response to SomeGuy
Although interesting, it strikes me as very challenging to predict the quality of an incoming recruiting classes and especially specific players based on reality - several years down the road.
Based on y-t-d performance, the best 10 to 12 players are just as likely to have been assigned a 2 star ranking as 3+ above. Additionally, what other colleges recruited these 10 to 12 players as to quality of the basketball program is also very scattered. Looking at the 3 potential IL POY candidates based on y-t-d performance, there is even less correlation as to H.S. rating/ranking and what other colleges recruited them. The absence of Towns and Aiken does distort the overall picture somewhat although injuries are always part of the mix.
It makes sense inherently to have as many 3+ star recruits as possible that were heavily recruited by strong basketball programs but there are many other issues that come into play including the growth or non-growth of players, player development by the coaching staff, their role and fit on a team etc. Some players simply have high basketball IQ even with somewhat limited skills.
I was watching the announcers rave about Makai Mason two nights ago against Oklahoma and thought about how IL teams may have viewed him as an 18 year old.
|
mrjames
Professor
Posts: 6062
Loc: Montclair, NJ
Reg: 11-21-04
|
Re: 2019 Recruiting Classes
01-30-19 02:28 PM - Post#274176
In response to bradley
You're touching on some key points here, but you may be looking at the wrong KPIs.
For instance, it's not the aggregate number of different types of recruits that compose the top players in the league, but the success rates by type that matter. If I have 10 3+ star recruits in my league and 100 2 and below star recruits in my league, and my Top 10 players are split 5 and 5, the correct interpretation isn't that top players are equally likely to be 2 star or 3+ star recruits. Rather, the correct interpretation is that 3+ star recruits are 10X more likely to be top players.
Your point about classes versus individuals is important and is borne out in the data. The model is solidly predictive for classes, merely based on number of recruits, recruits that eclipse that threshold "great recruit" level and number of non-zero (i.e. completely ignored by ratings services) recruits.
That isn't to say that it's impossible (or even difficult) to predict individual recruit performance with some accuracy. As folks have seen, getting reliable information on real offers and pairing that with trusted qualitative information can lead to decently accurate info on which incoming players are likely to have a big impact, including some players that the recruiting services have ignored. That information is often not readily publicly available, though, making it hard to build a reliable model on top of it.
|
bradley
PhD Student
Posts: 1842
Age: 74
Reg: 01-15-16
|
Re: 2019 Recruiting Classes
01-30-19 04:04 PM - Post#274188
In response to mrjames
Kinda see your point but based on y-t-d performance against H.S. ratings (Verbal Commits), the top performer list in some kind of ranked order might be:
Morgan - 2
Oni - 2
Cambridge - 2
Cannady -3
Brodeur - 3
Knight - 2
Lewis - 3.7
Bruner - 3.3
Choh - 2
Stephens -2
Barry -2
Bassey- 2.7
Stefani - 2
8 out 13 received 2 ratings and the top 3 POY candidates received 2 ratings and were not heavily recruited. I did look at the list for each IL team with player ratings greater than 2 and it is pretty hit and miss relative to actual performance. The list and order may obviously change over time.
If you look at the top 3 performing players for each IL team, there is some correlation to H.S. ratings but not much more than some correlation and in some cases none. It may be simply a function of not very accurate assessments of player capabilities as some of the comments regarding strengths and weaknesses by individual players were laughable, i.e. Desmond Cambridge.
Based on overall team H.S. ratings, Harvard would be the strongest team by far but they might well be the case if Aiken and Towns were playing.
I just think that there are other factors as important than H.S. ratings and who attempted to recruit a player. and at the end of the day, peformance. The reality check to various theories and analytics is peformance although injuries are certainly a wild card.
|
CrimsonBlood
Freshman
Posts: 46
Age: 33
Reg: 03-15-18
|
Re: 2019 Recruiting Classes
02-17-19 04:25 PM - Post#276996
In response to bradley
Chris Ledlum looks like the kind of player who can come in and make an immediate impact. Kind of reminds me of Kyle Casey-- someone who is long and can play inside/outside and throw down some ridiculous dunks. Chris is probably thicker/more physical.
Way too early but can't resist thinking about a starting 5 next year of:
1: Aiken
2: Bassey
3: Towns
4: Ledlum
5: Lewis
Seems like a top 25 team to me if they stay healthy and can gel (and assuming Towns can return to the same level of play).
|
HARVARDDADGRAD
Postdoc
Posts: 2691
Loc: New Jersey
Reg: 01-21-14
|
Re: 2019 Recruiting Classes
02-17-19 04:32 PM - Post#276999
In response to CrimsonBlood
Crimsonblood - suggest moving that over to the Harvard thread
|
mrjames
Professor
Posts: 6062
Loc: Montclair, NJ
Reg: 11-21-04
|
Re: 2019 Recruiting Classes
02-17-19 05:05 PM - Post#277007
In response to bradley
Just saw this now that the thread got bumped.
Makai Mason was No. 18 in NERR's 2014 New England rankings, just ahead of kids that went to Michigan, San Diego St. and WVU. There were pretty high expectations for him coming in as there are for all of the recruits that are Top 20 in New England.
I have a recruiting database that goes back to 2002 and incorporates a variety of ratings systems (it's much more useful for the 2008 class and beyond as the ratings systems became more mature and consistent after that point). It also includes actual production at the college level (win shares). Empirical evidence shows that ratings have some predictive value and that predictive value grows dramatically the higher the ratings are. Anecdotally, I imagine that true offers would increase that effect (Paul Atkinson and Devin Cannady spring to mind as the recent most notable "offers vastly outpace their rating and the rating never caught up" candidates).
But no, this isn't just my opinion I'm throwing out there - it's the result of putting together a database of 500+ Ivy recruits and comparing ratings to performance.
At the same time, you're absolutely correct that there are other things that matter to ultimate performance than ratings and offers. The model does a nice job of identifying players with a high likelihood of being productive. At the same time there will be a much, much, much larger pool given a low chance of being productive and while you wouldn't want to bet on any individual player in that pool, in the aggregate, you'd expect a handful of players to emerge from there each year.
For as much as people called the win share prediction around the 2016 Harvard class (~55-60) crazy, even with all the injuries, they're going to end up around 40 at the end of this year, within striking distance of hitting that number with one more season to go.
|
|
