with seble's patch in place to balance the early entries, i figured it would be worth taking a look at the whole class and seeing how it looked, and having a discussion on that. what seems to be working well? what, if anything, still seems to be a problem? and to clairfy, seble's change is this:

"Improved the logic to rank players for the NBA draft to provide better balance between the 5 positions. Also some players may be evaluated for a different position than their listed one if their ratings indicate a different natural position."


from this comment and his posts on the forums, the way I read this is that now players should be more balanced talent wise who go early - instead of having 20 bigs and only the top 2 guards. also, players who were, for example, PFs who were really SFs, who might not have the killer reb/ath some PFs do, but have a bevy of other strong ratings, may no longer be "hidden" from the draft, and should be drafted (assuming they are good enough) because seble is now considering them as SFs. my guess is he has some number of ranking algorithms, who knows if its 3 (1 for bigs, 1 for guards, 1 for SFs), or 5 (one per position), or more like 10 (one per major role). but basically i think he calculates the score for every player for every position/role/whatever, and then takes their "best" scores, and uses that to figure out if they should be drafted or not. now, to see how that (and previous changes worked) -

i started simply by going down the early entry list from tark, one by one, and identifying the school, player position, and player class of each early entry, to get a high level picture.

duke - soph   C
duke - junior  PG
duke - soph   SG
duke - junior  SF
duke - soph   PF
conn - junior C
conn - frosh  PG (not eye popping)
conn - junior PG (not eye popping)
texas - soph  SG
texas - junior PG
texas - junior SF
pitt     - soph  C
pitt     - soph  PF
wava - junior SF
wava - soph  PG
fl st    - junior PF
fl st    - soph  PF
io.st. - junior PF 
unc   - junior SG
indi   - junior SF (indiana was 4-23, but the sf was rated 1005)
gtow - soph  PF
mary - junior PF (maryland was 6-21, but the PF was rated 947 and strong)
notre - junior SF
vand - junior SF
wash - junior C
zona - junior PG
wa.st - junior PF
lville  - junior SF
mi st - junior SF


so, at a glance - 29 total early entries, 1 fr, 9 soph, and 19 juniors. 4 Cs, 8 PFs, 8 SFs, 3 SGs, 6 PGs

a few thoughts on that high level picture:

1) the distribution by position is a lot better than it was (when we were seeing like 22 EEs, 1 guard EE). 

2) the # of EEs is slightly higher than we are used to, but i expect that is the kind of thing that could settle down over time - seble's change, as i mentioned, should cause great players who were "hidden" by their listed position, to wind up going early. so i would naturally have guessed this could temporarily increase the # of EEs, as the "hidden" plays get found out and go pro.

3) the # of juniors seems pretty high, relatively. it seems juniors have been going more in other drafts too, which i think it fine. however, this difference (between juniors vs sophs vs fr) could also be over stated, for the same reason - great sophs who were "hidden" by their position are now juniors, and there are so many good juniors, it makes sophs less likely to go. over time, i would expect this to even out a bit.

4) only 1 freshman went pro. i think thats a good thing, in general. 9 sophs is really not that bad either.

5) although its not from that little 1 line snap shot i posted, i did look at, for at least a few seconds, every single player that declared early. all in all, i thought they were all VERY GOOD players, who, if you just knew the player (and not the situation), you wouldn't be surprised at all they left. in that regard, i think seble did a good job.

so, all in all, from the high level standpoint, i think things are looking pretty good. thoughts on that?

now to get into the nitty gritty a bit. when you dig deeper, there are a few things that should at least be considered. are they problems or not? i can't say for sure, but i'd love to know what you guys think. anyway, here are the interesting details that might not seem to fit:

6) while only 1 FR went pro, im surprised at who it is. like i said, i looked at every player, and really only 1 of them really stood out to me. UCONN lost a FR PG, after a final 4 run (which makes it better), but i was really surprised at the players ratings. you can check out some of the other early entries yourselves, but they have a lot of high 90s it seems or else are just very loaded, 950+ type guys. well, here are the stats on OR's lost fr guard - 91 ath, 90 spd, 25 reb, 82 def, 4 sb, 39 lp, 77 per, 89 bh, 82 pass, 40 we, 82 sta, 64 dur, C ft... that does not look like a guy who would go pro EVER (not with his current ratings), and least of all, as a freshman! maybe he was one of those guys who was like a 5% chance to leave and OR got a bad dice roll, i don't know, but i dont think its good for people to lose freshman rated like that. now 1, its hard to get up in arms about :) and it was off a final 4 school, at least.

7) there were 29 early entries. we all know at the top end of top programs, you are largely splitting hairs on your matchups, and often have many great players who have very slim margins above/below the guy they are facing. to me, this suggests that you should see a pretty even spread of EEs across teams. 

this years EE crop included - 5 lost by Duke, 3 lost by Texas, 3 for UConn, 2 for pittsburgh, 2 for Florida State, and 2 for West Virginia. now, to be fair - Duke was the runner up, and emy had one of the best recruiting classes ive ever seen recently, and some other great ones recently, so he was about as well set up to lose 5 as you could. UCONN made the final 4, Pittsburgh won the championship, W Virginia was an elite 8/ #3 rpi team, and Florida state was sweet 16/30 rpi. But texas. who is not a powerhouse program loaded with studs, was only 44 rpi/1st round.

now, I have 2 problems with that. first and foremost, a 1st round team who legitimately has a couple studs amongst some decent players should not be totally decimated. the coach had 3 fantastic players, which he doubtless worked very hard for. you cant reload on a B like you can on an A/A+ like most of the multiple-player losers. it just totally ruined everything that coach has worked for for the last few years, and he left the school, which i cannot blame him for. my personal opinion is that the single biggest issue in EEs prior to sebles work was the ability to lose too many players for the situation. so while each of those 3 players were surely draft worthy, it is neither fun nor good for the game for something like this to happen to a school like Texas.

its also a bit suspect teams with 4-23 and 6-21 records had guys go early, but it was 1 each, and they were really studly, so i guess i can live with that.

i also have a problem with emy/duke losing 5 early. just because he did a bang-up job recruiting, just killer, and only had 1 senior - doesn't mean he should not get to see the majority of his team leave. when you are a very legit runner up with 1 senior, you should have a shot at a title the next year. again, its not fun for anyone, nor is it good for the game, to decimate these programs so hard.

so, i will once again conclude that while looking better, the recruiting system still has a major flaw, the same one its had for a while. there are 2 fairly simple options in my opinion. first, put artificial limits in place. teams that dont miss the post season have chances of losing a stud early very very much diminished, unless they have multiple studs. i think one good player should not be taken away, from programs with already so much work to do to compete, when he is the only remotely draft-worthy player.

and really any team could be artificially capped similarly (if you want to keep it simple). figure out how many guys they have who are draft potential, and make sure you never have more than half leave, something like that - even that one simple rule would probably fix 95% of the duke/texas kind of situations. or you could just cap teams who aren't top 16 at 1 EE, cap teams who aren't top 5 at 2, and cap any team at 3, something simple like that.

however, id rather see a less artificial solution, not a cap, but its not as easy to implement, although based on the complexity of what seble has done, i think its rather doable. let your best guy go EE first, and then cut the already existing random factor on the EE chance on all other players on the team. the cut should be much more significant for teams who are not top 25 or so, and should be more significant after the 2nd, 3rd, 4th EE than the first. i dont think 2 its really that bad. but 3+ just sucks, no matter who you are, and its very rare there is a team so good that they are head and shoulders above competition after losing 2 of their best players. and really anything above 3 for the best teams seems overly harsh, while anything over 1 for non top 25ish teams seems overly harsh (3 at TX is just cruel!!)



anyway, those are my thoughts. looks to me like seble's latest change was good, but still did not address the major issue of how to handle multiple early entries at a school, to keep TX and Duke situations from happening.

Jeff, I agree with almost everything you wrote here. A couple things to highlight:

-The one point that we don't really agree on is that 19 juniors and 9 sophs seems a little skewed towards too many juniors. I actually think the opposite, and that almost all early entries should be juniors, with just a handful of really elite sophs. Because of IQ restraints, players are never great as freshmen and usually not great until junior year. So to have a large # of sophs leave to me doesn't make sense when viewed in tandem with the IQ piece.

-I agree that the frosh who left was a joke. And to be honest, given what I mentioned above, I simply don't think frosh should leave.

-I think that (by far) the biggest problem is the one you mentioned -- breakdown by team. As you pointed out above and I've preached to seble before, during and after he was making this change, there are loads of pretty comparable guys who could love -- some of whom do, and others don't. So to have 3/4/5 guys leaving from one team, and then have other comparably successful teams with comparably awesome players lose no one, makes no sense.

The system should be more fair and equitable, and that is a huge failure in the current setup.
You should hit seble up with a ticket on this point, because I think it's crucial.

Posted by girt25 on 7/26/2012 3:14:00 PM (view original):
Jeff, I agree with almost everything you wrote here. A couple things to highlight:

-The one point that we don't really agree on is that 19 juniors and 9 sophs seems a little skewed towards too many juniors. I actually think the opposite, and that almost all early entries should be juniors, with just a handful of really elite sophs. Because of IQ restraints, players are never great as freshmen and usually not great until junior year. So to have a large # of sophs leave to me doesn't make sense when viewed in tandem with the IQ piece.

-I agree that the frosh who left was a joke. And to be honest, given what I mentioned above, I simply don't think frosh should leave.

-I think that (by far) the biggest problem is the one you mentioned -- breakdown by team. As you pointed out above and I've preached to seble before, during and after he was making this change, there are loads of pretty comparable guys who could love -- some of whom do, and others don't. So to have 3/4/5 guys leaving from one team, and then have other comparably successful teams with comparably awesome players lose no one, makes no sense.

The system should be more fair and equitable, and that is a huge failure in the current setup.
You should hit seble up with a ticket on this point, because I think it's crucial.

TBH i would like to see mostly juniors leave too. i meant to say the ratio seemed high compared to what we've seen, and compared to what id expect, not high compared to what i think is ideal. seble sort of said that we shouldnt expect to see a significant change in # of early entries, and i thought he suggested the class distribution wouldnt change much either. i was just surprised it was so high (junior, and sophs being that low a % of ees), and was saying that i think that will probably change slightly when the system has been in place a while. i think the # of sophs was artificially lowered by the juniors who would have gone as sophs under the new formulas particularly... so i would expect this to be an unrepresentative sample, in that sense

ive never played in d1 and im just taking a guess that based on what u guys want is to make the game fun not like reall life which i understand i really dnt know anything about d1 in this game but in real life EE have a lot freshman and ketucky lost almost there whole starting lineup and same with unc so if this was suppose to reflect real life then more freshman and sophmores would leave and along with juniors and sometimes have players stay on the big 6 teams but that does not happen too much but again this is just a game and should be made fun and i dnt really know about d1 so could u guys tell me y freshman should never leave and barley have any sophmores?

and this is also not suppose to sound disrespectful to the peolpe that do play in d1 im just wondering the reason behind all of this ur saying?

I think gameplay would benefit from a rather simple adjustment that progressively reduced the odds of a 2nd, then 3rd, then 4th EE.  I am imagining that each player gets assigned by an algorithm a number that expresses his rank for purposes of possible EE.  We'll call that an EE rating.  I dont know if the system works this way, but the simple fix to me would be to run those numbers, then select the first EE, then reduce - lets say by 5% but some number - the EE rating of all players on the #1 guys team, then pick #2, do the same for the guys on #2's team - except if #1 and #2 EE both are from same team than all remaining guys on that team get their EE rating reduce by say 10%....continue to apply sequentially until all EEs have happened.

Even if Duke is losing three, and some other elite team with great players is losing no one, that's stupid. Five is just crazy extreme. It's a pretty easy fix that I think we can all agree would be more fair and better for game play, after the first guy you lose early, it should be progressively harder for each guy to leave.

That way instead of Duke and UCLA losing 3 (or more) and Kansas and Kentucky losing zero, they'll all probably lose 1-2. That's just a better, more fair game. (Those teams are just examples, not specific ones that did or didn't lose guys.)

I was the coach of Duke in the example above.  I expected to lose a couple of players early, but 5 was just a bit too extreme for my tastes.  Duke was my destination school from the first day I started playing this game and it took many, many seasons and perfect timing to finally get there.  I may re-apply just to see if my resume is good enough to get the team back (can't imagine anyone in Tark having a better one at this point in time), but with the way the job process works, it wouldn't surprise me a bit to see someone else get the job ahead of me.

Mason's post above is pretty accurate actually, I probably should have lost 2-3 two seasons ago and then 2-3 this past season and I could have lived with that.  But losing 5 at one pop is quite the kick in the groin.  In fact, that combined with a couple of other things is what caused me to finally drop the team.  Again, I am having second thoughts, but I'm still not 100% convinced I even want to go back at this point.

Tw, I don't think it can be that cut and dry, sometimes you are going to have extreme situations, I don't see any possible way you can make anything so consistent that you never have outliers. And in this instance, Duke's outlier is caused by this being the first season of the new logic. We may never see 5 EEs ever again because this type of situation could be erased from future seasons. Like emy said he probably should have lost a couple of guys the season before, I remember reading the ACC board after the season and emy pretty much saying he was shocked he didn't lose anybody. Not saying that I wouldn't be ****** losing 5 EEs, because I absolutely would, but this situation is not something that I think is going to be typical of the new process. 

West Virginia University is WVU not WaVa, what is next notre dame is nota damn

i like the idea of reducing the changes of players going early. i think seble like, takes the whole universe of players, and lines them out. then, the first guy decides to go, then the second, and so on - eventually players are more likely to say no because so many  better players say yes. 

well, if my understanding is right (i think its close at least), you could simply reduce the random factor for all future recruits for a team by say 25%. so you might have a PF who is #1 in the country, 95% chance to go, and also a SG who is #10 at 85% to go. well when you get to the SG, there is some random % we are rolling a rng against. maybe 8 of the players declared of the first 9, and the SG therefore is like 80% chance to go instead of 85%. well, because 1 team mate left, that 80% could get cut by 25%, which is 60% (i mean to reduce the current value by 25%, not just subtract 25% from the number).

say you also had a #30 C who was maybe 45% to go objectively. say 25 people above him went, which worries him (this step may not even exist honestly, it might always be 40%), and now he has a 40% chance to leave. well, assuming both earlier players left, now id give him a .4 * .75 * .75 = 22.5% chance to leave, which is significantly less. or say just the PF went but not the C, then id have him at .4 * .75 or 30%, which is still significant. 

in this case, even if you had super studs, say your 4th guy is 70% to leave, but the first 3 have left already, well now hed be a .7 * .75 * .75 * .75 which is just a 30% chance of leaving.

all in all, the major impact is that schools are less likely to lose multiple players, especially the ones who were closer to not leaving. i dont think we need the #1 player in the nation to leave less, but the #30 guy should leave less if three players already left. i mean come on, the program is going to be hurt enough from the first three, can't they keep their #30 dude so they at least have someone good?

I think that is more or less how it should work. Don't know the exact numbers, but in theory something like that.

And kmason, I understand Duke was extreme, but I've seen it before with that many guys leaving from one team. Duke was just one example. But I don't want to see a bunch of guys, even if it's 3 and not 5, leave from one team or a few teams, while other great teams with players just as good losing zero. That's hurtful and there's no reason it should be like that.