Future Drafts (1,598 views)

[eric]

Another reason we've talked about 5.0 is what to do when draft classes run out. Here are some other possibilities:

1. Parody / novelty classes. Pro wrestlers, WNBA, GMs, etc.
Pro: you could draft sim Bruns and immediately have the entire roster fail every drug test, instant tank
Con: limited pool to draw from, probably only five classes worth doing tops

2. 60s/70s classes
Pro: I noticed in doing draft analysis that we've barely had any of those classes, granted some players would have been used in the creation draft but we're missing 15 years so there's still plenty left to use.
Con: by definition limited pool to draw from, would be missing most of the big names

3. More college/Euro players
Pro: enormous pool, there are 350 D1 schools so one year alone gives us 50+ draft classes, is how IRL draft works
Con: once we get down to obscure schools like Texas Agricultural and Mechanical we might as well use random names

4. Just use random names / builds
Pro: unlimited pool, possibility of making draft master position algorithmic
Con: randomly generated players harder to connect with, harder to tell which players will be intentionally busted

5. Re-use previous classes
Pro: names we all know, will by definition always have enough players because there are only 20ish classes active at any given time
Con: if we use the same builds it reduces (but does not eliminate) the mystery, if we use different builds GM expectations won't be met

.

My feeling is that option (5) is the best way. What are everyone else's thoughts? Are there other options you'd like to propose?

[Majic]

why not do a combination of all them?

just do a random pick for the year of which one we use and go from there to build out the class

[eric]

Oct 19, 2016 13:30:08 GMT -6 Majic said:

why not do a combination of all them?

just do a random pick for the year of which one we use and go from there to build out the class

That's a good thought but I for one would rather set up some kind of circular apparatus so that we knew in advance what type of draft each year would be.

[Lazy Pete]

I wasn't part of 2.0 but I think they recycled classes and the reception was lukewarm. Maybe someone from late 2.0 can add their thoughts

[Cinco de Wardo]

option 2 imo.

[Heebs]

Option 2 (including players used in creation draft). Then Option 5.

[20s Navidad]

lol at trying to sneak WNBA into Option 1

It's not happening

[Delap]

I like a combo of 1, 2, and 5. A draft of non-NBA athletes and pro wrestlers would be great. Lots of good classes from the 70s to use. Then just start over.

I want Andre the Giant and Rey Mysterio running a PnR for the Sonics.

[Odin]

2>1>4>3>5

[andrewluck]

Option 1, IMO.

Andrew Avery has probably not made an appearance but I bet he'd beast at PG or SG.

Also, The Rock at SF doing Lebron-like things.

[Odin]

Oct 19, 2016 16:06:56 GMT -6 andrewluck said:

Option 1, IMO.

Andrew Avery has probably not made an appearance but I bet he'd beast at PG or SG.

Also, The Rock at SF doing Lebron-like things.

We had/have William Avery. He's an asshole.

[KwYawnza]

Yeah Option #2 would be my choice.

Reusing old classes and having a player X II was pretty bleh in 2.0

[IanBoyd]

Option 2 mixed in with some more fictional/high school/euros imo.

[Heebs]

As the new draftmaster I'm bumping this for further discussion.

We basically need 60 players per class. 60 novelty players per class seems like overkill. WNBA players were nixed by the commish.

I think there are three good options.

First, we can use the 60s/70s classes with or without the creation draft players. My preference is with creation players, which leads me to my next good option.

Second, recycle old classes. We'll keep the list of classes in the hopper at 20, adding a class each year. So we'll start with the first 20 draft years (the classes from 3001-3020) and randomly select one and replace it with the next available class (3021 in this example). That will provide some novelty to when we get the class, but we avoid having duplicate players. We can

Third, we can use undrafted college players. In the abstract this seems like the best choice, but I've not come up with a way to select a subset of players without overlap issues. The best option I've got so far is to use randomize the year and limit it to freshmen.

Thoughts appreciated.

[MistleTacoe]

Option 2 with a little Option 1 and Option 4 sprinkled in there.

[SPL]

Jan 18, 2017 8:29:34 GMT -6 Heebs said:

As the new draftmaster I'm bumping this for further discussion.

We basically need 60 players per class. 60 novelty players per class seems like overkill. WNBA players were nixed by the commish.

I think there are three good options.

First, we can use the 60s/70s classes with or without the creation draft players. My preference is with creation players, which leads me to my next good option.

Second, recycle old classes. We'll keep the list of classes in the hopper at 20, adding a class each year. So we'll start with the first 20 draft years (the classes from 3001-3020) and randomly select one and replace it with the next available class (3021 in this example). That will provide some novelty to when we get the class, but we avoid having duplicate players. We can

Third, we can use undrafted college players. In the abstract this seems like the best choice, but I've not come up with a way to select a subset of players without overlap issues. The best option I've got so far is to use randomize the year and limit it to freshmen.

Thoughts appreciated.

I would vote for option 2. Can we not just say it is the kid of the player drafted in that class and make them all jr.

[20s Navidad]

Heebs can you expound upon your third option? I'm not sure I'm fully grasping what you mean and what the process would be.

[SPL]

Jan 18, 2017 9:03:53 GMT -6 20s Navidad said:

Heebs can you expound upon your third option? I'm not sure I'm fully grasping what you mean and what the process would be.

3rd option would be players that played college basketball that season that were never drafted into the NBA. At least that is how I read it.

[Face-in]

Options 1 and 3 please

[20s Navidad]

Jan 18, 2017 9:06:45 GMT -6 SPL said:

Jan 18, 2017 9:03:53 GMT -6 20s Navidad said:

Heebs can you expound upon your third option? I'm not sure I'm fully grasping what you mean and what the process would be.

3rd option would be players that played college basketball that season that were never drafted into the NBA. At least that is how I read it.

I get that. What I don't fully get is the process he described which is why I want more details. 

[Delap]

I'd say start with #1 (the classes are there, why not use them?) and then go to #2.

[Heebs]

Jan 18, 2017 9:03:53 GMT -6 20s Navidad said:

Heebs can you expound upon your third option? I'm not sure I'm fully grasping what you mean and what the process would be.

First, the issue with undrafted college players. The issue is that undrafted players show up in multiple years. So if we were selecting from college players in 2000, there would be a lot of overlap on rosters for 1998, 1999, 2001, and 2002. That overlap causes two problems. First, the draft master has discretion to mine the big names and leave the scraps for later draft masters to comb through. Second, it's a pain in the ass to keep track of which players have been used - the big names will be easy but most players will never be big names in the league. I don't want to (and don't want to create a system in which my successor will have to) search through old draft classes to make sure each player they're picking hasn't been created before.

The proposed solution would be to randomly select the year and say only freshmen from that year can be used. The draft master can draw from their entire college career in creating their grade set and/or profile, but they have to have been a freshman in the random year to be part of that class.

I considered using four year periods, but that doesn't solve the overlap.

We could create a thread with players that have been used and do a quick cntl+f - that'd be easier than clicking through all of the old draft threads and cntl+f'ing. I guess that's a solution to the second problem - but we've still got the first problem.

[FectaDEEZ]

Jan 18, 2017 9:12:35 GMT -6 Delap said:

I'd say start with #1 (the classes are there, why not use them?) and then go to #2.

These are my thoughts exactly. Not a fan of re-using classes.

[Face-in]

Jan 18, 2017 9:23:56 GMT -6 Heebs said:

Jan 18, 2017 9:03:53 GMT -6 20s Navidad said:

Heebs can you expound upon your third option? I'm not sure I'm fully grasping what you mean and what the process would be.

First, the issue with undrafted college players. The issue is that undrafted players show up in multiple years. So if we were selecting from college players in 2000, there would be a lot of overlap on rosters for 1998, 1999, 2001, and 2002. That overlap causes two problems. First, the draft master has discretion to mine the big names and leave the scraps for later draft masters to comb through. Second, it's a pain in the ass to keep track of which players have been used - the big names will be easy but most players will never be big names in the league. I don't want to (and don't want to create a system in which my successor will have to) search through old draft classes to make sure each player they're picking hasn't been created before.

The proposed solution would be to randomly select the year and say only freshmen from that year can be used. The draft master can draw from their entire college career in creating their grade set and/or profile, but they have to have been a freshman in the random year to be part of that class.

I considered using four year periods, but that doesn't solve the overlap.

We could create a thread with players that have been used and do a quick cntl+f - that'd be easier than clicking through all of the old draft threads and cntl+f'ing. I guess that's a solution to the second problem - but we've still got the first problem.

You could pick from the rivals 150.  They go back to 2003.

n.rivals.com/prospect_rankings/rivals150/2003

[SPL]

Jan 18, 2017 9:28:49 GMT -6 Face-in said:

Jan 18, 2017 9:23:56 GMT -6 Heebs said:

First, the issue with undrafted college players. The issue is that undrafted players show up in multiple years. So if we were selecting from college players in 2000, there would be a lot of overlap on rosters for 1998, 1999, 2001, and 2002. That overlap causes two problems. First, the draft master has discretion to mine the big names and leave the scraps for later draft masters to comb through. Second, it's a pain in the ass to keep track of which players have been used - the big names will be easy but most players will never be big names in the league. I don't want to (and don't want to create a system in which my successor will have to) search through old draft classes to make sure each player they're picking hasn't been created before.

The proposed solution would be to randomly select the year and say only freshmen from that year can be used. The draft master can draw from their entire college career in creating their grade set and/or profile, but they have to have been a freshman in the random year to be part of that class.

I considered using four year periods, but that doesn't solve the overlap.

We could create a thread with players that have been used and do a quick cntl+f - that'd be easier than clicking through all of the old draft threads and cntl+f'ing. I guess that's a solution to the second problem - but we've still got the first problem.

You could pick from the rivals 150.  They go back to 2003.

n.rivals.com/prospect_rankings/rivals150/2003

I would be a big fan of this

[Heebs]

That would definitely be something I use if a class comes up in that era under the "freshman only" plan.

[20s Navidad]

What I am thinking right now is we continue with 4.0 until the final player from the currently posted draft clssses retires (so the last player from the 3047 and earlier drafts). I like this because it doesn't give a hard deadline of a last season so there would still be some suspense and intrigue waiting to see retirings, but it would still give a rough timeline so people can act accordingly.

We would then start 5.0 with a new creation draft and throw all the other classes back in the randomizer. Thoughts?

So whatever solution we ended up going with for drafts after 3047 would just be filler classes until the real ones wash out.

[Lazy Pete]

I like the idea of playing out the string w/ the 47 class, although I don't know what that would mean for the trade market in those waning years. Also, what would we do for the creation draft? Would we block out a decade of real classes again like we did w/ 4.0?

[20s Navidad]

Jan 18, 2017 10:19:14 GMT -6 Lazy Pete said:

I like the idea of playing out the string w/ the 47 class, although I don't know what that would mean for the trade market in those waning years. Also, what would we do for the creation draft? Would we block out a decade of real classes again like we did w/ 4.0?

Would choose a specific year of real life NBA and use that as the player pool. Would have to look at the draft class distribution to determine how many classes it would completely kill. 

[Odin]

i think it would be neat to do a couple of things:

1. make a large list of players and pull 70 random players from it each season. tmac, barkley, and lanier in the same class. we could do a variant of it after the 47 class to see if it's something that we might want to do in 5.0. I've seen other leagues do it and it seemed to work well.

2. start 5.0 with all fake players so we have more draft classes to work with