#SLOE: People's Basketball Sim League

A few people have talked about this very briefly over the years. And yes, we're only in 2012 right now so we still have time. But I really do want to bring up the impending question of what happens after the 2017 draft. I don't think any of us want to quit, so that's out of the picture. The natural idea is to start using generated players. I'll be honest, that has no appeal to me at all. Part of the joy is making teams based on players we've heard of and looked up to.

My suggestion: Starting in 2018 draft, we put players from the 1960 draft. Or 1970, but some year that's far enough back for us to take a while to catch up to where we were so no one is in the league at the same time. That way, there are more players that we've heard of (although a lot of them we won't know) and part of that real-life attribution is still there. I've run this by some people in private, but I wanted to put it up to the whole league to see if anyone likes this idea.

I'm open to other suggestions, but if we can find a way to avoid using fake players, I'm all for it.

I'll offer my thoughts here:

1. The last year FBB has "real life rookies" for is the 2014 draft (the game was released before the 2015 and 2016 drafts had happened, and it would be a herculean effort to manually create the 60+ players from those drafts each year). Creating the 2015 and 2016 draft classes will be tougher than usual because we will have to extrapolate what we've seen so far to produce player ratings - does anyone REALLY feel confident in how much potential a guy like, say, Ben Simmons should have? Essentially, we will be creating fictional players slightly based on current players and slapping current names on them.

2. We're already seeing FBB get a little "out of sync" with trying to guess how good rookies will be. This is because rookie potential is generally based on "best season in the league" and since FBB only has stats through 2014, anybody who busted out in 2015 or 2016 looks awful in the rookie draft because they haven't "hit their peak" yet. The 2013 draft is a prime example of this (the Greek Freak) - it's only going to get worse in 2014 when rookie potential will be based on rookie season before we adjust it and of course in 2015 we'll have to manually create the draft class.

3. FBB gives us only the following options: "Historical Debut" (meaning guys debut in the same year they were really rookies), "Random Debut" (where classes are randomly generated from across NBA history - which means we could get a re-debut of guys we've already seen in the league), or "Fictional Players." Pulling players from the 1960 draft (as an example) has only slightly less "huge workload" attached to it that "manually creating the 2015 and 2016 rookie classes" will have - while these players do exist in FBB's database and do have their "peak stats" already there, we will have to manually create the draft classes by inserting players anyway.

Having set those three items out there, I think the most important questions are:

* Who is going to be tasked with spending the time and effort needed to create the draft classes (note: I really don't have the time and energy to do this even if I wanted to)?

* Is the cost in time and effort involved in creating "real player" draft classes worth the amount of extra enjoyment creating them would bring?

I think each person is going to have to ask themselves about the answer to "is the cost worth it" for them to volunteer to take on this project year after year. Further, the answer to "who is going to be tasked" pretty much has to be "someone volunteers to do it" - I'm always very wary of deciding on ways to spend other peoples' time for them. So I guess what we REALLY are asking here is, "is there anyone in the league that wants to spend the time and energy to do this?" And we have to be mindful that if at any point, every owner in the league answers, "no thanks, that's more work than it's worth" we're going to either have to go to "fictional" or "random debut" by default.

Full disclosure: I personally prefer leagues with fictional players because I think players often get over- or under-valued based on real life name recognition but that's my opinion and JNR's opinion that "Part of the joy is making teams based on players we've heard of and looked up to" is equally valid. Either way, this is a question that we need to get a consensus (probably not unanimous agreement) on pretty quickly here.

Perhaps a dumb question, but could we perhaps just focus on the top-30 rookies in a class and let the rest be randomized/fictionalized? With very few exceptions, I don't know any of those guys anyway. I can dedicate time to working with the draft file, cross-referencing actual stats, peaks, rookie years, etc. But I would prefer to work with 'Real' as opposed to fictional players.

My vote would be to go back to 1960 or 1970 starting as soon as the 2015 draft class.

to help make this decision first of all (because it's going to take an enormous amount of work regardless) I think we should take a poll on whether the league would prefer 'real' or 'fictional' players moving forward.

One option that could be on the table is that I could create a spreadsheet that would generate a draft file for old drafts by scraping stats off the internet and using some sort of equation to derive attributes and potentials. I would then deliver that file to Wig or whoever to tidy it up. It would be a lot of work up front but I also like having real players and it would be fun to put together.

I don't want to sound stupid cause i don't know how the fbb works that well. However, couldn't we get like 5-10 of us to manually create the players from that draft?

kucoach7 wrote:One option that could be on the table is that I could create a spreadsheet that would generate a draft file for old drafts by scraping stats off the internet and using some sort of equation to derive attributes and potentials. I would then deliver that file to Wig or whoever to tidy it up. It would be a lot of work up front but I also like having real players and it would be fun to put together.

I've been thinking about it a bit and (1) I would be excited about doing this and (2) I think I could do it fairly easily. I think it would find an incoming player's rookie year and compare it to historical performance from our league's history and choose something like the attributes of the player for the year most closely related to the incoming player's rookie year (of course we'd have to discount it for guys like Oscar Robertson). Then it would look at the players best year IRL as determined by PER or WS or something and then find the historical performance from the PBSL that best matches it and assign potentials accordingly.

kucoach7 wrote:

kucoach7 wrote:One option that could be on the table is that I could create a spreadsheet that would generate a draft file for old drafts by scraping stats off the internet and using some sort of equation to derive attributes and potentials. I would then deliver that file to Wig or whoever to tidy it up. It would be a lot of work up front but I also like having real players and it would be fun to put together.

I've been thinking about it a bit and (1) I would be excited about doing this and (2) I think I could do it fairly easily. I think it would find an incoming player's rookie year and compare it to historical performance from our league's history and choose something like the attributes of the player for the year most closely related to the incoming player's rookie year (of course we'd have to discount it for guys like Oscar Robertson). Then it would look at the players best year IRL as determined by PER or WS or something and then find the historical performance from the PBSL that best matches it and assign potentials accordingly.

Maybe we could pilot this by having you do a year we've already gone through (say 1996) and compare the to what our actual draft file looked like. Assuming it was close enough, we go with this?

I think that sounds like something that could work!

TheSyndicate wrote:Maybe we could pilot this by having you do a year we've already gone through (say 1996) and compare the to what our actual draft file looked like. Assuming it was close enough, we go with this?

I would be fine running a test on a draft we've done with one caveat. A lot of the data I'll have for the test draft will not be available for many years if we started with 1960 players. ORB and DRB were not separated until 73. No 3P data until 79. No STL or BLK data until 73. TO in 77. So a files for the 60s would rely heavily on extrapolating (most likely just pulling atts ratings from players that looked like them in in PBSL in the other attributes I can assign) for those numbers.

I'm kind of a fan of fictional players, but the idea of randomly inserting old players would be cool, too. If they double up with active players, we could invent a cool time-travel backstory to PBSL.

If not, what kind of draft files do we have to work with. All I need is a spreadsheet of currently active players and currently inactive players from the past, and I can randomly generate a draft class of any number of players using SQL with pretty relative ease.

Id recommend to whoever takes on the task to get with Scott and Wig, as both have extensive experience in building players based off stats. They also have a good feel for how attributes relate to one another and translate into the projected stats. There's also two parts in this. The ratings are only half. You also need to pull bio data for every player too which can be time consuming as well. So splitting this across 2 people might be good too.

Sent from my SM-G935P using Tapatalk

IamQuailman wrote:Id recommend to whoever takes on the task to get with Scott and Wig, as both have extensive experience in building players based off stats. They also have a good feel for how attributes relate to one another and translate into the projected stats. There's also two parts in this. The ratings are only half. You also need to pull bio data for every player too which can be time consuming as well. So splitting this across 2 people might be good too.

Sent from my SM-G935P using Tapatalk

The bio information would also be be compiled automatically by my tool.

You guys with your fancy programming languages. I love it.

kucoach7 wrote:

IamQuailman wrote:Id recommend to whoever takes on the task to get with Scott and Wig, as both have extensive experience in building players based off stats. They also have a good feel for how attributes relate to one another and translate into the projected stats. There's also two parts in this. The ratings are only half. You also need to pull bio data for every player too which can be time consuming as well. So splitting this across 2 people might be good too.

Sent from my SM-G935P using Tapatalk

The bio information would also be be compiled automatically by my tool.

Sent from my SM-G935P using Tapatalk

An alternate idea I had re: fake players...

As part of the offseason, we add a player backstory element where we come up with some additional details about our first round draft pick(s). We'd probably want clearer guidelines (and a word cap or something to limit how much of a chore this would have to be) on what goes into a player's backstory. Anyone who participates would get points, and we could have some kind of voting element where whoever puts together the best backstory gets a free player training for that player.

It seems like the fake players are the easiest route to go (plus they'll help take real life biases out of the game as Wig stated earlier), but this would at least add some flavor to who these randomly generated names would be.

garbageman wrote:An alternate idea I had re: fake players...

As part of the offseason, we add a player backstory element where we come up with some additional details about our first round draft pick(s). We'd probably want clearer guidelines (and a word cap or something to limit how much of a chore this would have to be) on what goes into a player's backstory. Anyone who participates would get points, and we could have some kind of voting element where whoever puts together the best backstory gets a free player training for that player.

It seems like the fake players are the easiest route to go (plus they'll help take real life biases out of the game as Wig stated earlier), but this would at least add some flavor to who these randomly generated names would be.

Thats actually a pretty cool idea. I kinda like it

Sent from my SM-G935P using Tapatalk

I've actually considered quitting the league and doing this full time as garbage man pointed out; essentially I'd manually make draft files for you guys and create a forum or a post for updates weekly updates on those players .. maybe even do media for the league

ballsohard wrote:I've actually considered quitting the league and doing this full time as garbage man pointed out; essentially I'd manually make draft files for you guys and create a forum or a post for updates weekly updates on those players .. maybe even do media for the league

You can't quit until you get a 5th ring.

I'm going to leave this open for discussion until the end of the week; look for a poll next Monday.

Can we extend this to the end of this week (Sunday night)?

Bump...

For reference, FBB provides the following options for automatically creating draft files (does not count "use historical debut" since we're discussing the problem of passing the end of the historical database). All of these have the "pro" of not requiring someone (e.g., kucoach and syndicate) to do the work to manually put a rookie file together.

I took the time to do four samples of each so you can compare "apples to apples" what a draft using each of these methods would look like (these hypothetical draft files were created for the 2014 draft). I did four rather than just one of each to make sure you had multiple examples so a "really good" draft or a "really bad" draft using a given method wouldn't be the sole representative of that method.

1. Historical Players, Random Debut
The game pulls a roster of players from all eras of the NBA so George Mikan and Gary Payton might be in the same draft class.
PROS: Gives us a much larger player pool to pull from before exhausting "real" players.
CONS: Duplication of previous players (e.g., you could get another Gary Payton in the league, while I think it is unlikely, I can't guarantee against multiple players of the same name in the league at the same time). This might require manual intervention (which kind of defeats the purpose of this option versus having someone do the draft file manually).
OTHER NOTES: Lets us pit players from different eras against each other (some like this, some don't).

http://pbsl.ijbl.net/randomdebut1/
http://pbsl.ijbl.net/randomdebut2/
http://pbsl.ijbl.net/randomdebut3/
http://pbsl.ijbl.net/randomdebut4/

2. Fast Break College Basketball Import
We run a FBCB league "underneath" the PBSL and let it export draft classes, which we then import into FBB.
PROS: Draft Classes can be imported from FBCB after the trade deadline, meaning nobody knows what the draft file will look like so tanking to get a particular player is impossible. FBCB was built to create a "farm system" for FBB. Players have more history when they get to the league, including multiple years of college stats if they stayed in college multiple years and awards from High School and College follow players to the pros (see the "Achievements" section on this player page: http://pbsl.ijbl.net/fbcbimport3/players/player42.htm).
CONS: College performance is generally no indicator of pro potential (just current ratings) so extra player history does not mean extra ways to scout players - except my looking at college stats you DO get indicators of the player's foul rate (helpful to avoid drafting a foul-prone guy) and fondness for three-point shooting.
OTHER NOTES: We could run FBCB sims concurrent with the FBB sims or maybe on the off-days. Could also give another Sim Vegas venue, particularly with FBCB's March Madness. Talent level in drafts seems a little higher with this method than with other methods.

http://pbsl.ijbl.net/fbcbimport1/
http://pbsl.ijbl.net/fbcbimport2/
http://pbsl.ijbl.net/fbcbimport3/
http://pbsl.ijbl.net/fbcbimport4/

3. Fictional Draft Classes Generated directly by FBB
We simply allow FBB to generate its own random draft classes (no imports, no historical players).
PROS: Quick and easy.
CONS: Draft classes are accessible from the start of the season, which could lead to tanking. Draft classes tend to be very top-heavy and shallow. Sometimes generates player names with foreign characters that don't display well on this web server (see http://pbsl.ijbl.net/fictional3/players/player613.htm ).
OTHER NOTES: None, really.

http://pbsl.ijbl.net/fictional1/
http://pbsl.ijbl.net/fictional2/
http://pbsl.ijbl.net/fictional3/
http://pbsl.ijbl.net/fictional4/

I'll put up a poll at the start of the playoffs, but this gives everyone a week or so to look at what the various options that don't require manual intervention would actually look like.

Omg I've been wanting to do college sim for a while now.


Sent from my iPhone using Tapatalk

For the FBCB, are we also working with an infinite pool of fake players, but with college info?

Yeah - after looking at the examples, I've been swayed. Love the college idea and those draft files look noice.