Caliber of Players Created Each Year in Amateur Draft are Mediocre
 

I have a fictional league with 32 teams, but set up as closely with the MLB structure as possible so as to not unintentionally create any issues with my outcomes. I like playing fictional league, with fictional teams, and fictional players, but I want it run as closely with the MLB as possible so financials and what not aren't overly skewed.

That said, my league is set up as such: ML, AAA, AA, A, R. I read on the OOTP forums that it is ideal in the amateur draft to create about five rounds per league level. Since I have five league levels, I should have about 25 rounds in the first year draft.

However, I've noticed lately that in the last few drafts, I don't really see any high caliber players in the first few picks of the 1st round. Most of the newly minted players have mediocre or slightly above average potential ratings.

I had read on another OOTP thread that the logic that drives the draft system may have something to do with the number of rounds you have. In other words, the more rounds you have, the higher percentage of high potential players will be created; player quality seems to be proportional to the quantity of rounds you have.

So I'm wondering if I'm not having enough draft rounds in order to create more of these high potential players each year. It's concerning, because as I play through more seasons, the high caliber players that were initially created are filtering out (retiring), and they aren't being replenished by other high caliber players in the amateur drafts. Thus, I'm about to have a league with mostly mediocre players and no super stars.

Has anyone else figured this system out by chance? What's the ideal amount of draft rounds (based on your league structure), and how many rounds of players do you make available? For instance, if you have 25 rounds, do you create enough players for 26...27...30 rounds? Does it make any difference?

I draft 35 rounds & have noticed the quality of amateurs very lacking after about 10 years.

that said, I haven't made it very far because I play out every game & just now staring to see a few players drafted around the 10th turn into good players (so far) so could be wrong.

my bad, I play MLB that is just now turning fictional.

CPU creates all players for the draft.

it's based on # of teams at your ML level... at least for a draft. i think feeders are about quantity? i really don't know on those, if they are different or not. i can promise this about players generated for an amatuer entry-level draft.

increasing rounds will help a bit, but only through increased opportunities for TCR-luck for the most part.

if op is transitioning from real 2018 to fictional, things may look difference after that point. dealing with 2 different animals -- real vs fictionally created. in this case, you simply have to adjust what your eye believes to be ~normal. make sure to autocalculate your modifiers during this transition. after it's ~100% fictional created players, you won't have to auto-calculate after that unless you choose to for other reasons.

there are potentially really bad draft years, fwiw. you can have really bad years for a string of consecutive years, if unlucky. if you don't change modifiers, this will cause the league to naturally ebb and flow with stats-- or probability of resulting stats barring good/bad luck based on how it plays out that particular run through.

also if i recall, the first draft with the last remnants of real amatuers in DB is absolutely pitiful -- even more so to anything mentioned before. that shouldonly be 1 year? may not even happen anymor... if i had to guess it just filled in scrubs with that last year of a few RL amatuers, or i had some really bad luck in past at same point in time of a league. (don't play this way much anymore, so it's a while ago)

i strongly remember the first few amatuer drafts after expending all the RL amatuers from dB as being truely horrible years for talent. it may be purposeful in order to nudge ai into using RL players a bit longer? i know they choose different AI-eval #s for that very reason.

'19 looks different to the eye from '18 too. i think they made it slightly fatter at the bottom fo the distribution curve... a trend in recent years, i think. this is a good thing, once you get your eyes used to the new normal.

the elite players still look very similar, but the ~average guys are probably looking a bit less yellow and a bit more orange on the color code. make the middle a bit more ubiquitious... jsut about luck amongst those that are rated that way to have a good year.

This is a completely fictional league from the beginning (started in year 2000). In the beginning, when I created the league, there were what seemed like ample star player distribution across the league. As I've played through a number of seasons (now 2018), it seems like as star players have retired out of the league, they have not been replenished with other star players through the draft or international free agent signings. In other words, the star player pool (at least according to the ratings) is drying up as I play through more seasons. This has left me with a league of mostly average players (per their ratings).

Does this correct itself over time?

left-handed, I only play one league... no minors... use the reserve roster.

64 teams... 8 divisions with 8 teams each.

I noticed that after about 10 years, my rookies did not perform well... the Rookie of the Year competition did not have many qualified players.that had good stats.... that when I first noticed the lack of talent.

I finally started creating my own draft class using the free agents screen to create new players each season... the draft created too many players that never amounted to anything.

I still create the rookie class each year and assign them randomly to each team... 2 players each year per team... this works fairly well...except the pitchers do much better than the position players... but I noticed not too many rookie performed well this season.

I create a variety of talent... a few stars... some good players...mostly average players and a lot of bench players.

The game does not seem to do a good job of replenishing talent each year.

This is another area where feeder leagues help. Players from feeders have between 3 and 8 seasons of playing games and of course having OOTP games player development improving them every season.

I noticed the same thing about "created" draft classes a little while ago. It's been a problem over the last two versions at least. After 20 seasons or so there is a serious regression to the mean in draft class quality. I did a little experiment. I created two leagues in two different games and simmed 20 seasons. The first game used a standard "created" draft class, the other 120 feeder teams in four leagues.

The created draft class had 1050 players of which less than 10 were five star potential, Roughly 20 were 3.5 to 4.5, the rest 3 or less.

The feeder draft class had 1128 players with 16 five star potential and between 70-80 (don't remember exactly) in the 3.5 5 to 4.5 range.

Another 20 years of sim saw each draft class with slightly less of each but proportions about the same.

This is a problem that probably should bee looked at but using feeders is a way to circumvent it. Of course, the downside of that is now you are using feeders.

Could these be combatted with talent change randomness, dev/aging speed, and say adding another 15-20 rounds to the draft class?

Feeder sim speed is a deal breaker for me I think. As I play with a complete universe.

My concern with feeder leagues is how much that would potentially slow down my sim, as I am playing on a Macbook Pro, not a gaming rig. That's why I've been using a draft class instead. Used to do 20 rounds (21 rounds of created players). But I'm moving to 25 rounds (26 rounds of created players) to see if that helps any with the talent levels that are created.

Left_handed, you say you're playing on a mackbook, I have no idea what it's
capable of next to a desktop PC? Is it just time or space too?

I don't run the whole universe in OOTP but back when I played FM I ran some pretty huge games. With OOTP I run MLB, 5 levels minors, and 4 feeders (2 college, 2 HS) with 120 teams. It's been awhile since I've tried it but IIRC my 6 year old xps spits out a full season in 2-5 minutes. 300 teams, ~13,000 players, 4,000 personnel, and 13,339 games. With this setup I see the same type of results as Mr. Marlin, IE I don't have issues with talent in my league with default settings. I do tick the box where players that retire without ever seeing MLB are deleted. With that ticked undrafted feeder players can drift around a while but they will eventually retire and be gone along with the "also rans" in the minors. Oh yeah, I do use Ghost players too to keep the minors at a more manageable level.

It's easy for me to say this playing on a desktop and not knowing the capabilities of other machines. I do wonder though, what is "slow"? Are we talking an hour to run a season using feeders compared to
minutes without?

Just throwing it out there in case it helps in any way.

As I have a pretty mid range AAA gaming desktop here is my experience
6 cores/12 threads, 32gb ddr 4, nvme 1tb, rx 580 8gb

OOTP 19. I play full default universe. That full independents, international and their minors. Plus the Arizona fall league.Except for Cuba and Australia as they are both bugged and produce MLB players and Cuba can't enter free agency or defect so no point in them. However I do solely play in challenge mode. So I can't change Cuba and Australia settings.

Well in challenge mode you can't say sim an entire year. Its mostly day, week or if their is a special date you can sim to that. But no advance simming time.

With feeder leagues on top of the full universe its this daily sim time that is super slow. As opposed to say non challenge mode where you can sim a month or months etc.

I tried it and it was a no go. Not sure if this is cause challenge mode is more about daily/weekly sim time or what.

I guess this reveals something about what kind of geek I might be, but I'm suddenly picturing Left-handed's macbook as some sort of TARDIS.

if you didn't change number of teams or implement a feeder system or something, then it's normal. the darft pool will ebb and flow quite a bit. if you had a bunch of bad drafts, this really can happen... not often.

if you magically restarted league (or a backup from day one), i'd wager this doesn't happen a 2nd time in a row in same amount of time. longer the time it runs, the more likely you encounter this sort of bad luck.

good luck can happen too. so, 'golden' eras can naturally occur, too. (within reason - proably capable of dipping more than rising, i'd think... easier to "suck" in high proportions than the opposite.)

in case something might be actually wrong:

scouting accuracy is a potential culprit for a false positive of such a problem, fwiw. could turn 100% accuracy on to look at dpeth to ensure you are correct (hindsight, doesn't help know but could start looking for later conclusions). i'd crash the game after doing that, so you don't see the correct data on scouting reports. also, you can customize view to remove name and any other potential identifier. so, it won't lift the veil for you.

i bet it's just some ulnucky years. don't assume anythign at first, if you do this... you don't have a baseline of depth for your draft if you've only seen it through the eyes of inaccurate scouting. everything you've seen so far with scouting accuracy turned on is bloated ratings in the draft.

------

if you don't want the stats to slip due to lesser talent levels, you can auto-calculate the modifiers. this will rais a talent-poor environment's overall numbers to the baselines of the totals in Stats and AI settings.

in times liek these, i occasionally find a few elite guys that put up better numbers... it's not all bad :p also, if there's enough to fill your team, you dominate the others even more than average too.

more often than not, my team slips in these times too... even if i have the best players possible. i think it gets a bit more bunched up in distribution, so there's not as much difference from ~meh and ~elite. less differentiation means they don't dominate as well.

Well just for comparison here is what I have...

dell xps, i7 4th gen 4 core 8 thread, 24 gig ram, nvidia 1050, 1T ssd

I play out all of my games in modern 3d with all other games going on in the background. When I hit "finish day" or whatever it's called it takes 1-2 seconds.

Maybe straying a bit off original topic but thought I'd share.

it's not quite 1:1... as in double leauge due to feeders won't make it 2x as long, but it does take longer, nonetheless.

you'd do better comparing a year-long sim times to determin actual affect. 1 day is volatile.

e.g. i have all my mil end on the same day.... it takes a while for that one day. ties in together when i say feeders won't be 1:1 too, because i don't think it has to re-organize it more than 1/year or just a couple times etc... it's not comparing thousands of players inovolved in a call up on sept 1 etc.. or trading deadline deals.

playing games out is easy work compared to choosing between 2 players let alone the more complicated break-even analysis involving all the player shifting around for one promotion or process of selcting players in a trade etc etc... that stuff isn't happening at feeder level? i think.

i'd bet figuring out one trade potentially takes longer than a game played as far as sim time. if enough players have to be compared and substituted etc etc... wouldn't take long... that includes alll the processing needed to re-organize all the affected depth charts too.

if accuracy didn't alleviate fears, then you need more teams or if feeder systems are 100% scalable, then that would work, once you dial in size needed for the level of talent you want to see in a draft.

I guess I shouldn't underestimate its ability.

It's a 2.7 GHz quad core gen-8 i7 processor (Turbo Boost up to 4.5 GHz) with 16GB of 2133 MHz RAM, a 256 GB SSD. It's capable, I just wasn't sure how much it'd slow down my sim, if any at all. It's plenty fast with the current 5-tier universe I have at the moment. But adding enough feeder leagues to make my draft classes more rich in talent seems like it could potentially bog down the sim speeds. Maybe I'm wrong about that.

Apparently it's also bigger on the inside.
(Okay, I promise, no more Dr. Who references. At least not anymore tonight.)

bwahahah i think you are far beyond diminishing returns with that thing and ootp...

gen 8 and only quad core? you sure? still going to last a long time. cpu is rarely the bottleneck for most game uses. maybe with db-type pure number-crunching work it'd reach its potential (cpu's potential that is).... or video encodeing, but i think the gpus help now on that too.

biggest thing -- don't buy cheap parts. like RAM. not all similar sized RAM has the same latency. it's a much larger factor of speed than mhz. e.g. you go from 2133 to 2400 and it's not a proportional increase in speed... it's probably insignificant, actually. you improve on the latency specs and it's much better.

same with the GPU.. not all of the same model are the same... avoid the no-name brands and products taht seem like a steal.. they likely are not. about 10 years ago i bought a cheaper GPU and saved ~$20. the heat sink was so crappy it couldn't even fully utilized without major heat issues. what's the point of that nonsense? lol...

I know it has been 2 months since the last post in this thread, but I would love for an OOTP dev to help us understand this all.
I have 2 leagues that are almost the exact same setup as you Left Handed in your very first post.
I am not asking for a draft of 10 80 rated players, but a couple 70-80, a bunch of 50's with some 60's sprinkled in would be nice.
32 teams draft 25 rounds each year. After the first round, you are already down to 50 rated players overall potential.
So by round 3, I am drafting guys my scout says will never see the playing field at a ML level.

Would be nice to improve this a little.