What are people’s thoughts on Freshmen and Sophomores with less than 10% odds of leaving (based on Cub’s calculator) still being eligible to go? Fair part of the game or unnecessarily frustrating and a case where pure randomness makes too much of an impact on the following season’s outcome?

Adopting a system like what’s used for recruiting, where there is a sort of actual cut-off below a certain departure probability seems like a potential solution, but could also lead to gamification if it’s too transparent where the line is, as coaches can just manipulate their players’ ratings to stay below the line.

In my head, a more steep tapering off would maintain some element of randomness while reducing the heartbreak. Currently, Sophomores level out at roughly 10% odds of leaving regardless of how low you go on the board and for Juniors it’s roughly 30%. Rather than flattening out, perhaps it’d make sense for the odds to more linearly progress to zero?

Though we’re unlikely to see changes here anytime soon, I’m curious to hear the community’s thoughts.

To start, I don't really like basing discussion off a user-created odds-calculator, no offense to cub. Those odds don't describe the way the system is built, or the actual odds assigned to each player in the particular instance when their number is called, they just describe outcomes we see, and that's not a great basis for discussion. There's a lot we don't actually know about this topic (by game design) but we make assumptions based on common usage.

I do think the system works pretty well, and though there are some potential changes I could get behind, I think it's pretty fair to see some long-shot players leave. It's pretty straight forward to avoid early entries. You don't have to recruit early entry caliber players, and if you do, it's very easy to see where they are at on the big board throughout their career. Now that they've expanded the big board to 150, there is really no excuse at all. If your freshman pops up on the BB at any point, and you want to avoid a possibility of EE, you keep them down under 100 or so. Now is it possible a "staying" freshman starts out higher and you just can't drop him low enough to get him out of danger, and maybe he leaves on a .5% roll (or whatever, I'm not opening cub's calculator) for the last spot in the draft? Sure, it's possible. I don't think removing that risk is smart. The better change, if any change is made, would be to increase the likelihood of some players to leave early, and make that a scoutable preference. Increase one-and-dones, have them come in with higher across-the-board IQs (starting at C to B- instead of F to D+) and let us roll the dice with eyes open.

Posted by shoe3 on 4/15/2023 9:43:00 AM (view original):
To start, I don't really like basing discussion off a user-created odds-calculator, no offense to cub. Those odds don't describe the way the system is built, or the actual odds assigned to each player in the particular instance when their number is called, they just describe outcomes we see, and that's not a great basis for discussion. There's a lot we don't actually know about this topic (by game design) but we make assumptions based on common usage.

I do think the system works pretty well, and though there are some potential changes I could get behind, I think it's pretty fair to see some long-shot players leave. It's pretty straight forward to avoid early entries. You don't have to recruit early entry caliber players, and if you do, it's very easy to see where they are at on the big board throughout their career. Now that they've expanded the big board to 150, there is really no excuse at all. If your freshman pops up on the BB at any point, and you want to avoid a possibility of EE, you keep them down under 100 or so. Now is it possible a "staying" freshman starts out higher and you just can't drop him low enough to get him out of danger, and maybe he leaves on a .5% roll (or whatever, I'm not opening cub's calculator) for the last spot in the draft? Sure, it's possible. I don't think removing that risk is smart. The better change, if any change is made, would be to increase the likelihood of some players to leave early, and make that a scoutable preference. Increase one-and-dones, have them come in with higher across-the-board IQs (starting at C to B- instead of F to D+) and let us roll the dice with eyes open.

Thanks for sharing shoe. I’d be a big supporter of something like a pro interest preference being introduced in recruiting.

I think increasing odds of departures across the board would also be a welcome change, as it would reduce the super extreme results of having players sub-100 on the big board leave early.

Both of these would limit chaotic randomness in my opinion while the first would introduce another strategic element to recruiting.

Something else that’s been long discussed is introducing some new recruits during the late period. I am ambivalent on this one because I’m not sure whether it would increase or decrease parity between high and low baseline schools (probably decrease),.

"To start, I don't really like basing discussion off a user-created odds-calculator, no offense to cub."

This is a bad statement in the sense that the calculator is entirely correct, at least regarding Trail's point. Odds to leave do form asymptotes at 10/30%, that is just fact. Based on the data, dropping from 40 to 60 obviously lowers the chance of leaving by much more than dropping 80 to 100.

What are the limitations?
Some believe that there is another layer of big board movement we can't see (so maybe a guy at 76 is bumped up to 70 because of post-season success or whatever), so obviously the calculator can't figure that out. But the calculator trained on a lot of data (1-2% CI in most situations). That does not apply in this discussion.

Posted by cubcub113 on 4/15/2023 2:01:00 PM (view original):
"To start, I don't really like basing discussion off a user-created odds-calculator, no offense to cub."

This is a bad statement in the sense that the calculator is entirely correct, at least regarding Trail's point. Odds to leave do form asymptotes at 10/30%, that is just fact. Based on the data, dropping from 40 to 60 obviously lowers the chance of leaving by much more than dropping 80 to 100.

What are the limitations?
Some believe that there is another layer of big board movement we can't see (so maybe a guy at 76 is bumped up to 70 because of post-season success or whatever), so obviously the calculator can't figure that out. But the calculator trained on a lot of data (1-2% CI in most situations). That does not apply in this discussion.

"Bad statement" lol. Thanks for your evaluation, cub. I'll be sure to file that away.

I'm sure you have crunched copious amounts of data in ways that meant a lot to you. It's still outcome oriented, and is not focused on the system process or reasoning, or gameplay goals, and that is why I don't like basing this discussion on it, whether you think my preference is "bad" or not.

i would be ok with them bringing back an NBA preference, i kinda enjoyed it the first time around. i feel like they should do it differently if they do it this time though. its kinda tricky, make it too modest and people will be annoyed its useless, make it too strong and then you have folks trying real hard for those guys and furious if they leave anyway...

i would like to see it be more of a 100/0 proposition maker. maybe a guy with high propensity to leave goes 100% if he's going to be a first round pick, and a low propensity guy is 100% to stay if he's a 2nd round pick. it could be more complex than that, maybe the high guy leaves as a jr in either round too. sometimes those hard edges cause their own issues (drat, he left first pick of the second round!), but i feel like knowing what to expect would be kinda nice. just making it more or less likely up and down the whole spectrum, i don't know, i think its going to be hard for people to not be frustrated by that.

I totally agree Trail. Long shot EEs add absolutely nothing to the game from a gameplay perspective. It's randomness for the sake of randomness.

Why does the game reward lucky coaches and punish unlucky coaches who have nearly zero impact on the results.

All guys who are top 20ish should leave 100% of the time and those below top 100 should have a 0% chance of leaving. Seems simple enough

I had no idea the NBA preference was in the game previously. How did it work? I like that idea if its done correctly.

Also Shoe... you are playing a simulation game. Why are you disregarding a monte carlo sim and outcome-oriented math? How do you think WIS was built? I don't see why we can't use that as one more thing to analyze when discussing this topic. The underlying data seems fine.

Posted by Basketts on 4/18/2023 7:48:00 AM (view original):
I had no idea the NBA preference was in the game previously. How did it work? I like that idea if its done correctly.

Also Shoe... you are playing a simulation game. Why are you disregarding a monte carlo sim and outcome-oriented math? How do you think WIS was built? I don't see why we can't use that as one more thing to analyze when discussing this topic. The underlying data seems fine.

Because it makes discussion wonky and produces "forum facts" which get spread around. The developers have been intentionally tight-lipped about how early entries are actually determined. If you go to CS and complain about your soph who left early from 104 on the board only having a .5% chance, they will probably tell you that is not an accurate figure - ie, that wasn't the probability of leaving assigned to the player when his number was called - because that's not how the system works. The figure describes probability in a broad sense, but is not looking at what has actually happened, because it doesn't know the system.I don't like having a discussion on potential changes to the system based on ignorance.

What I do know is - like I said - it is very straightforward to deal with early entries. You can limit the number you recruit, or avoid them altogether, you can see where they are on the big board at all times, and you can manage their development to keep them where you want them, if that's important to you. This is all very good approximation of dealing with the tightrope of recruiting elite talent in college athletics. It's never been a black or white binary, where all the top leaves every year. There should be a possibility of top guys coming back, that's good for the game, as long as it can't be manipulated beyond just doing what you can with development (which is a known risk/reward choice) and crossing your fingers.

Well the developers probably would because .5% isn't the chance "when his number is called." It's the likelihood that a player will go in the grand scheme of 10,000 simulations. If there's still an open draft spot when they call his number at 104, then the odds are much higher he gets drafted in that moment of time. So I guess it is dangerous to have precedents set on the forum based on viewing that model wrong. It should be used as exactly you say, a broad sense look at probability. The big board is there for us to see. We have the draft results. The only margin of error is whether movement happens between the last big board and when the draft occurs. And I don't think that invisible movement is drastic enough to throw the entire model out the window as inaccurate.

That's not to say anything from your second paragraph is wrong. I think those are great conversations and good practices to have when it comes to EEs. I tend to use both best practices and the monte carlo simulation to plan ahead. It's worked for me at least.

Posted by Basketts on 4/18/2023 11:30:00 AM (view original):
Well the developers probably would because .5% isn't the chance "when his number is called." It's the likelihood that a player will go in the grand scheme of 10,000 simulations. If there's still an open draft spot when they call his number at 104, then the odds are much higher he gets drafted in that moment of time. So I guess it is dangerous to have precedents set on the forum based on viewing that model wrong. It should be used as exactly you say, a broad sense look at probability. The big board is there for us to see. We have the draft results. The only margin of error is whether movement happens between the last big board and when the draft occurs. And I don't think that invisible movement is drastic enough to throw the entire model out the window as inaccurate.

That's not to say anything from your second paragraph is wrong. I think those are great conversations and good practices to have when it comes to EEs. I tend to use both best practices and the monte carlo simulation to plan ahead. It's worked for me at least.

I don't think we disagree much at all. I'll put it this way, I think it's a great use of Monte Carlo sim for developers to check to see if the designed system is working as intended, to see if the outcomes are lining up with overall expected goals. As a user, I have less use for it when I don't know exactly how the system is set up. I am also not concerned with meta for my own gameplay, and for folks who are, I get that it can be useful for their purposes to that end. But that's an entirely different discussion.

Yeah--just to be clear to everyone reading this who is newer to the EE world. The reason you push your player down from 90 to 110 is NOT to decrease his chance of leaving if 60 players aren't drafted ahead of him. It is to ensure 60 players listed ahead of him leave. A player will never go EE and not be drafted.

Posted by shoe3 on 4/18/2023 8:22:00 AM (view original):

Posted by Basketts on 4/18/2023 7:48:00 AM (view original):
I had no idea the NBA preference was in the game previously. How did it work? I like that idea if its done correctly.

Also Shoe... you are playing a simulation game. Why are you disregarding a monte carlo sim and outcome-oriented math? How do you think WIS was built? I don't see why we can't use that as one more thing to analyze when discussing this topic. The underlying data seems fine.

Because it makes discussion wonky and produces "forum facts" which get spread around. The developers have been intentionally tight-lipped about how early entries are actually determined. If you go to CS and complain about your soph who left early from 104 on the board only having a .5% chance, they will probably tell you that is not an accurate figure - ie, that wasn't the probability of leaving assigned to the player when his number was called - because that's not how the system works. The figure describes probability in a broad sense, but is not looking at what has actually happened, because it doesn't know the system.I don't like having a discussion on potential changes to the system based on ignorance.

What I do know is - like I said - it is very straightforward to deal with early entries. You can limit the number you recruit, or avoid them altogether, you can see where they are on the big board at all times, and you can manage their development to keep them where you want them, if that's important to you. This is all very good approximation of dealing with the tightrope of recruiting elite talent in college athletics. It's never been a black or white binary, where all the top leaves every year. There should be a possibility of top guys coming back, that's good for the game, as long as it can't be manipulated beyond just doing what you can with development (which is a known risk/reward choice) and crossing your fingers.

JR at 104 actually has about a 29% chance of leaving early if it gets to him, according to the data (likely end of 2nd round pick). This is super high, considering a Junior at 45 (likely end of 1st round pick) only has a 55% chance of leaving early. Very counterintuitive to a lot of new coaches.

Posted by Benis on 4/17/2023 10:12:00 AM (view original):
I totally agree Trail. Long shot EEs add absolutely nothing to the game from a gameplay perspective. It's randomness for the sake of randomness.

Why does the game reward lucky coaches and punish unlucky coaches who have nearly zero impact on the results.

All guys who are top 20ish should leave 100% of the time and those below top 100 should have a 0% chance of leaving. Seems simple enough

I think this is the essence of my position as well. In the best games, skill dominates due to a long learning curve, and luck creates opportunities on the margins. For the most part, I think HD gets high marks in this regard. There are dice rolls in recruiting, but you control which battles you enter. Outcomes of games are uncertain, but the variance in team performance is realistic and predictable. In most places, you can prepare for or hedge against randomness.

With EEs, there's certainly an argument to make that you can hedge against randomness by not signing NBA-caliber players, as Shoe has laid out. However, for any coach who's trying to build the best team in their world, this is not a viable option. We knowingly manage our risk through holding back player development to keep our fringe players from ascending the big board. But every season, due to the sheer number of dice rolls that occur, a few low probability events will occur, some benefiting coaches and some hurting coaches.

I don't necessarily agree with Shoe that there is information the community is not aware of that sways EE probabilities. But even if there is, the fact that it's unknowable means it's a missed opportunity to add another strategic element to the game and increase the learning curve, which would make the game even more fun and engaging.

As it stands, EE dice rolls can tremendously sway the balance of power among the top teams. Because they occur following the early signing period, there is generally minimal opportunity for a coach to react and adjust their recruiting strategy to compensate for a highly improbable event.

I admit that I'm particularly motivated about this because I generally take on a couple walk-ons, which makes my teams more vulnerable to these chance occurrences. This is a strategic decision to gain an advantage in recruiting, and I'd agree with anyone who says, "If you want to be less vulnerable to this, start running deeper rosters." Nonetheless, I believe a bit more strategic influence and less arbitrary randomness here would be a benefit to the game, increasing the learning curve and decreasing user frustration simultaneously.