People say the percentages will even out, but I'm 2 for 20 on ties 40% to 64% with most being me leading. Ironically the 2 that I have won were on ones that I did NOT lead. It really feels to me like the dice roller is backwards. My prestige was B- to B in most of these so your best bet is avoid A or greater schools all together.

the rolls have been weird for me, too. with a bunch of folks playing, it has to be weird for a bunch of people, but i do wonder... mostly because i am reluctant to rule out anything without a really good reason, not because i think its actually off or because my own experience has been that far from normal (it hasn't been). but perhaps we should get a collective of like 10-20 folks together who religiously record every dice roll on every team for a year and then check the math?

Probably a few reasons why this is a common theme with people. First - you just don't hear about the success stories of recruiting. When you feel entitled to a recruit and you win with the higher percentage, it was supposed to happen so why come on the forums and mention it? But more importantly, each person has a very small amount of rolls every season, even if you tend to get into multiple battles. Small sample size in statistics will always have high variation that could seem skewed from the actual percentages. This is why when you go to the roulette table in Vegas they will almost always display the last 10 or so spins. They know people can't help but notice patterns and when they see that 8 out of the last 10 times hit red, they bet heavy on black because they think its due. But its just a way to manipulate. The probability to hit black was always the same.

It's possible that there is a coding error, but I find it hard to believe that a game built on simulation would get the most basic percentage roll wrong.

Posted by Basketts on 9/22/2021 12:11:00 PM (view original):
Probably a few reasons why this is a common theme with people. First - you just don't hear about the success stories of recruiting. When you feel entitled to a recruit and you win with the higher percentage, it was supposed to happen so why come on the forums and mention it? But more importantly, each person has a very small amount of rolls every season, even if you tend to get into multiple battles. Small sample size in statistics will always have high variation that could seem skewed from the actual percentages. This is why when you go to the roulette table in Vegas they will almost always display the last 10 or so spins. They know people can't help but notice patterns and when they see that 8 out of the last 10 times hit red, they bet heavy on black because they think its due. But its just a way to manipulate. The probability to hit black was always the same.

It's possible that there is a coding error, but I find it hard to believe that a game built on simulation would get the most basic percentage roll wrong.

Exactly. Most VH-H battles are in the neighborhood of 66-34. If the percentages are correct, you’ll lose about 1/3 of those given enough samples. You’ll likely remember each of those losses a whole lot better than every win. And the recall discrepancy compounds on the forum and on CC, because folks who naturally complain about the bad beats see confirmation for their biases in reports of people who are all doing the same kinds of things. So unless you are *religiously* collecting the data over a lot of battles in a lot of seasons - and the folks I’ve seen who have done that, who I trust to get it mostly right, tend to indicate the system works fine - all we have is an anecdote, which human nature tells us will be skewed a certain way.

agree 100%. its far more likely human bias and the statistical realities (someone has to get 1 in 1000 unlucky if there are 1000 of us! more or less) are the cause, than the actual RNG being off. i am just an extremely skeptical person - which means while i am generally skeptical of all conspiracy theories or far fetched ideas - i am also reluctant to rule them out without a really good reason. plus any true skeptic is skeptical of themselves above all else. and my initial inclination was, its definitely human bias. but now that i've internalized that, and i don't really have anything resembling proof, i can't help but wonder.

i will say this though. RNGs themselves are definitely far from perfect, especially if you are running on an older langauge/platform. developer usage of RNGs can also be fairly suspect in general. now, you might say, the sim engine uses RNG like crazy - it can't be a problem. but that's not true. for example, a lot of older RNGs, and even some modern ones, you have to explicitly seed, and often that seed is the same every time. from there the RNG generates the numbers based off that seed... sometimes in 100% deterministic fashion. well, always in 100% deterministic fashion, technically speaking, but from a laymen standpoint, i wouldn't worry about that. anyway, what is important is, seeded RNGs tend to work reasonably well *over the long haul*, but are fairly predictable on the early numbers. like over the course of 500 flips in a sim engine run or something, it evens out. but if you do 500 runs of quantity 1, the results can be EXTREMELY far from random. so the act of flipping 500 coins for 500 independent battles, is fundamentally very different than flipping 500 coins in a series for 1 game sim - and that's really a general statement, the seed thing is just an example.

i suspect seble got it right, or close enough. but he'd be like the millionth guy to get it wrong, if he did. so i sort of think it would be worth... getting a little consortium together... to test it. the recruiting info shared would all be after the fact and it seems like no concerns collusion wise or any of that.

Super appreciate this conversation and all of the thoughtful contributions. For what it's worth, before posting, I did think of many of the things mentioned here: these are all anecdotal without date/research, we remember the losses more than the wins, a small sample size can not be trusted when randomness is a factor, and so on. I have won 1 or 2 recruiting battles during this bad luck stretch that I thought I should have prolly lost, though who knows what the other team offered.

As for avoiding battles with A prestige squads, going big on a guy rated around 50 as a guard (not overall) is exactly trying to avoid battling an A. Doesn't matter when they swoop in later, maybe cause they've lost out elsewhere, and stroll off with my hidden treasure.

The one wild card I'm left wondering about--that no one has mentioned here--is baseline prestige and how much this plays a role. Am inclined to believe that if I took a big time conference team I'd have an easier shot at recruiting big time players. This is, of course, how real life works. But real life makes for terrible gaming (and storytelling).

Posted by gillispie on 9/22/2021 11:36:00 AM (view original):
the rolls have been weird for me, too. with a bunch of folks playing, it has to be weird for a bunch of people, but i do wonder... mostly because i am reluctant to rule out anything without a really good reason, not because i think its actually off or because my own experience has been that far from normal (it hasn't been). but perhaps we should get a collective of like 10-20 folks together who religiously record every dice roll on every team for a year and then check the math?

I like this idea. If we can start a forum I’ll post all of my dice rolls in it… if a bunch of others did the same we can have some real data to work with

Posted by rls1 on 9/23/2021 10:56:00 AM (view original):
Super appreciate this conversation and all of the thoughtful contributions. For what it's worth, before posting, I did think of many of the things mentioned here: these are all anecdotal without date/research, we remember the losses more than the wins, a small sample size can not be trusted when randomness is a factor, and so on. I have won 1 or 2 recruiting battles during this bad luck stretch that I thought I should have prolly lost, though who knows what the other team offered.

As for avoiding battles with A prestige squads, going big on a guy rated around 50 as a guard (not overall) is exactly trying to avoid battling an A. Doesn't matter when they swoop in later, maybe cause they've lost out elsewhere, and stroll off with my hidden treasure.

The one wild card I'm left wondering about--that no one has mentioned here--is baseline prestige and how much this plays a role. Am inclined to believe that if I took a big time conference team I'd have an easier shot at recruiting big time players. This is, of course, how real life works. But real life makes for terrible gaming (and storytelling).

Baseline doesn’t have a direct effect on recruiting. The recruit is only looking at what the programs functional prestige actually is. Baseline just acts like a tether, making it easier for high baseline teams to maintain B and above level prestige than the low and mid-majors. So it matters to the extent that a high baseline team can have a mediocre stretch and still maintain a B+ functional prestige, while a B level team that misses the NT a few seasons in a row will dip down below B. But the recruit only cares about the functional prestige.

I keep track of all my rolls. I'll participate.

And for stormfury, when you get to 6 roll wins and 44 roll losses, it'll turn for you. It did for me! Hang tight brother. It'll change for you

i don't know about an open forum thread. seems we want to find a closed groups of relatively committed folks so we aren't getting semi-random, biased-influenced posts from folks. i think it would have to be people committed to sharing the data every season for the involved teams, for like a year or something, perhaps even with a commissioner who is checking as worlds recruit to try to make sure we get everyone. if there's much potential for bias in the results, seems like they will be meaningless.

i would like to see the end results shared to anyone interested, but i doubt anyone cares outside of 'does it line up or not'.

also thinking about the info sharing side a bit more... there is some gray area in there, like if i posted a battle % early in RS1, someone could in theory use that info against my opponent or something. the safest is posting at the end of RS2, but you risk people missing it - that is a big dead zone when folks tend to take off for a while. i think we could figure out how to make it work, not exactly sure how, but it does seem like there have to be some guardrails or something, and having a bunch of folks randomly posting on an open forum seems impossible to manage appropriately, i guess.

Posted by gillispie on 9/24/2021 4:54:00 PM (view original):
i don't know about an open forum thread. seems we want to find a closed groups of relatively committed folks so we aren't getting semi-random, biased-influenced posts from folks. i think it would have to be people committed to sharing the data every season for the involved teams, for like a year or something, perhaps even with a commissioner who is checking as worlds recruit to try to make sure we get everyone. if there's much potential for bias in the results, seems like they will be meaningless.

i would like to see the end results shared to anyone interested, but i doubt anyone cares outside of 'does it line up or not'.

also thinking about the info sharing side a bit more... there is some gray area in there, like if i posted a battle % early in RS1, someone could in theory use that info against my opponent or something. the safest is posting at the end of RS2, but you risk people missing it - that is a big dead zone when folks tend to take off for a while. i think we could figure out how to make it work, not exactly sure how, but it does seem like there have to be some guardrails or something, and having a bunch of folks randomly posting on an open forum seems impossible to manage appropriately, i guess.

That’s my fear. With a public forum you could also get people who are frustrated with a loss so they post it but neglect to post their wins