Posted by Benis on 10/8/2018 12:51:00 PM (view original):

Posted by Benis on 10/3/2018 2:01:00 PM (view original):
Probability, for the lack of a better word, is good. Probability is right, probability works. Probability cuts through and captures the essence of the evolutionary spirit. Probability, in all its forms. Probability for life, for money, for love, knowledge has marked the upward surge of mankind.

And probability, mark my words, will not only save Hoops Dynasty, but that other malfunctioning corporation called WhatifSports.

A little disappointed no one got this movie reference.

'Wall Street'

Posted by dacamel14 on 10/8/2018 3:32:00 PM (view original):

Posted by Benis on 10/8/2018 12:51:00 PM (view original):

Posted by Benis on 10/3/2018 2:01:00 PM (view original):
Probability, for the lack of a better word, is good. Probability is right, probability works. Probability cuts through and captures the essence of the evolutionary spirit. Probability, in all its forms. Probability for life, for money, for love, knowledge has marked the upward surge of mankind.

And probability, mark my words, will not only save Hoops Dynasty, but that other malfunctioning corporation called WhatifSports.

A little disappointed no one got this movie reference.

'Wall Street'

Ding ding! Not the most famous quote ever but I think it fit in nicely here.

Posted by Benis on 10/3/2018 2:01:00 PM (view original):
Probability, for the lack of a better word, is good. Probability is right, probability works. Probability cuts through and captures the essence of the evolutionary spirit. Probability, in all its forms. Probability for life, for money, for love, knowledge has marked the upward surge of mankind.

And probability, mark my words, will not only save Hoops Dynasty, but that other malfunctioning corporation called WhatifSports.

Is part of that a quote from GWF Hegel?

Posted by mullycj on 10/8/2018 10:14:00 AM (view original):
I don't know why you feel the need to quash anyone's man crush on here topdogg!!

I know! I didn't mean to interfere. Maybe I was getting a little jealous because no one loves me that way!

(I never did get a return sitemail, shoe!)

Posted by topdogggbm on 10/10/2018 9:29:00 PM (view original):

Posted by mullycj on 10/8/2018 10:14:00 AM (view original):
I don't know why you feel the need to quash anyone's man crush on here topdogg!!

I know! I didn't mean to interfere. Maybe I was getting a little jealous because no one loves me that way!

(I never did get a return sitemail, shoe!)

LOL, I can’t even go 3 days without getting a “where u at?” text now. I see where this is heading.

Posted by shoe3 on 10/11/2018 11:06:00 AM (view original):

Posted by topdogggbm on 10/10/2018 9:29:00 PM (view original):

Posted by mullycj on 10/8/2018 10:14:00 AM (view original):
I don't know why you feel the need to quash anyone's man crush on here topdogg!!

I know! I didn't mean to interfere. Maybe I was getting a little jealous because no one loves me that way!

(I never did get a return sitemail, shoe!)

LOL, I can’t even go 3 days without getting a “where u at?” text now. I see where this is heading.

Well, I tried stopping by your work first. But they said you were off. So I figured as a second option, I'd try to find you here!

Kind of late in the post but I agree completely with the original post. I have played this game for 12 years in various worlds under various names and have noticed a very definite tendency to make things conform to expectations. Not just games but even individual stats to some extent. I find it curious when I am working on a rebuild and play teams I'm not suppose to beat, I will lead at half time and get completely blown out in the second half. Once in a while is chance. This is not once in a while. You can almost predict when you will get a bunch of fouls all clumped together, for instance, because you haven't gotten enough to that point according to game logic. I have noticed these things looking at the pbp for years. Everything looks good in the box score but the way it gets there is frequently clearly not completely random. Just my observation and opinion of course.

Bump: last year's discussion on in-game adjustments to nudge actual game results toward expected results.

For gillispie and indiana and anyone else with insomnia or a few hours to kill.

Oh, fun thread. Thanks for bumping. I have way too much time on my hands.

I love the casino analogy. Dahs mentioned the 1s and 0s example as a counter to flipping a coin. To simplify, if you had a bag full of 1s that represented favorable outcomes for team 1, and 0s that favored team 2, if more 1s happened to be drawn in the first half and were not put back in the bag, then there would be a natural correction and probability that more 0s would be drawn in the second half for team 2. This was written off as a red herring.

Before that example, it was previously asked: "Would you play poker or blackjack at a casino that engineered “expected” results like this?"

Blackjack is exactly the 1s and 0s example in real world practice. Once a card is dealt, you do not put it back in the shoe. Your winning percentage as a player is dependent on the cards previously drawn and discarded. When more face cards or aces have been discarded, the deck favors the dealer and the expected result is lower. More of them? You have a hot deck and better chance of winning. Identifying that by counting cards is how those MIT students won millions. Without being able to see the cards in a blackjack shoe, you could easily sit at a table with a cold deck and lower odds of winning. At this point, the deck is stacked against you. Does this mean you started playing a casino game with perverse probability? No. It’s still statistics. When the decks are exhausted, the house’s winnings will be close to the mean.

FTs keep getting brought up since it is easy to understand. At least in a sim game, free throw shooting is the roulette of the casino. Each spin is an independent action, just like flipping a coin. You have the same exact probability of the ball landing on black every spin. It does not matter when you step up to the table. You will get the same odds. Forcing this type of calculation on all aspects of the game engine is flawed. There are too many variables at play.

The idea that a probabilistic simulation is tainted because it uses dependent events rather than independent events doesn’t make any sense. I can simulate a hand of blackjack just as I can simulate a spin of a roulette table. Just because blackjack relies on dependent events that changes the “weight” of the “coin flip” for each hand played based on previous events, doesn’t mean that the game itself is moving further away from the mean and screwing its players. Blackjack has the best odds of any casino game.

TLDR: I’m guessing the engine is a complex simulation of conditional probability that calculates both dependent and independent events. This doesn't seem to be creating an unacceptable amount of bad beats.

im pretty sure there was no conditional probability prior to the introduction of feedback. every event was independent just like free throws.

It was a completely different game engine. The previous engine builder paid attention to standard deviations of results and would tighten the probability of each individual event.

I believe the current engine has a wider range of probability but then introduced the "correction feedback" to get the final game outcomes back to the expected ranges. Same end result for the most part but a different way of getting there.

Mully, that doesn't make sense. There's no way to "tighten the probability of each individual event" when virtually all of the events involved are all-or-nothing events. If you consider your average 2 point FG to be about a 50% probability event, individual shots have a standard deviation of .5. The only ways to reduce deviations are to increase the number of events or to do what has been done in the current engine and move away from a true series of independent events. The old engine just had broader distributions of outcomes.

I don't think it was much broader, though, which is why I really don't understand why people have gotten so bent out of shape about this.

Dahs - Ill be clearer. Tarek would look at game results across worlds and actually change/tweak the engine every couple seasons to make results more aligned to college results. He was also known to tinker with the recruit generation from season to season. He had much more time to tinker with the game than the current programmer.

I haven't been around that long so I'm not sure how that old engine worked, or this one for that matter. I'd have to think, unless the engine was very simple, that there was still dependent events. Let's look at a few plays:

Player X misses a two pointer.
Player Y makes a 3 at the other end.
Player X makes a two pointer.

These are independent events. If you assume the players on the court didn't change, the ratings and iq of offense and defense dictate the probability of players making a shot. Ratings and IQ don't change once a game starts, just like someone's free throw rating. So why has the probability of player x shooting a two pointer changed slightly between the first and second shots? It's because we have a variable dependent in stamina that changes play to play. This is already a bit more complex than free throw shooting.

Now let's change this up a bit:

Player X misses a two pointer.
Player Y turns the ball over.
Player X makes a two pointer.

Same players on the court, Player X now shoots a two pointer after a TO. The probability of player x making that two pointer has increased significantly. This is a conditional event that only has a chance of occurring after a TO. One independent event is dependent on the previous event. Unless the old engine didn't increase the percentage of making a shot after a turnover, then it had conditional probability.

The heart of this discussion though is the variable modifier that we suspect is in this version of the engine. This is a modifier, much like stamina, that runs throughout the game. Unlike stamina, it isn't tied to any rating. This is why shoe3 calls it godmode. NBA Jam is a good example of an extreme version of this modifier. Why was it extreme? Because it's roots started as an arcade game. By design, arcade games are trying to get you to replay it so that you put another coin in the machine. Someone realized that blow outs don't get replays and last second shots for the win do.

Still trying to wrap my brain around how this less extreme version of the modifier might work in the WIS engine.