I'm sure this has been asked before, but it's early in the morning.

Scenario:
The final score of a game is a 10 point loss; the play by play shows a 4 point game with 1 minute to go; the play by play shows you start intentionally fouling and eventually lose by 10.

Question:
Does that mean the engine simulated the game as a 4 point game, then since you weren't playing real-life Northern Iowa, the end-game strategy just didn't pan out? Or does it mean the engine simulated it as a 10 point game with all user settings taken into account into one large formula?

What I'm basically asking is whether there is any relevance to the play by play. Do I know for sure that my end game strategy was deployed or is the entire play by play truly just 100% window dressing?

Thanks,

So many responses!!

each position is simmed in order, your end of game setting matter

I know they matter. My question is whether everything (including the end game settings) are all part of 1 equation that spits out the 10 point loss in my example. Meaning the play by play is just window dressing to get to the final result. Alternatively, does the 4 point margin with 1 minute to go actually have some meaning?

Granted this would be just 1 game. But there is theoretically a big difference between playing a team to a 4 point game and said team just hitting FTs at the end to ice it, vs playing a team to a 10 point game and knowing nothing else about the result.

Posted by ekswimmer on 3/21/2016 10:26:00 PM (view original):
I know they matter. My question is whether everything (including the end game settings) are all part of 1 equation that spits out the 10 point loss in my example. Meaning the play by play is just window dressing to get to the final result. Alternatively, does the 4 point margin with 1 minute to go actually have some meaning?

Granted this would be just 1 game. But there is theoretically a big difference between playing a team to a 4 point game and said team just hitting FTs at the end to ice it, vs playing a team to a 10 point game and knowing nothing else about the result.

NO, to the question you're asking. However, do I think there are other wonky settings that go into effect in the last minute of the game, yes. But to cover or (even weirder) make a spread, then I would definitely say no. In my humble opinion.

Every individual play happens.. It is not one sim. It is a complete series of plays, one after the other.. In order. Each play has a percent chance something will happen, and there is a generated number that looks up that play. Think dice roll and a board game.. One play at a time.

Awesome. Thanks.

Just 1 point of clarification:
So I should assume the play by play accurately reflects the simulation as it runs 'event' by 'event' (TO, pass, shot, rebound, etc.)?

yes,

I wondered this just this morning. I lost a game last night where my opponent tied it up on a 3-pointer, which sent us to OT and then lost by 7 in OT. I was wondering if I was always going to lose by 7, or if singular events in both regulation (e.g., missed ft, offensive rebound, etc) mattered. Interesting stuff that I had never really thought of before.

This is how it works. This is specifically for the NBA sim game but HD uses the same general (with different components) process.

https://www.whatifsports.com/knowledgebase/KB_Article_Details.aspx?kbid=582

Right, exactly like that. And HD does not have historical shooting rates to determine the chance a shot is for player A or Player B, etc. That is where distribution comes in. So, if you have 5 guys in the game, with distro settings of 4, 1, 1, 1, 1.. Then the guy with 4 has a 4/8 chance of being the shooter, each other player has a 1/8 chance in being the shooter. In HD, the distro rating can be modified by double teaming. (making the Double teamed guy shoot less by lowering his distro).

Once they pick the shooter, the game proceeds event by event.. The outcome for the game is not predetermined, things like guys picking up fouls and sitting (so your leave in longer or take out quicker) have an impact, etc.

Each play happens individually, in series, totally independent of any other play. If a game is played 100 times, there will be 100 completely different outcomes.

It is also why a guy with a 1/8 (12.5% chance) of taking a shot based on distro might take many more or less than 12.5% of the teams shots in a given game. He has a 12.5% of taking a shot... If the team takes 1000 shots while he is in, he will take about 125 shots. But just like you CAN flip a coin 10 times and get 10 heads (each flip is independent and not tied directly to any other flip).. You can have a guy, over a short period (like one game or one quarter) take many more or fewer shots than his distro would predict. But in the long term, that percentage will balance out, just like in the long term (over thousands of flips) a coin flip will turn out very close to 50% heads and 50% tails.

There are things that do not count as normal shots as well.. Offensive rebound put backs are not part of distro (normal) shots. Fastbreaks from steals are not, etc.

Posted by ekswimmer on 3/22/2016 12:07:00 AM (view original):
Awesome. Thanks.

Just 1 point of clarification:
So I should assume the play by play accurately reflects the simulation as it runs 'event' by 'event' (TO, pass, shot, rebound, etc.)?

although there are some aspects of the play by play that are hopelessly inarticulate and dont make clear what happened - esp at the end of the game