What Happens Next

[Kingswood]

Since you creators are emphatic that your sim is realistic explain this:

Columbia Cardinals 40
Indianapolis Immortals 13
Columbia Indianapolis
First Downs 13 20
3rd/4th Eff 3-10 5-14
Total Yards 315 345
Passing Yards 240 140
Comp/Att 23/33 17/25
Comp % 69.7 68
YPA 7.29 5.61
Passing TDs 3 1
Passer Rating 108.2 78.8
QB Pressure % 42.5 46.2
QB Sacked 5 (32) 11 (84)
Rushing Yards 75 205
Rushing Att 11 26
YPC 6.79 7.88
Rushing TDs 0 1
Turnovers 1 5
Interceptions 1 1
Fumbles(lost) 4 (0) 9 (4)
Penalties 2 (10) 2 (20)
Time of Possession 28:33 31:27

Somehow we had similar pressure rates and yet I had over double the sacks. 11 sacks... not unheard of but it happens once every 5-10 years, not once every other day in challenges.

Let's not also forget to look at fumbles. The last time a team fumbled the ball 9 times in a game was in 1990... it's been over 3 decades. We're not even in the same millennium.

None of us believe this is an accurate sim at this point. The only thing realistic is the player names and positions, and even some of the names and positions are wrong. We've all tried to find small ways to manipulate it in our favor, like QB runs, position shifts, etc. As GM's it's our responsibility to find a way to win. The problem is we have no base to go off of. I can't base it off of the NFL because we've established this is unrealistic. You said yourself this only generates off of 1 year of production so it's too limited to be realistic in that regard. We don't have enough stats to go off of so I can't base it off my own inconsistent production. We can't see our opponents' depth chart before or after the game so we have no idea what matchups are winning and which are losing. We don't have any ratings because you don't believe in ratings, which I don't necessarily disagree with.

We do have the challenge games though, which would be great if performances were consistent. I can make one tweak, maybe swap two corners, then challenge the same team 3 times straight. Those 3 challenges will all yield different results. In one game I can be blown out, in the next I could dominate, and the 3rd we could tie or one of us could squeak bye. If you compare those games then for one you obviously have too much inconsistency to utilize any of the data. Secondly, if you attempt to compare those games to the games before the adjustment was made then all you would be doing is comparing inconsistencies.

I love the game and I love the concept, but it would be nice to know that these future updates are going to address these issues. True, football is unpredictable, "Any Given Sunday", blah blah blah. But it's not that unpredictable, and it should be even less unpredictable on a computer, without weather and without in game injuries, imho. After all, I think we deserve to know what we're paying for. Is there a list of changes we can expect in the gameplay/strategy/stats categories? Can we expect more of the same?

And Congrats to DJ and both his left nuts. Dominating the Challenge board like that is impressive. You always give me the best competition.

[cdcox]

Hi Kingswood,

I will address some of your questions. A discrepancy between pressure rates and sacks can be explained by three factors: 1) certain pass rushers are more likely to convert a pressure to a sack, 2) certain QBs are more likely to escape pressure than others and 3) luck. I agree that 11 sacks by one side in a game is an example of things that happen rarely in the NFL that happens too frequently in Sandbox. This is a signal that something in the algorithm needs tweaking. Same with fumbles, and a few other game outcomes. I intend to put some "curbs" in the simulator to avoid some of these extreme outcomes.

I agree that the simulation results need to be presented in a way that enables Sandbox owners to have a better idea of what is working and what is not. That is a goal, when I rewrite the simulator. You mention playing three different challenge games with three different outcomes. That is to be expected as evidenced by interdivision play in the NFL. Many times the outcomes are quite different. The goal for the new version of the simulator is that you may be able to get an idea of which players are winning match ups most often, and where the weaknesses are that you can work on to improve your roster.


I agree that congrats to DJ are in order. His teams consistently do well. Good luck/bad luck affect all of us equally. But DJ appears to be able to consistently put together rosters that can overcome the bad luck to be a consistent winner.

[cdcox]

One other nerdy thing I will mention about the next version of the simulator has to do with coding architecture. I am a self-taught hobbyist software engineer. I'm proud of what I have been able to put together from a software point of view, but because I'm self-taught there are aspects of the code that I would do differently, given what I have learned in recent months about how professional software engineers go about their work. The current simulator is just shy of 6000 lines of code, organized into 75 functions. It still has many aspects of spaghetti code, which refers to code that is difficult to understand, test, and modify. As a result I expect that there are a few "ghosts in the machine" or hidden bugs beyond those that show up regularly in the simulations. Spaghetti code makes it difficult to track down and modify those bugs without introducing new ones.

The new simulator will be a complete rewrite that relies on more disciplined use of object oriented programming and test-driven design. The 75 functions in the current code will be reorganized into probably something like 200 individual classes. Each of the classes will have its own unit test, that allows it to be tested independently of the rest of the code. This will result in the code doing what I intend it to much more reliably. It will also allow me to modify some core logic deep in the simulation, without affecting other parts of the simulation. In other words, I should be able to make modifications in response to things that appear to be in accurate.

The new simulator is a major project, but one I'm excited about and anxious to get started on.

[cdcox]

One final note (again on the nerdy side) is that Brian Burke did some research years ago in which he concluded:

"The actual observed distribution of win-loss records in the NFL is indistinguishable from a league in which 52.5% of the games are decided at random and not by the comparative strength of each opponent."

Here is a links to the first article in a series of three that he wrote leading to this conclusion:

https://www.advancedfootballanalytic...-outcomes.html


You can use the site navigation on the Advanced Football Analytics to fined parts 2 and 3 of this series. I know as an NFL fan, that I don't like to think that luck plays such a significant role in game outcomes when I am rooting for my favorite team. But this randomness is baked into the NFL as well as most sports. Sandbox has this randomness baked into the simulations.

[Kingswood]

One final note (again on the nerdy side) is that Brian Burke did some research years ago in which he concluded:

"The actual observed distribution of win-loss records in the NFL is indistinguishable from a league in which 52.5% of the games are decided at random and not by the comparative strength of each opponent."

Here is a links to the first article in a series of three that he wrote leading to this conclusion:

https://www.advancedfootballanalytic...-outcomes.html


You can use the site navigation on the Advanced Football Analytics to fined parts 2 and 3 of this series. I know as an NFL fan, that I don't like to think that luck plays such a significant role in game outcomes when I am rooting for my favorite team. But this randomness is baked into the NFL as well as most sports. Sandbox has this randomness baked into the simulations.

Thanks! That looks like it will be a good read. I appreciate you taking the time to humor me. I'm glad that I'm not completely crazy.

[Wind From West]

Not to be a ? but I appreciate the crazy inconsistencies, lol.

If you get a DJs nutz or Kingswood that has it figured out, we're all toast. I challenge only the top teams but I feel it won't matter for some of us?

I like the fact none of my D gameplans work for my other team but yet that first team kills the QB and forces turnovers when I deploy it.
Sacks & turnovers don't always mean a W tho.

[Kingswood]

Not to be a ? but I appreciate the crazy inconsistencies, lol.

If you get a DJs nutz or Kingswood that has it figured out, we're all toast. I challenge only the top teams but I feel it won't matter for some of us?

I like the fact none of my D gameplans work for my other team but yet that first team kills the QB and forces turnovers when I deploy it.
Sacks & turnovers don't always mean a W tho.

**** inconsistencies. the game should be skill based. I'm not paying to play poker.

[Kingswood]

One final note (again on the nerdy side) is that Brian Burke did some research years ago in which he concluded:

"The actual observed distribution of win-loss records in the NFL is indistinguishable from a league in which 52.5% of the games are decided at random and not by the comparative strength of each opponent."

Here is a links to the first article in a series of three that he wrote leading to this conclusion:

https://www.advancedfootballanalytic...-outcomes.html


You can use the site navigation on the Advanced Football Analytics to fined parts 2 and 3 of this series. I know as an NFL fan, that I don't like to think that luck plays such a significant role in game outcomes when I am rooting for my favorite team. But this randomness is baked into the NFL as well as most sports. Sandbox has this randomness baked into the simulations.

Definitely an interesting read but all in all I'd call BS. There's too many factors involved in football to chalk up 50% of outcomes to luck. If that's the case then why bother watching 3 hours of football when it can be decided with a coin toss? Is this luck factored into the code as well? Are half of our games decided by luck? Can they calculate just how ****ty my luck has been in your sim?


New Orleans River Cats 48
Newton Broncos 14

New Orleans Newton
First Downs 21 22
3rd/4th Eff 4-10 3-7
Total Yards 335 307
Passing Yards 176 221
Comp/Att 17/19 24/31
Comp % 89.5 77.4
YPA 9.27 7.13
Passing TDs 2 1
Passer Rating 140.4 93.6
QB Pressure % 50 48.8
QB Sacked 3 (22) 8 (47)
Rushing Yards 159 86
Rushing Att 35 12
YPC 4.55 7.18
Rushing TDs 4 1
Turnovers 0 7
Interceptions 0 1
Fumbles(lost) 1 (0) 7 (6)
Penalties 7 (40) 4 (33)
Time of Possession 38:36 21:24

[Wind From West]

**** inconsistencies. the game should be skill based. I'm not paying to play poker.

Id like to think I'm OK playing these games but if a team constantly wins, its frustrating for teams with strong rosters who apparently can't gameplan.
I absolutely see your point. I'm the devils advocate type tho.

[Kingswood]

Id like to think I'm OK playing these games but if a team constantly wins, its frustrating for teams with strong rosters who apparently can't gameplan.
I absolutely see your point. I'm the devils advocate type tho.

Game planning isn't luck though. That's strategy. I'm all for the devils advocate approach. I prefer to see things from every angle possible. Strong rosters with weak game plans should lose just as much as they win imo. What I consider excessive bad luck would be getting sacked 8 times, with 7 fumbles, and 6 of them lost lol. I'm sure their new parameters will address this. I'm looking forward to the next edition.