NOTICE: This league is using the BLEEDING EDGE game engine. For more information, click here.

The new user interface is in preview!

Want to check it out? Click here! (If you don't like it, you can still switch back)

League Forums

Main - Community Help Forum

Re: Play Selection

By ArmoredGiraffe
11/03/2022 7:24 pm
Action-Jackson wrote:
Thanks for the info. It's definitely not sample size. I'm talking multiple teams over multiple seasons, downs and formations. I will look more into GP but I thought that just modified short / medium / distance not play selection.


Definitely double check your sliders. I just checked a team of mine with the same play calling and have an even split over 13 games with the hurry up/slow down sliders at 0

Re: Play Selection

By Action-Jackson
11/03/2022 10:07 pm
It has to be one of these things. All my plays are selected by rules but I doubt that's it. What it has to be is that I was under the impression that there was a 50% chance a run would be called because well I selected 25 & 25 and it just looked at that and then the play later. I have more pass plays than runs in the per down rule so it must factor that in thus the logic of your previous comment about checking the percentages. Sound correct?

Re: Play Selection

By Action-Jackson
11/04/2022 3:20 pm
I really want to stick with the game but I'm throwing constantly down the stretch with this setting. Like under 2 minutes.

Slow Down - Min remaining per score
Same as hurry up, except this switches you to clock-draining mode. You will rarely pass the ball and will take as much time between snaps as possible. This is calculated on each play, so if your opponent scores and takes you out of the set range, you will exit slow down mode.
5

Re: Play Selection

By Mcbolt55
11/04/2022 3:25 pm
You said before you exclusively call plays based on rules? In that case wouldn’t they always override this general slider? Perhaps there is something you need to edit in there.

Re: Play Selection

By Action-Jackson
11/04/2022 4:27 pm
I think you're right. I think the rule may over ride that setting altogether.

Re: Play Selection

By CrazySexyBeast
11/04/2022 8:45 pm
raymattison21 wrote:

3. Chooses a play based on weights (larger weights = higher probability)
4. If no play was available (i.e. you don't have any inside running plays with non-zero personnel and the selection was an inside running play), a play is randomly selected.

Looking at a log from Victory league the I see #4 coming into play.


Logic would dictate a return to (3) and dismisses 4. At no opportunity should the MFN AI overrule the plays selected by the User in preference to any play not selected by the user in said offensive or defensive plays selected by the user. (WHEW!)

This is both a coding loophole and a code absolute.
If 3 fails, then 4, submit to three.
If any play selected by the user fails to fill option 3 - then the AI chooses a random or play matrix optimal based on user selected plays available.
Either/or; A/B (or A/B+C; A/B+Cx) selection processes.

The current code only moves forwards. When the code reaches no answer, it is forced into the current 4.6 MFN AI processing, which in turn can select "undefined" (plays the user did not select for inclusion).
This issue goes at least back to 4.3, when I started playing.
"The basest of the base."
"The beast of two backs."
~Shakespeare

The fix would require the code to move backwards from its current endpoint (or, "thinking"), and then choose from plays selected only by the user after eliminating any and all non-user options based on defensive LB/DB and personnel sliders (with user created rules being the priority).

In this circumstance, the end result logic chain will not and cannot ever be ideal for the user, but at least the end result is something the user chose to include as viable. Devil be damned, the user got what they wished for, from plays the user thought and choose to be ideal.
Rules exist ;)
^eliminates user complaints

In simple terms, right now the MFN AI only asks how there is no play, but never asks why there are no user options (via a process of elimination based on slider/rules/other current MFN parameters involved).
...And therefore never chooses any user options selected.

^ Regarding the above optimal user engagements currently available in MFN, a new code process/logic chain would be required to force the AI into:

If three [#3 above] fails (X) times, then (Y). With Y being a play chosen by the AI that is IN the users playbook - and not possible to select outside of user choices.
...and thus eliminating the current AI option #4 (above); replacing it with a viable, A/B + C (orCx) logic based coded intuition, and user derived option 4 moving forward within the current available play matrix definitions.
...and a welcome, new definition of the current and untenable AI #4 options.

This would be a gigantic fix.

Most of the backing code required is already present.

TBD by JDB. <serious coding requiring a real deal and absolutely dedicated, thoroughly limited beta group running play sims not game sims. <this would require a new beta MFN "outside" yet within whatever current code is in mass use.

It definitely would not be the current Beta league structure - which has proven to be disastrously biased with 4.6.

In no way should the AI ever be choosing offensive or defensive plays outside of what the user has selected to be included.

Ever.

Fix the engine or I quit.
Cache/categorize/fix/progress.
Good for dev, good for humanity!
:)
Never perfect.

So many of us want to help, just give us the proper playgrounds to be your hamsters.

#deepsheepmaynard
*edited for the LCD and yer welcome.
Last edited at 11/04/2022 10:11 pm

Re: Play Selection

By CrazySexyBeast
11/04/2022 8:50 pm
Tron, bros. User basics Jr High level.
Seriously, we've been here b4.
If y'all be Galactica, then so shall we be again.
So say we all.

Last edited at 11/04/2022 8:50 pm

Re: Play Selection

By CrazySexyBeast
11/04/2022 10:19 pm
*qualifiers:
20 yrs with Digital Addiction/Nioga as a CMO and Code Dev Leadership/Base Code Project management 1999-2019.
Please, Google it. Especially all that press from Wired, PC MAgazine, Inquest, and Wizard publications.

I know this stuff.
I really, really do.

Use me, dammit.
Dvorak accomplice circa fall 2000.
aka duckynine/duckyeight
The f'ing OG ducky rolling back to 1992.

PS: Y'all are welcome to look up sanctum on Patreon, we are deving 2.0 3d 12x12 awesomeness with 600+ current cards available.
Random unlimited variables?
PFFFT.
Cake walk, cue the LP.
Last edited at 11/04/2022 10:28 pm

Re: Play Selection

By CrazySexyBeast
11/04/2022 10:25 pm
^ Me throwing me for freelance work yet a billionth time since joining MFN in 2019.
Did I say FREE?
I did.

Re: Play Selection

By Action-Jackson
11/06/2022 4:13 pm
Using rules definitely over rides the slow down rule and I suspect it also affects play selection somehow. I have switched back to using sliders so time will tell.