New Sticky Note

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

League Forums

Main - Beta Chat

Re: [0.4.6] Version 09d60821

By raymattison21
9/17/2020 10:24 am
Smirt211 wrote:
(anyone know why if you key on Man OLB Flat Zone every week your team stays at 0% or drops to a negative % for familiarity).


It's a desire to make zone defense function-able. Bring it to the forefront. Awhile back blitzing was given that over-use penalty and then endless efforts were poured into giving credibility to zone defensive plays to push that up while pushing blitzing down.

But, yeah, I don't like it. I've been at -75% on that play when someone mashed it at me. It rewarded the mashing with each and every usage of it.




By jdavidbakr - Site Admin
6/05/2019 9:05 pm
I think the bars suddenly becoming empty is a rendering issue, possibly due to a rounding error causing it to display greater than 100%.

i never heard anything except this.....

Re: [0.4.6] Version 09d60821

By Smirt211
9/17/2020 10:33 am
I just pulled up that set over in No Holds Barred.

2 of the 4 plays I'm using and all the rest in the set are near/at full bars but 50% of the plays I use I had to build up from the 2 RB/3 TE set.

I'm at like 33% and 38% on them; clipping off 5 yards per play.

Last edited at 9/17/2020 10:34 am

Re: [0.4.6] Version 09d60821

By raymattison21
9/17/2020 10:35 am
Smirt211 wrote:
2 RB/3 TE set.

Think of it as a starting basis/gift. It's a main goal-line set; thus supposed to be a widespread every team uses it function. Therefore, for years (real time!) you're given that as a starting point to build around. 2 RB/3 WR is given some plays at full red bar, too, right?

I think it's cool to give something to operate off of but I do want to say if my memory is correct newer leagues like CJ's No Holds Barred the 2 RB/3 TE set started off @ 0% on all plays.

I may be wrong.




By jdavidbakr - Site Admin
6/23/2015 9:57 pm
When a league is auto-drafted the familiarity is just as random, through the season you'll build a very solid familiarity base. The players have the familiarity set before the draft, but IMO it wouldn't be worth drafting players based on that since they will learn your playbook soon enough. Drafting for your HC's style is a much more solid strategy


Re: [0.4.6] Version 09d60821

By Smirt211
9/17/2020 10:37 am
I guess nods but basically you'll only get full bars on 2-1-2 HB Blast, 2 RB/3 TE and a few 2 RB/3 WR plays. I haven't memorized it.

Mostly it'll be blank slate or very thin red bars!

Thank you for the info. :)

Re: [0.4.6] Version 09d60821

By TheAdmiral
9/17/2020 6:15 pm
OK Here's cross league date on the Team passing stats.

Looking at

Passing Yards
Yards/Attempt
Completion%
Interceptions
Sacks

I have collated the data from

MFN-1: BETA (last 5 seasons only)
Victory Football League: BETA (one full season, currently drafting in year 2)
League of Legends: 0.4.5 (six full seasons from dispersal draft)
No Holds Barred: 0.4.5 (one full season from dispersal *No Punt Block rules in place)
Franchise Football: Roster Only (two and a half seasons of data, currently season 3, midweek 9)

I think there is a bug in the reporting of Sacks and Interceptions

Feedback appreciated.

I haven't really analysed the data yet so I'm interested to see how you read it

https://docs.google.com/spreadsheets/d/13orAbhEFkdLcvtO6dNiucuyer4LoeY32ZaNuhTIrCxM/edit?usp=sharing

*I haven't updated all the highest and lowest outliers for each league yet - hoping to get around to that tomorrow*
Last edited at 9/17/2020 6:15 pm

Re: [0.4.6] Version 09d60821

By setherick
9/17/2020 6:20 pm
I don't understand the purpose of doing data analysis of beta code. For one, the code base changes too often for the data to be reliable. And, for two, I don't understand doing manual data analysis for something that should be being covered by an automated testing suite.

Really, each play should be tested against all other plays in isolation 10,000 times and then the data agg'ed. What it would show is what all good game planners already know. Most offensive plays do not work ... at all ... against any defense. And the ones that do work are because of broken interactions between players.

Re: [0.4.6] Version 09d60821

By TheAdmiral
9/17/2020 6:35 pm
setherick wrote:
I don't understand the purpose of doing data analysis of beta code. For one, the code base changes too often for the data to be reliable. And, for two, I don't understand doing manual data analysis for something that should be being covered by an automated testing suite.

Really, each play should be tested against all other plays in isolation 10,000 times and then the data agg'ed. What it would show is what all good game planners already know. Most offensive plays do not work ... at all ... against any defense. And the ones that do work are because of broken interactions between players.


Agreed, but then you also need to cover strength of offense v strength of defense, quality of coaching and a million other variables (exaggeration).

I did this because I was interested to see whether the discussion on interceptions being up and sacks being down was anecdotal perception or if it was genuinely trending that way across all leagues.

Following the switch to 0.4.6 as and when it happens, this will surely be helpful, no?

Re: [0.4.6] Version 09d60821

By setherick
9/17/2020 6:38 pm
Actually, you don't need to account for those things. All you need to account for is player attributes, which are easy enough to stratify over 10,000 tests. Testing plays in isolation against each other will expose many of the issues with both offense and defense if you create the right set of metrics.

Re: [0.4.6] Version 09d60821

By raymattison21
9/18/2020 5:19 am
TheAdmiral wrote:
setherick wrote:
I don't understand the purpose of doing data analysis of beta code. For one, the code base changes too often for the data to be reliable. And, for two, I don't understand doing manual data analysis for something that should be being covered by an automated testing suite.

Really, each play should be tested against all other plays in isolation 10,000 times and then the data agg'ed. What it would show is what all good game planners already know. Most offensive plays do not work ... at all ... against any defense. And the ones that do work are because of broken interactions between players.


Agreed, but then you also need to cover strength of offense v strength of defense, quality of coaching and a million other variables (exaggeration).

I did this because I was interested to see whether the discussion on interceptions being up and sacks being down was anecdotal perception or if it was genuinely trending that way across all leagues.

Following the switch to 0.4.6 as and when it happens, this will surely be helpful, no?


I like stats and kinda remember the code of what I was trying to exploit at that time , but just look at infinity s last game ...7 ints ? Push the sim to the max and it will respond.

On the other side mwd ran a low blitz heavy nickel pass defense and got ran all over. Like both you are saying is why? Or should it be so powerful either way?

As I blitzed a lot and shut down the run but can’t really pick off a pass for the life of me. There’s got to be a balance.

If the little parts are working than the big numbers will fit but plan has to be accounted for . Simply said air yards is way off. It’s been for awhile. Our total number might be close but YAC is a bit high. Not due to broken tackles but those deep cover guys jumping routes.

We never have a defender playing too deep. Well unless it on those rare comeback passes. Which is our longest air yard passes.

Re: [0.4.6] Version 09d60821

By Infinity on Trial
9/18/2020 7:26 am
Conference championship passing:

We combined for 39/78 (50%) for 379 yards (4.85 per attempt) and a whopping 7 interceptions (all by my opponent).

But hey, no sacks!

I still have no rules or overrides in place. My opponent knows what he's doing and has a great team. It's difficult to see how 4.6 is going to be an improvement.