41 coaches online • Server time: 16:14
Forum Chat
Log in
Recent Forum Topics goto Post Creating a custom to...goto Post Secret League Americ...goto Post DOTP Season 4
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
Desultory



Joined: Jun 24, 2008

Post   Posted: Aug 23, 2016 - 20:27 Reply with quote Back to top

mrt1212 wrote:

So in Box there is new team protection. Basically the algorithm makes it so that the suitability of a match between two teams, one with over 30 games played and one with under 30 games played is very unlikely if not impossible at times.

I feel that the lifespan of a team is probably a much stronger indicator of how well the team is built and how competitive it is rather than of pure TV but I'd want to do an analysis of how much of a predictive value lifespan is in terms of winning.


Ok. That's great that it has theoretically already been implemented. Mathematically, what does 'very unlikely' actually mean. Because I got the game...

One doesn't really need to analyse that deep to realise that a +ST, guard black orc blocker isn't worth 150k, but a +ST,guard,block,stand firm blocker is probably worth 190k. That maybe would be the difference between 135 games and 17games...

More game teams are obviously stronger, but development time is not factored in to TV. Isn't Fanfactor (failing) meant to help with that.
mrt1212



Joined: Feb 26, 2013

Post   Posted: Aug 23, 2016 - 20:49 Reply with quote Back to top

As far as I'm concerned, Fan Factor is both a benefit and a yoke on teams. On the one hand it's TV that is situationally utilized to benefit yourself on FAME but more often than not it just adds to your TV in a way that might give your opponent inducements etc. So some days 140 TV in Fan Factor gives you +2 Fame, you get an additional reroll on one kickoff and it's provided some value. Other days it's about 50 to 80 TV of deadweight relative to your opponents fan factor that doesn't result in any positive outcomes for the match.
garyt1



Joined: Mar 12, 2011

Post   Posted: Aug 23, 2016 - 20:53 Reply with quote Back to top

DarthPhysicist wrote:
Farming is not equated to firing/rehiring. Farming would be that you only play against all halflings or all snot teams in a preconstructed way, just so you could get spp.

Yeah maybe that is it but Benji plays decent opponents with cr over 150 so it doesn't really make sense in this case.

_________________
“A wise man can learn more from a foolish question than a fool can learn from a wise answer.”
keggiemckill



Joined: Oct 07, 2004

Post   Posted: Aug 24, 2016 - 03:13 Reply with quote Back to top

mrt1212 wrote:

So in Box there is new team protection. Basically the algorithm makes it so that the suitability of a match between two teams, one with over 30 games played and one with under 30 games played is very unlikely if not impossible at times.

I feel that the lifespan of a team is probably a much stronger indicator of how well the team is built and how competitive it is rather than of pure TV but I'd want to do an analysis of how much of a predictive value lifespan is in terms of winning.

I thought this was at 15 games? Ive been matched up several times with a team with over 50 games, vs my team with 10 or less games. This is not only recent, but after the changes. It has helped curb some of the bad match ups, but hasn't curbed them completely.


Desultory wrote:

Ok. That's great that it has theoretically already been implemented. Mathematically, what does 'very unlikely' actually mean. Because I got the game...

One doesn't really need to analyse that deep to realise that a +ST, guard black orc blocker isn't worth 150k, but a +ST,guard,block,stand firm blocker is probably worth 190k. That maybe would be the difference between 135 games and 17games...

There was an interview on cost of players explained in a 3 die block interview with Tom Anders. Everything is based off of a Human Lineman, for the cost of 50k. I would have to listen to it again to fully understand it, but it makes sense when you here it broken down.

Desultory wrote:

More game teams are obviously stronger, but development time is not factored in to TV. Isn't Fanfactor (failing) meant to help with that.


I don't understand what this question is. TV is supposed to reflect a teams growth, and/ or worth.If a rookie team gains after one game 2 ST ups, the team move up 10 TV points. from a 100 to a 110. If a team loses a 50 k Lineman, it goes down to 950 TV. Its exact mathematics.

_________________
The Drunker I get, the more I spill
Image
"Keggie is the guy with the bleach blond hair that gives answers nobody else would think of."
Jeffro
mrt1212



Joined: Feb 26, 2013

Post   Posted: Aug 24, 2016 - 03:24 Reply with quote Back to top

I could be wrong on specifics but something exists.
DarthPhysicist



Joined: Jun 14, 2015

Post   Posted: Aug 24, 2016 - 05:00 Reply with quote Back to top

So the worst case of farming I've seen, was in BB2 where a game was listed as 9-0, so I watched the replay. A wood elf player was playing his team against all linerats and would knock down the entire front line, run out a catcher, stand up the linerats again, knock them down, and then score a TD. Rinse and repeat for TDs and casualties. To me, THAT is farming. Regulating your TV for maximum effectiveness, especially in B, isn't farming; it's just team management. In fact, given that the teams can be played indefinitely, it seems natural to be a bit more selective in your player skills. There's a certain team that we all know and love (somebody thought they saw them in Iraq but turns out, they were just on FUMBBL) and it is fairly well (and annoyingly) trimmed and managed. This isn't farming (at least according to my anecdote).
mrt1212



Joined: Feb 26, 2013

Post   Posted: Aug 24, 2016 - 05:05 Reply with quote Back to top

Exactly DarthyPhysicist, what most consider farming is so egregious that it's laughable to use it as something that invalidates the other coaches team. You'd have to do some digging before you could legitimately lob that feeble grenade and then what...nothing! We have admins to deal with legit abuse.
Kryten



Joined: Sep 02, 2003

Post   Posted: Aug 24, 2016 - 06:38
FUMBBL Staff
Reply with quote Back to top

The details of how the scheduler works can be found here. The relevant bit is this:

"8. Calculate rookie protection factor:
r_games = ( Min(games_1, games_2) + 1 ) / 15, if either team has less than 15 games and the opponent 15 or more
r_games = 1, otherwise"

Basically, the suitability is reduced if one team is below 15 games and the other is above. A team with zero games has the largest protection with the suitability being multiplied by 1/15. That protection goes down as the team plays games, so in keggie's example of a team with 10 games, the suitability would be multiplied by 11/15. For example, if the suitability would have been @950, it would be @696 after adjustment for rookie protection. Enough for the scheduler to try hard to find a different matchup, but still it's going to happen sometimes.
pythrr



Joined: Mar 07, 2006

Post   Posted: Aug 24, 2016 - 06:40 Reply with quote Back to top

team is perfectly ok

your opponent is an idiot

_________________
Image
Image
Wreckage



Joined: Aug 15, 2004

Post   Posted: Aug 24, 2016 - 08:37 Reply with quote Back to top

The algorythm and balancing in general has not overly much meaning for the scheduler in 4 team schedulings. For the health of the draw to be good there would need to be very many coaches that activate with very many teams.
At best you won't get the straight out worst matchup you could get.

And the needs of the most coaches go towards: 'Quality, what is this? I don't need this curset thing.'
Badoek



Joined: May 17, 2009

Post   Posted: Aug 24, 2016 - 09:17 Reply with quote Back to top

Dalfort wrote:
I thought "farming" was when you are dominating play and start making passes or safety net for a hand off to an AG2 (1) player to score. I hadn't realised that picking was part of the farming process and that I have been enabling more farmers than the UK government...


You're crop Dalfort. Crop Laughing
ArrestedDevelopment



Joined: Sep 14, 2015

Post   Posted: Aug 24, 2016 - 09:47 Reply with quote Back to top

keggiemckill wrote:

I thought this was at 15 games? Ive been matched up several times with a team with over 50 games, vs my team with 10 or less games. This is not only recent, but after the changes. It has helped curb some of the bad match ups, but hasn't curbed them completely.


Kryten basically explained this, but yes, the algorithm does not completely prevent old teams vs newer teams (it also doesn't actually prevent an entirely rookie team facing an older one, it just makes the suitability score so low as to make it almost untenable, but some matchups will get through).

The algorithm for rookie protection as it were (sub 15 games) also doesn't appear to care if the >15 game opponent is 16 games old or 600.

keggiemckill wrote:
There was an interview on cost of players explained in a 3 die block interview with Tom Anders. Everything is based off of a Human Lineman, for the cost of 50k. I would have to listen to it again to fully understand it, but it makes sense when you here it broken down.


You can see pretty much the process and costings for CRP here: https://fumbbl.com/help:CreateRoster


keggiemckill wrote:
I don't understand what this question is. TV is supposed to reflect a teams growth, and/ or worth.If a rookie team gains after one game 2 ST ups, the team move up 10 TV points. from a 100 to a 110. If a team loses a 50 k Lineman, it goes down to 950 TV. Its exact mathematics.


Pretty obvious - TV isn't directly correlated to team ability/strength (and yet we are supposing it is, when we use TV to match-make), since certain skills/stats don't have the same immediate impact on the pitch compared to other skills/stats or stacks of skills. We already know that.

Desultory supposed FF is supposed to act somewhat as a balancing factor in growth, but that simply isn't true, except actually for the new team - over time, most teams will end up sitting around 8-12 FF (in my experience). A new team can hit that in its first 8 games, even through just a bunch of tied games. So there, it's inhibiting growth by giving less/more inducements (league) or a harder match-up (tv-based match-making, assuming even TV).
This gets thrown right out of whack in a perpetual environment.

None of this is new.

_________________
Image
Desultory



Joined: Jun 24, 2008

Post   Posted: Aug 24, 2016 - 10:16 Reply with quote Back to top

ArrestedDevelopment wrote:

The algorithm for rookie protection as it were (sub 15 games) also doesn't appear to care if the >15 game opponent is 16 games old or 600.

Pretty obvious - TV isn't directly correlated to team ability/strength (and yet we are supposing it is, when we use TV to match-make), since certain skills/stats don't have the same immediate impact on the pitch compared to other skills/stats or stacks of skills. We already know that.

Desultory supposed FF is supposed to act somewhat as a balancing factor in growth, but that simply isn't true, except actually for the new team - over time, most teams will end up sitting around 8-12 FF (in my experience). A new team can hit that in its first 8 games, even through just a bunch of tied games. So there, it's inhibiting growth by giving less/more inducements (league) or a harder match-up (tv-based match-making, assuming even TV).
This gets thrown right out of whack in a perpetual environment.

None of this is new.


I agree entirely with what you've written.

I did incorrectly suppose fan factor was meant to be a balancer.
Some sort of algorithm that uses some kind of scalar differentiating would be nice, but I don't think there would be enough active teams in blackbox to really make it worthwhile.

25~30 games seems better to me than 15, but I assume then we get in to problems of matching >30 completed game teams?

It's important to roll good (skills and in game) in those first 15 games then. Maybe I might start retiring my healthy teams on 14 games..... Smile


Last edited by Desultory on Aug 24, 2016 - 10:22; edited 1 time in total
Desultory



Joined: Jun 24, 2008

Post   Posted: Aug 24, 2016 - 10:19 Reply with quote Back to top

P.s. On a big guy from https://fumbbl.com/help:CreateRoster
Bone-head and Wild Animal do not change the price.
Mighty Blow does not change the price.

The TV is the flaw. I find it's calculation overly simple.
I'm sure a group of skilled players could revamp it so TV reflected actual ability better. (More IF/THEN factors)

i.e. CPOMB together is 200k instead of 70k approx.


Last edited by Desultory on Aug 24, 2016 - 10:27; edited 1 time in total
harvestmouse



Joined: May 13, 2007

Post   Posted: Aug 24, 2016 - 10:27 Reply with quote Back to top

The problem is that TV works pretty well for resurrection. The only large problem (that I can see) with resurrection TV is that there are a lot of skills that are underutilised.

In my mind, there's absolutely no way that a team should be sitting on over a million gold and a 11 players like the OPs. That is just wrong and limiting the scope of the game.

I think to fix TV, you first would have to recognise different formats of the game and their needs. Then work out a handicapping and rating formula for that format............it seems we have this conversation at least 1 a month.
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic