62 coaches online • Server time: 20:30
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
smeborg



Joined: Jan 04, 2019

Post   Posted: Nov 04, 2021 - 06:46 Reply with quote Back to top

Hi All - this is a simple plea, not a complaint.

This morning I played a match with my newly formed rookie Chaos Renegade team (TV99.5). Without checking (therefore analogous to Box), I accepted a challenge from a Chaos team with TV105 (thinking it might be a 1-game team with Apoth). The Chaos team turned out to be as follows:

4 Warriors, 7 Beastmen, 2 RRs, Apoth
6 random skill-ups, all handy (M-Blow, B-Tackle, Block, Wrestle, Tackle, Kick)
4 games played

No complaints about the conduct of the game, my opponent was very sporting, and we had good banter in chat.

But it felt like like a mis-match (5 eminently usable skill-ups in this match for the TV cost of a RR - a trade not available to rookie teams). This sort of thing will surely happen in BB2020 Box. I suggest it will deter some players, especially newbies who will be "ambushed" by teams built along these lines. If you add re-cycling of players for an optimal mix of random skill-ups, the mis-match would likely be greater.

So my plea is to have 1 different general environment for pick-up games. As has been suggested before, using games played (or range of games played) might be one way to do it.

Hope that helps the discussion along!
mrt1212



Joined: Feb 26, 2013

Post   Posted: Nov 04, 2021 - 06:50 Reply with quote Back to top

Nuffle hears your plea and condemns you to always finding a fly in the ointment of a good time, based on false notions of fairness and why you offer ritual to Nuffle. Enjoy.
Kondor



Joined: Apr 04, 2008

Post   Posted: Nov 04, 2021 - 06:54 Reply with quote Back to top

For this rule set, I think what you are describing is a feature rather than a flaw.

By the way, to roll 6 random skills over 4 games and have what you describe as the outcome is to face a very lucky coach.
sebco



Joined: Feb 14, 2005

Post   Posted: Nov 04, 2021 - 08:15 Reply with quote Back to top

smeborg wrote:
So my plea is to have 1 different general environment for pick-up games. As has been suggested before, using games played (or range of games played) might be one way to do it.

Hope that helps the discussion along!


Thats helps ! (well, ok, I'm subjective, I totally agree with your statements / pleas)

_________________
I like cheese but don't call me skaven !
Kondor



Joined: Apr 04, 2008

Post   Posted: Nov 04, 2021 - 08:23 Reply with quote Back to top

sebco wrote:
smeborg wrote:
So my plea is to have 1 different general environment for pick-up games. As has been suggested before, using games played (or range of games played) might be one way to do it.

Hope that helps the discussion along!


Thats helps ! (well, ok, I'm subjective, I totally agree with your statements / pleas)


Actually, it would not help. It would just be a different meta game situation. You would have teams with random skills playing against larger TV differences in teams that are choosing skills, or, the team that is saving up for a couple of "better skills." gets a game like the one above.

i am actually in favor of complete random match making. Ignore everything and let the inducements do what they are meant to do.
stej



Joined: Jan 05, 2009

Post   Posted: Nov 04, 2021 - 08:33 Reply with quote Back to top

Make all skill rolls random. You get what nuffle gives you and you damn well like it!
mekutata



Joined: May 03, 2015

Post   Posted: Nov 04, 2021 - 09:41 Reply with quote Back to top

Kondor wrote:
sebco wrote:
smeborg wrote:
So my plea is to have 1 different general environment for pick-up games. As has been suggested before, using games played (or range of games played) might be one way to do it.

Hope that helps the discussion along!


Thats helps ! (well, ok, I'm subjective, I totally agree with your statements / pleas)


Actually, it would not help. It would just be a different meta game situation.


I tried that method in C and had good games for both sides that way.

_________________
Image
tussock



Joined: May 29, 2011

Post   Posted: Nov 04, 2021 - 09:54 Reply with quote Back to top

Also, just join a league if you want league play.

Like, SWL does it better, but there's lots of them.

_________________
ImageImage
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Nov 04, 2021 - 10:23 Reply with quote Back to top

A 0-game 995k team would feasibly get matched against a 4-game 1050k team, no matter what type of scheduler we had

They've been very lucky to get 6 useful skills in such a short space of time. Which was just unlucky for you
koadah



Joined: Mar 30, 2005

Post   Posted: Nov 04, 2021 - 10:28 Reply with quote Back to top

tussock wrote:
Also, just join a league if you want league play.

Like, SWL does it better, but there's lots of them.


Keep it Secret! ([SL]2020)
You'll get more than a measly 1350! Twisted Evil

Getting lucky on random rolls is still getting lucky on random rolls though.

_________________
Image
O[L]C 2016 Swiss! - April ---- All Stars - Anniversary Bowl - Teams of Stars - 13th March
argos_72



Joined: Mar 02, 2007

Post   Posted: Nov 04, 2021 - 11:02 Reply with quote Back to top

smeborg wrote:

This sort of thing will surely happen in BB2020 Box. I suggest it will deter some players, especially newbies who will be "ambushed" by teams built along these lines. If you add re-cycling of players for an optimal mix of random skill-ups, the mis-match would likely be greater.



I believe that this problem can be easily solved by better defining the association rules of the Box's Match Making Algorithm.
Currently, they look like this

https://fumbbl.com/p/notes?op=view&id=522

The Box matchmaking mechanism now only looks at the TV and not the number of games played.

Rather than proliferating other Game Finder systems, perhaps the Box algorithm could be refined to try to make associations that also take into account the number of games played (I quite agree with you on this).

Keep in mind, however, that the goal is to maximize the number of games that can be played so there is always a tradeoff to be made between available "offer" (number of coaches who want to play with certain teams) and reality (games that are played).

In any case a system based on GF cherrypick assumes that a coach will look at the roster of his opponent before accepting a challenge and this is the reason why you very rarely have very unfavourable matches (but this can also reduce the number of matches and races played).

I hope that when the Black Box is activated at least 50% of the coaches will decide to use the Box mechanism and 50% other systems (GF, League or tournaments). A good balance between the association systems is very fair.
Christer



Joined: Aug 02, 2003

Post   Posted: Nov 04, 2021 - 15:13
FUMBBL Staff
Reply with quote Back to top

The thing with match making Blood Bowl is that people don't want to play games that have mismatching TV. It doesn't matter if you say "you've played the same number of games" or not. If someone gets paired against an opponent that's 100TV up, they get upset and feel the match is unfair.

I've tried alternative ways to schedule matches in the past and it causes huge issues in the perception of people, and essentially kills the format. TV-based scheduling is the only thing that works on a large scale.

TV is far far from being a good indicator, but it's what people know.
argos_72



Joined: Mar 02, 2007

Post   Posted: Nov 04, 2021 - 16:19 Reply with quote Back to top

Christer wrote:
The thing with match making Blood Bowl is that people don't want to play games that have mismatching TV. It doesn't matter if you say "you've played the same number of games" or not. If someone gets paired against an opponent that's 100TV up, they get upset and feel the match is unfair.

I've tried alternative ways to schedule matches in the past and it causes huge issues in the perception of people, and essentially kills the format. TV-based scheduling is the only thing that works on a large scale.

TV is far far from being a good indicator, but it's what people know.


yes and this is perfect for me and I think for most of person the TV gap is the first parameter to check for a matchmaking in Box division.

Today with the redraft we will also reduce the risk of big TV gap (which in any case could be possible in BB2016 because
For 10 to 29 games played, max TV difference allowed is (15 + (games - 10) * 2)%
For 30 or more games, any opponent is allowed
and this naturally was a problem in a permanent league where you could have a team with TV 1600 and 30 game played and another TV 2200 with 200 games played

as I mention today i think with the redraft this will be no more a problem so the matchmaking algoithm will be even more "fair" which is a good news Smile

The only further optional point we can think to add in the BB2020 Box Matchmaking algorithm could be to the number of game played , because it is true that you can play at TV 1000 with a team that on the paper is TV 1000 but which contains maybe one superplayer because it is a recovery match where legend/starplayer are out

example
https://fumbbl.com/p/team?team_id=1017349

someone could be surprised to meet CD team at TV 1040 with one CLAWMB and 3 guard (as the 2 bull centaurs are out)
But I think this will be exceptions

The drawback to add too many "filters" in the Box Matchmaking algorithm is risk that no association is done and this is not good where there are minimum number of coach (4) which are activating.
So of course the matchmaking algorithm should be a trade off and I think that as it is implemented today plus the introduction of the redraft will generate a Box division more fun and balanced!

So Looking forward for the Box BB2020 launch (and maybe we can celebrate with a nice Trophy! )
Wink

thanks Christer again for the great job
JanMattys



Joined: Feb 29, 2004

Post   Posted: Nov 04, 2021 - 19:01 Reply with quote Back to top

Christer wrote:
The thing with match making Blood Bowl is that people don't want to play games that have mismatching TV. It doesn't matter if you say "you've played the same number of games" or not. If someone gets paired against an opponent that's 100TV up, they get upset and feel the match is unfair.

I've tried alternative ways to schedule matches in the past and it causes huge issues in the perception of people, and essentially kills the format. TV-based scheduling is the only thing that works on a large scale.

TV is far far from being a good indicator, but it's what people know.


This is probably very true, but do you think things might change with the new ruleset? After all it is undeniable that bb2020 put a lot of effort in the inducements, in terms of quality, quantity and diversity.

I know inducements do not (or do not always) cover the differences in TV and they are situational at best, but with such a rich selection, wouldn't it be possible that larger than average TV gaps might have become a little more appealing/acceptable now?

_________________
Image
smeborg



Joined: Jan 04, 2019

Post   Posted: Nov 04, 2021 - 20:40 Reply with quote Back to top

Christer wrote:
The thing with match making Blood Bowl is that people don't want to play games that have mismatching TV. It doesn't matter if you say "you've played the same number of games" or not. If someone gets paired against an opponent that's 100TV up, they get upset and feel the match is unfair.

I've tried alternative ways to schedule matches in the past and it causes huge issues in the perception of people, and essentially kills the format. TV-based scheduling is the only thing that works on a large scale.

TV is far far from being a good indicator, but it's what people know.

Thanks, Christer, and I well understand.

I will see if I can gather my thoughts to formulate a simple team-matching logic in line with my tentative thinking.
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