shadow46x2
Joined: Nov 22, 2003
|
 Posted:
Jan 15, 2009 - 13:25 |
 
|
| Paragon wrote: | | I'm curious what kind of stats other coaches are getting in the box. |
my chaos are fairly balanced....but then again, i'm not concerned about a balanced schedule with a team that can handle most races easily....
my delfs, on the other hand, are a completely different story, and this is what pisses me off...
orc - 7
khemri - 5
dwarf - 4
chaos - 3
norse - 2
woodies - 2
amazon - 1
delf - 1
helf - 1
lizard - 1
undead - 1
| Paragon wrote: | | My sample size is small, however, so it would be nice to see some more actual numbers from other coaches. |
take a look at brasky's numbers..he's the resident addict...
brasky's chaos..
orc - 40
dwarf - 19
khemri - 14
cdwarf - 13
chaos - 12
skaven - 10
helf - 7
human - 6
woodie - 5
delf - 4
lizards - 4
amazon - 3
ogre - 3
undead - 3
necro - 2
fling - 1
norse - 1
now, keep in mind...these two teams don't mean the system is entirely borked...nor does it mean it is perfect...it's just two examples of hideously unbalanced "diversity"...
| Paragon wrote: | | One other thing I find interesting is that the number of [R] games has increased while the number of [B] games has decreased over the past couple of weeks. If coaches are abandoning [B] in favor of [R], it's worthwhile asking why they're doing that, and how big a factor racial diversity has been in that choice. |
i can say i've written off blackbox on several occasions because of just that, the lack of racial diversity....
the only reason i continue to play it is because i have some sort of sentimentality invested in the teams, and want to play the teams...i'm only using the blackbox because that's the system they're built under....
i know quite a few coaches who have moved back to ranked for several different reasons, one of which is the lack of diversity....
i don't hide the fact that i rail against the lack of diversity quite a bit, but from everything that i see, the bot matchups, my team's opponent history, etc...i see a larger than average amount of one type of team, and that's a big turnoff point for me....
--j |
_________________
| origami wrote: | | There is no god but Nuffle, and Shadow is his prophet. |
  |
|
Thomcat
Joined: Jul 20, 2004
|
| Posted:
Jan 15, 2009 - 14:05 |
|
| funnyfingers wrote: | | +10 TS Elves should beat a bashy team. Elves can pull out TDs with 3 players left on the field. Try that with Chaos, Orcs, Khemri, or Dwarves. |
yeah for one game in a row - but when all they meet is these kind of teams and many off them without any FF they just wont last that long without a lot of luck |
_________________
Og inviterer hermed alle danskere MED godt humør ind i #fumbbl.dk |
|
Unstoffe
Joined: Aug 22, 2004
|
| Posted:
Jan 15, 2009 - 14:18 |
|
Mmm... looking at BillBrasky's Chaos, that is a horrible record indeed, nearly 1 game in 3 vs orcs.
A lot of that is down to the fact that above 200TS or so, there really are a lot of orcs, maybe as much as 50% of the teams. And the killer is : the current (probably unnecessary) bias in the scheduler against mirror matches means that if there are 2 orc teams activated and 2 non-orc in the same bracket, it will tend to make the orc/non-orc matchups every time. Might be a good one thing to modify in the scheduler there.
Looking at his 5-6 game all orc runs, I wonder if the 'reduce the odds of the same race twice in a row' tweak suggested elsewhere might not be good too. |
_________________
British or thereabouts? Check out the White Isle League |
|
JanMattys
Joined: Feb 29, 2004
|
| Posted:
Jan 15, 2009 - 14:42 |
|
Just want to point out a simple fact (dunno if it has relevance with Shadow's numbers, but I'll point it out anyway:
My hellfishes (a team that I am proud to consider non-picky) have played 122 of their 700 games vs orcs. This makes 1 game against orcs every 5,73 games.
This is simply because Orcs are the most popular race around, and you find plenty of orc opponents in R too. So, asking for an "even spread" of race choices in DivB is somewhat futile... even R, where you can "pick" opponents, has more Orc coaches than (say) Norse ones.
My 700 games sample in R, while obviously biased (due to the fact that, to an extent, I pick my opponents and opponents pick me), is meaningful (if nothing else, because 700 is a big enough sample) and clearly shows that it's "ordinary" to have a orc opponent every (say) six games on Fumbbl. The DivB makes it 1 Orc every 4 games... that's not so distant from the "ordinary" average of the site.
Of course, Dorfs and Khemri benefit from DivB more than Orcs because Dorfs and Khemri are dodged more than orcs in regular divisions... but that's a problem of regular divisions more than it's a Blackbox problem, imho.
And yes, I still think that 100% Orcs and 100% dofs DivB coaches have no balls... but that's entirely my point of view  |
_________________
 |
|
CircularLogic
Joined: Aug 22, 2003
|
| Posted:
Jan 15, 2009 - 14:51 |
|
| Unstoffe wrote: | 50/50 win odds doesn't seem to be the aim... currently it is, (% win chance for team A -50) = (% outbashing chance for team B -50) *0.3
Seems to work pretty well to me, barring the TS formula imbalances which a lot of good work is being done on, by SillySod in particular.
@Westerner, yes you're right, those two coaches are clearly experts. If as a less skilled coach you wanted to get to the top 30 then orcs are the way to go. Still worth noting that while presumably there are good coaches out there with orcs, their record even at the highest levels is not as good as agility teams.
I'd guess the reason for the preponderence of bashers is that many coaches see their ideal game as one where they both beat and outbash the opposition, hard to achieve that with elves. I can't see a solution to this that I like more than just living with the current mix of teams though...
Edit - changed 50/40 to 50/50...doh  |
The underlined part is NOT true. Rather a complete misunderstanding how the scheduling works. The BBR factor is PER COACH and should be rather seen as a modifier to the coaches blackbox rating (BR - akin to CR in ranked).
| Ash wrote: | | The more elf team you got, the less bashy opponment you draw... |
Also I want to clarify, that this just means if more elves in total are in the box, every coach is more likely to play vs elves. It should NOT be read as "The more you play elves, the higher chances that you get paired against other elves if available".
There is alot of misunderstanding about the scheduler in terms of BBR and racial diversity. Just wanted to clear that up. |
Last edited by CircularLogic on %b %15, %2009 - %14:%Jan; edited 1 time in total |
|
SillySod
Joined: Oct 10, 2006
|
| Posted:
Jan 15, 2009 - 14:54 |
|
He has a point though, the existence of BBR means that the box is not a 50-50 thing, it has evolved beyond that already. |
_________________
Putting the "eh?" back into Sexeh.
"There are those to whom knowledge is a shield. There are those to whom it is a weapon. Neither view is balanced." |
|
CircularLogic
Joined: Aug 22, 2003
|
| Posted:
Jan 15, 2009 - 14:55 |
|
| SillySod wrote: | | He has a point though, the existence of BBR means that the box is not a 50-50 thing, it has evolved beyond that already. |
The box isn`t a 50/50 thing since the TS-caps were introduced. And the bias BBR introduces is per coach and not per team. So the scheduler thinks that Synn`s pro elves would outbash my claw/RSC chaos while my chaos outscores him. |
|
|
Unstoffe
Joined: Aug 22, 2004
|
| Posted:
Jan 15, 2009 - 15:20 |
|
Circ...
Did I mention BBR?
The calculation of suitability based on the win and bash probabilities is exactly as I stated.
For sure, you can argue that because BR and BBR are per coach, they may well be inaccurate when applied to individual teams and so those win and bash probabilities are themselves inaccurate, you may well be right. Certainly if you're correct in assuming that there is no consideration of racial factor when calculating BBR, or that if there is then the same factor as for BR (i.e. based on winning not bashing) is used, then yes, you will be right about BBR being more of a modifier of BR than a realistic measure of 'bashiness'.
But... I should say that assumption is wrong. It makes no sense to me to use racial factor in the calculation of BR and the win %, but not in BBR and bash %. And if you are using a racial factor, then surely it would make sense to derive a 'bash' factor from the stats, rather than use the win factor as used in [R].
Does that make sense? Obviously some clarification from the big C would be nice. Somehow I doubt he reads every one of the myriad posts about [B] scheduling, balance etc though |
_________________
British or thereabouts? Check out the White Isle League |
|
pac
Joined: Oct 03, 2005
|
| Posted:
Jan 15, 2009 - 15:21 |
|
Hmm, by mentioning the 50-50 thing I was meaning to refer to it in the sense of a guiding target, in contrast to the target of roster diversity (as I don't think you can have both at once), and not in the sense of it having been achieved. I wouldn't claim that the 50-50 target is met atm - although it's a lot closer to being met than any roster diversity target! |
|
|
westerner
Joined: Jul 02, 2008
|
| Posted:
Jan 15, 2009 - 16:46 |
|
@Unstoffe:
I think your point about lots of Orcs above 200 is well taken. In this environment, racial diversity is probably dependent on TS. At 100TS you probably have lots of races well represented and even up to 150 you can expect to encounter agility teams. But beyond that they start to get increasingly rare. It would be interesting to look at diversity at different TS levels and see how they compare.
I'm not sure about the lack of mirror matches being a problem though. I've played 8/147 mirror matches in B, or about 5%. Which feels about right. |
_________________
\x/es |
|
Unstoffe
Joined: Aug 22, 2004
|
| Posted:
Jan 15, 2009 - 16:57 |
|
Yeah, you're probably right about the mirror match thing - just goes to show the danger of arguing from a single case. Taking a wider look at the top 10 or so teams by games played, and the orc teams seem to be getting a perfectly fair share of mirror matches.
Bill's Chaos is simply an extreme example... |
_________________
British or thereabouts? Check out the White Isle League |
|
pac
Joined: Oct 03, 2005
|
| Posted:
Jan 15, 2009 - 17:12 |
|
| westerner wrote: | | I'm not sure about the lack of mirror matches being a problem though. I've played 8/147 mirror matches in B, or about 5%. Which feels about right. |
Depends what your teams have been. Your [B] Dwarf teams have, collectively, 3 games vs Dwarves in 68. That is indeed about 5%, but that is also very low given that Dwarves are quite popular in [B] - Brasky's Chaos above having 19 games against them from 147, or about 12-13%.
Neither sample is large enough to prove anything, naturally, and the prevalence of different teams at different TR levels complicates the whole thing. |
|
|
treborius
Joined: Apr 05, 2008
|
| Posted:
Jan 15, 2009 - 18:18 |
|
| Unstoffe wrote: |
(...)
But... I should say that assumption is wrong. It makes no sense to me to use racial factor in the calculation of BR and the win %, but not in BBR and bash %. And if you are using a racial factor, then surely it would make sense to derive a 'bash' factor from the stats, rather than use the win factor as used in [R].
Does that make sense? Obviously some clarification from the big C would be nice. Somehow I doubt he reads every one of the myriad posts about (B) scheduling, balance etc though |
you're spelling out, what i've been thinking about quite a bit - didn't get
involved into the fundamentals behind B until about 4 weeks ago and tried to
catch up on some discussions that have been going on since mid-2008, but i couldn't
really find any proper hint by Christer or someone who knows about how exactly that
BBR is calculated - i've seen someone arguing, that it would have to incorporate
racial modifiers and that makes a lot of sense to me...
...i.e. i'd expect BBR to account for the relative bashiness of a coache's playing-
style regardless of the race - meaning you could easily get a BBR playing only
Pro-Elfs which would match a BBR of someone playing Khemri if you only played equally
"bashy" as compared to the average bashiness of those races.
maybe this has been answered before, but does anyone know any details whether this is a
correct assumption or not?
btw: is your avatar meant to oppose pac's or what's the deal here?  |
|
|
Unstoffe
Joined: Aug 22, 2004
|
| Posted:
Jan 15, 2009 - 18:26 |
|
Heh, I put that up when I was about to play him in the WIL as I recall. Should probably get around to changing it...
Maybe to the 'post consumption' pair of disembodied eyes that ran about |
_________________
British or thereabouts? Check out the White Isle League |
|
westerner
Joined: Jul 02, 2008
|
| Posted:
Jan 15, 2009 - 18:32 |
|
| treborius wrote: | ...i.e. i'd expect BBR to account for the relative bashiness of a coache's playing-
style regardless of the race - meaning you could easily get a BBR playing only
Pro-Elfs which would match a BBR of someone playing Khemri if you only played equally
"bashy" as compared to the average bashiness of those races.
does anyone know any details whether this is a
correct assumption or not? |
Circ understands this better than I do, but I believe BBR (being purely per coach) does not correct for race or team. Team based BBR has been requested numerous times previously. |
_________________
\x/es |
|
|
|
| |
FUMBBL HAIKU'S
Search
