110 coaches online • Server time: 22:09
Index Search Usergroups Profile
Log in
Recent Forum Topics goto Post Pink Slip Death Matc...goto Post cdassak must be stop...goto Post TOP25 2514 Rankings
 
FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   ProfileProfile 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
VoodooMike



Joined: Nov 07, 2010

Post   Posted: Dec 04, 2011 - 14:24 Reply with quote Back to top

Carnis wrote:
The graphs are quite a bit more descriptive than the final numbers you've presented. For instance, the amazons are only playing TV 1000-1500 basicly, so they are sweetspotting.. Lizardmen in your graph are mostly in the 1000-1600 range as well.

No, the graphs just feel more descriptive than numbers, but as I've said in the past graphs are mostly for folks who like pretty pictures and imagine they mean more than they do - mostly because they don't know what to make of numbers in general.

The graphs, for example, don't tell you how games are distributed across TVs. Most games are played closer to 1000 than to 2000, so races that are stronger in the ranges where most games are played, will certainly be more successful than those in ranges that are infrequently played.

Likewise, you might want to look up what the R-squared part means before you take any "best fit" lines as evidence of something important, even with CI lines thrown in.

Carnis wrote:

Maybe you could display your average win% graphs for each race/TV-range overlayed in groups of 6? 24 would probably be a bit unreadable..

Look harder at the Y axes and then think about it.

Carnis wrote:
Also there are other trends beyond the win% going sky high for high tv chaos, which you already presented. Dark Elves seem to do better high tv as well.

Are there really trends? What do you base that on... eyeballing best fit lines?

Carnis wrote:
Dwarves also go steeply below 40% at the higher TV ranges, but due to sweetspotting in the final numbers they end up a tier1 race despite their obvious weakness in this ruleset. So there's obviously some need for interpretation into the numbers..

Is there? Again, win percentages deal with games actually played, not theorized games that would happen if everyone were playing at TVs they aren't playing at routinely, or the results of those games would apply more heavily to overall wins.

The data was all games from when LRB6 was introduced to FUMBBL, until sometime in October (I believe.. we'd have to go back and see what hitonagashi said the cut-off date was). So, that's about 9 or 10 months of blackbox games (around 50,000 games).

Do you feel there will be a change in the general TV range at which most games will be played? If so.. what do you base that on? If not, then the ranges at which games are being played regularly are factored into the weighting of the win percentages, and are thus quite valid - no alternate interpretation is really necessary.
VoodooMike



Joined: Nov 07, 2010

Post   Posted: Dec 04, 2011 - 14:41 Reply with quote Back to top

koadah wrote:
A doctor (expert) cannot actually feel what the patient is feeling. The doctor may be seeing the person for the first time where as the patient has lived with that body all their life.

A gynecologist knows more legitimate information about the female reproductive system than the typical woman does, even if the doctor is not a woman himself. I'm sure women know plenty about what it feels like to have a uterus, but it doesn't give them any inherent technical knowledge related to the biology.

Fields like medicine are based on experimentation and data, as well as statistical analyses. You come in with a headache and the expert takes that symptom and compares it (either mentally based on past knowledge of the spread, or looks it up) with things it could be, and how frequently the causes manifest, and works their way through the problem that way. Maybe you have a super-rare form of brain cancer! More likely, you just need an asprin. If you guessed cancer you might be right, but most likely you're wrong, regardless of your intuition.

If it turns out you have cancer that doesn't validate intuition, it just means a broken clock is still right twice a day.

It's also why they call gambling a tax on stupidity.

koadah wrote:
Scans don't show up everything.

They show up more than psychic powers, which is basically what you're using when you trust your intuition.

koadah wrote:
The person who accepts the reassurance of the experts and their data is often diagnosed too late and dies. Whereas the person who kicks up a fuss because of their gut feel/intuition can get their diagnosis on that 3rd or 4th opinion.

There goes that broken clock again. Kicking up a fuss over minor symptoms is a good way, I might add, to have future complaints taken even more lightly. Something about a boy and a wolf.

koadah wrote:
Of course the government accountants will say don't run those expensive tests because the data shows that 99% of the time it's nothing. But that's no comfort to the 1%.

Actually, they say "don't run expensive tests unless there are symptoms suggesting there is a reason to do so". People who live their lives based on "what if" scenarios related to 1% aren't really living lives at all... certainly not rational ones.

koadah wrote:
Question everything.

Why should I?
Carnis



Joined: Feb 03, 2009

Post   Posted: Dec 04, 2011 - 14:48 Reply with quote Back to top

VoodooMike wrote:
Carnis wrote:
The graphs are quite a bit more descriptive than the final numbers you've presented. For instance, the amazons are only playing TV 1000-1500 basicly, so they are sweetspotting.. Lizardmen in your graph are mostly in the 1000-1600 range as well.

No, the graphs just feel more descriptive than numbers, but as I've said in the past graphs are mostly for folks who like pretty pictures and imagine they mean more than they do - mostly because they don't know what to make of numbers in general.

Your numbers are just one number, out of context, win % over ALL games played at ALL tvs divided between race & race context ignored, tv ignored, just take what goes in and take the simplest descriptive statistic available (average win%). This will inevitably lead to overinterpretation, simplicifation (title: Black box race tiers).

Your pretty pictures have this info divided by TV range, providing more useful info for us coaches who play TVs other than 1000 too. The R2 values are low, yet even so the best fit curves are more informative than any overall win% indifferent of TV range.

Now they are pretty hard to compare as they are all in different X & Y axis', which is why pretty pics of more than one race overlayed would infact be useful in my opinion.

We all knew that dwarves are tier1 in tv1000.

Anyway, what Im trying to say is your numbers may not be describing anything useful.

Anyway, gratz on being completely incapable of having any sort of civil discussion online.
Carnis



Joined: Feb 03, 2009

Post   Posted: Dec 04, 2011 - 14:51 Reply with quote Back to top

koadah wrote:
Scans don't show up everything.

The person who accepts the reassurance of the experts and their data is often diagnosed too late and dies. Whereas the person who kicks up a fuss because of their gut feel/intuition can get their diagnosis on that 3rd or 4th opinion.

That's a horrible life philosophy Wink. Dont go to a doctor, if you're gonna trust your own intuition anyway Wink.
koadah



Joined: Mar 30, 2005
Location: London, UK

Post   Posted: Dec 04, 2011 - 15:27 Reply with quote Back to top

Carnis wrote:
koadah wrote:
Scans don't show up everything.

The person who accepts the reassurance of the experts and their data is often diagnosed too late and dies. Whereas the person who kicks up a fuss because of their gut feel/intuition can get their diagnosis on that 3rd or 4th opinion.

That's a horrible life philosophy Wink. Dont go to a doctor, if you're gonna trust your own intuition anyway Wink.


When did I say don't go to a doctor? I just said don't trust em. Wink

How do you get a 3rd or 4th opinion without going to a doctor?

Sorry dude that's life.

If you cannot fix your own car you will need a mechanic. If he charges you a fortune for doing nothing...

_________________
Image
CPOMB Nerfs - Open [L]eagues - Brand new teams only! - Places still free
Carnis



Joined: Feb 03, 2009

Post   Posted: Dec 04, 2011 - 15:44 Reply with quote Back to top

koadah wrote:
Carnis wrote:
koadah wrote:
Scans don't show up everything.

The person who accepts the reassurance of the experts and their data is often diagnosed too late and dies. Whereas the person who kicks up a fuss because of their gut feel/intuition can get their diagnosis on that 3rd or 4th opinion.

That's a horrible life philosophy Wink. Dont go to a doctor, if you're gonna trust your own intuition anyway Wink.


When did I say don't go to a doctor? I just said don't trust em. Wink

How do you get a 3rd or 4th opinion without going to a doctor?

Sorry dude that's life.

You didn't.

I said don't go.

Why not then? Well, since you won't take their answer anyway until it fits your gut feeling, better save your money & selfdiagnose.

Chances are the 4th and in your view correct opinion ends up being a guy who agrees with you to be nice Wink.
divino77



Joined: Jun 23, 2011

Post   Posted: Dec 04, 2011 - 16:39 Reply with quote Back to top

[A] [B] [C] [D] [E] [F] [G] [H] [I] [J] [K] [L] [M] [N] [O] [P] [Q] [R] [S] [T] [U] [V] [W] [X] Tot
Amazon [A] · · · · · · · · · · · · · · · · · · · · · · · · 0
Chaos [B] 1 1 4 · 1 · · 1 · · 1 · · 1 · 1 · 2 · 1 · · · 1 15
Chaos Dwarf [C] 3 2 1 1 1 1 · · · · 1 · 3 · · · · · · · 1 · · · 14
Chaos Pact [D] · · · 1 · · 1 · · · · · 1 · · 1 · 1 · 1 1 · · · 7
Dark Elf [E] · · · · · · 1 · · · 2 · · · · · · 3 · 1 · · · · 7
Dwarf [F] · · · · · · · · · · · · · · · · · · · · · · · · 0
Elf [G] · 2 2 · · · · · · · · · 1 · · · · 2 · · · 1 · 1 9
Goblin [H] · · · · · · · · · · · · · · · · · · · · · · · · 0
Halfling [I] · · · · · · · · · · · · · · · · · · · · · · · · 0
High Elf [J] · 5 1 1 · 1 1 · · 2 3 · · · 1 · · 1 · · · · · · 16
Human [K] · 1 2 · · · · · · · · · · 1 · 2 · 1 1 1 · · · · 9
Khemri [L] · · · · · · · · · · · · · · · · · · · · · · · · 0
Lizardman [M] · · 3 2 · 1 1 · · · · 1 · 1 · 1 1 2 · · · · · · 13
Necromantic [N] · · · · · · · · · · · · · · · · · · · · · · · · 0
Norse [O] · 4 3 2 · 3 · · · 2 1 1 1 1 · · 1 1 · 1 1 · 1 · 23
Nurgle [P] · 4 · 1 · 3 · 1 · · 1 3 · 2 2 · 1 · · · · · · · 18
Ogre [Q] · · · · · · · · · · · · · · · · · 1 · · · · · · 1
Orc [R] · · · · · · · · · · · · · · · · · · · · · · · · 0
Skaven [S] · · 2 · · · · · · · · 1 1 · 1 · · · · · · 1 · · 6
Slann [T] · · · · · 2 · · · · · · 1 2 · · · 1 · · 2 · · · 8
Undead [U] · · · · · · · · · · · · · · · · · · · · · · · · 0
Underworld [V] · · · · · · · · · · · · · · · · · · · · · · · · 0
Vampire [W] 1 3 4 2 2 2 1 · 1 · · 1 · · 2 2 · · 1 1 1 · · 2 26
Wood Elf [X] 3 4 3 2 1 1 · · · · 3 1 1 · · · · 3 1 1 1 1 · · 26
Total 8 26 25 12 5 14 5 2 1 4 12 8 9 8 6 7 3 18 3 7 7 3 1 4 198

_________________
Nothing is only something if it is contained otherwise it is eveything.
koadah



Joined: Mar 30, 2005
Location: London, UK

Post   Posted: Dec 04, 2011 - 16:53 Reply with quote Back to top

Carnis wrote:
koadah wrote:
Carnis wrote:
koadah wrote:
Scans don't show up everything.

The person who accepts the reassurance of the experts and their data is often diagnosed too late and dies. Whereas the person who kicks up a fuss because of their gut feel/intuition can get their diagnosis on that 3rd or 4th opinion.

That's a horrible life philosophy Wink. Dont go to a doctor, if you're gonna trust your own intuition anyway Wink.


When did I say don't go to a doctor? I just said don't trust em. Wink

How do you get a 3rd or 4th opinion without going to a doctor?

Sorry dude that's life.

You didn't.

I said don't go.

Why not then? Well, since you won't take their answer anyway until it fits your gut feeling, better save your money & selfdiagnose.

Chances are the 4th and in your view correct opinion ends up being a guy who agrees with you to be nice Wink.



Well, I've seen enough doctors and medical staff **** up to not trust them.

We've had one person struck off for continually ****ing up. But it took years and they were practising all that time without any improvement.

I've seen medical staff standing around staring at the a person that they know has ****ed up but not daring to break ranks.

I've looked into the eyes of someone who knows that they have crippled someone for life due to a bad decision.

I don't know what job you do but my job is fairly complicated. We often have different ideas on what do to after looking at the same data. People call us experts but sometimes we're going to wrong.

Sometimes we're going to know the correct solution but the budget won't run to it.

That's where VoodooMike is right. It's not cost effective to run 100 expensive tests when you're only likely get 1 positive. But people who put too much trust in the experts and their data get fobbed off especially when resources are short.

Sorry to be selfish but if they are only going to run so many tests or offer so many places there times when it's going to be worth your while to make sure that they pick you.

_________________
Image
CPOMB Nerfs - Open [L]eagues - Brand new teams only! - Places still free
koadah



Joined: Mar 30, 2005
Location: London, UK

Post   Posted: Dec 04, 2011 - 17:15 Reply with quote Back to top

Carnis wrote:

I said don't go.

Why not then? Well, since you won't take their answer anyway until it fits your gut feeling, better save your money & selfdiagnose.

Chances are the 4th and in your view correct opinion ends up being a guy who agrees with you to be nice Wink.


Oh and yes we have been correct on at least two self diagnosis that the so called experts had been denying for years.

I know the doctors won't want to hear this but thank **** for the internet.

_________________
Image
CPOMB Nerfs - Open [L]eagues - Brand new teams only! - Places still free
VoodooMike



Joined: Nov 07, 2010

Post   Posted: Dec 04, 2011 - 21:08 Reply with quote Back to top

Carnis wrote:
Your numbers are just one number, out of context, win % over ALL games played at ALL tvs divided between race & race context ignored, tv ignored, just take what goes in and take the simplest descriptive statistic available (average win%). This will inevitably lead to overinterpretation, simplicifation (title: Black box race tiers).

The numbers are perfectly within the context of blackbox as they deal with games actually played in the specified division. It doesn't matter if, say, khemri are king at TV 3000+ while they perform very poorly at 1000 to 1200. Why? Because the games that are actually played in this division are far, far more frequently in the range of TVs that khemri are not king at, which is reflected in their overall win percentage.

Carnis wrote:
Your pretty pictures have this info divided by TV range, providing more useful info for us coaches who play TVs other than 1000 too. The R2 values are low, yet even so the best fit curves are more informative than any overall win% indifferent of TV range.

Actually no, some of the R-squared values are quite high, all things considered. That said, the only thing the graphs really do at all is possibly suggest to players of a given race whether they should try to keep their TV down, maximize it, or strive to stay in a certain broad range. That said, without knowing this information people have been winning and losing games just fine with the races in question, so it seems to pan out the way it has been reported even without coaches knowing these supposed "trends".

Carnis wrote:
Now they are pretty hard to compare as they are all in different X & Y axis', which is why pretty pics of more than one race overlayed would infact be useful in my opinion.

Well have at it, then - the data I used is readily available from the stats thread in general. Such a thing would be an entirely different area of information, however, unrelated to discussing racial tiers, because it doesn't much matter whether a given race is only good at TV ranges where games actually happen.. that's where the games happen!

Carnis wrote:
Anyway, what Im trying to say is your numbers may not be describing anything useful.

It's pragmatic information rather than theoretical information, so yes, I know it won't be of use to a certain segment of the population. It does, however, tell people who might be considering making a new team in blackbox, and who wish to do well in it, which races are best suited to winning in that division.

Carnis wrote:
Anyway, gratz on being completely incapable of having any sort of civil discussion online.

Hey, if you can justify your position I'd be less sardonic... but you just keep saying that how often teams are actually winning is irrelevant. I'm baffled as to what you believe IS relevant.
blader4411



Joined: Oct 18, 2009

Post   Posted: Dec 04, 2011 - 21:45 Reply with quote Back to top

You know what else is relevant?

Playing some games on FUMBBL before posting threads like this.
Seeing as you have 0 games logged, we can assume you to either be inexperienced, or just a blatant attempt at trolling.

-Blader
koadah



Joined: Mar 30, 2005
Location: London, UK

Post   Posted: Dec 04, 2011 - 22:01 Reply with quote Back to top

blader4411 wrote:
You know what else is relevant?

Playing some games on FUMBBL before posting threads like this.
Seeing as you have 0 games logged, we can assume you to either be inexperienced, or just a blatant attempt at trolling.

-Blader


It's not relevant at all. The data is what it is.

I have the data. So does does Dode and Hitonagashi and anyone else who chooses to pull it out of the Fumbbl API.

If you choose to disregard the message because of who the messenger is that is your choice. Wink

_________________
Image
CPOMB Nerfs - Open [L]eagues - Brand new teams only! - Places still free
dode74



Joined: Aug 14, 2009

Post   Posted: Dec 04, 2011 - 22:08 Reply with quote Back to top

Can I call it or what? Wink
http://fumbbl.com/index.php?name=PNphpBB2&file=viewtopic&p=489396#489396

Can anyone shed some light on what divino77 posted? It looks a bit like a one-time pad, but I doubt that people are using the FUMBBL boards to pass encryption keys Razz
Garion



Joined: Aug 19, 2009

Post   Posted: Dec 04, 2011 - 22:16 Reply with quote Back to top

It was just this - http://fumbbl.com/FUMBBL.php?page=coachstats&coach=131154&division=10

games he has played in the box, with and against, every can see this by going to their coach page, then statistics, then selecting a division from the drop down menu, not sure why he posted it mind Confused

_________________
Unnerf Master chef 2014!!!
JimmyFantastic



Joined: Feb 06, 2007

Post   Posted: Dec 04, 2011 - 23:06 Reply with quote Back to top

Well even though I can see the point of all the players, to be fair to VoodooMike this does show the entire story. It shows that Zons are vicious minmaxers, it shows that Dwarfs are great at low tv and so are lizzies etc. It shows a lot of stuff but it shows it indirectly.

_________________
Garion wrote:
Jimmy is always right Razz

Purplegoo wrote:
Jimmy is, indeed, always right.

PurpleChest wrote:
Jimmy is, as always, right again.
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
Powered by PNphpBB2 1.2 © 2003 PNphpBB Group
Credits