Dos mil quinientos

[Trenton.]

What can I say? It's been a.... Horrifying four years.

 

I'll see you all again in 2500 more.

 

Yes. I mean years. I'm immortal.

 

 

Can only hope I'm not that same n00b that annoyed the crap out of everyone back then If I'm not, It's only because I hung around people as awesome as you.

 

 

Although we all know that I'm the awesomest one here currently most of the time forever.

[Goldengirl]

[Triumph]

I think Spiderweb is suffering from milestone thread inflation. As the number of milestone threads for increasingly arbitrary and trivial numbers (since when did anyone care about 2500???) multiplies, the value of such threads is diminished. This devalued thread, for example, earns 37.4% less approbation. I call on the chairman of the postcount reserve to take decisive action against this threat!!!

 

If nothing is done...

 

#occupypostcount

[Callie]

Enhorabuena! But can you swim with it?

 

I think Spiderweb is suffering from milestone thread inflation. As the number of milestone threads for increasingly arbitrary and trivial numbers (since when did anyone care about 2500???) multiplies, the value of such threads is diminished. This devalued thread, for example, earns 37.4% less approbation. I call on the chairman of the postcount reserve to take decisive action against this threat!!!

 

If nothing is done...

 

#occupypostcount

If the action is too extreme, we might fall into a period of postcount deflation. I suggest that the chairman institute a policy of quantitative post easing within the next few minutes.

[Ohsky]

Congratulations.

[The Reverend]

We of the 99% demand postcount milestone threads for all, not just the fat-cat wall-street high-post-counters!

 

I hearby occupy this thread, and all future milestone threads, in the name of all posters with less than 1000 posts, and declare it OUR milestone thread!

 

340 posts, FTW!!!

[Ohsky]

23 posts

[nikki.]

"Look at me, I can't even be bothered to write my own message! I am truly the more optimal spammer!" - Nikki

[Alorael at Large]

That's technically untrue. You could be bothered to write your own message, just not more than once.

 

—Alorael, who can't be bothered to write a signature for this post. Okay, that's obviously not completely true, but it's true insofar as he

[Randomizer]

Congratulations

 

I suggest a mass banning to reduce milestone topics.

[Actaeon]

Given that the mods are the most frequent offenders, that seems unlikely to help.

 

I would say that unusual milestones are acceptible so long as you don't also partake of the usual ones. I, for instance, ignored 1000 and went for 1024. I am assuming I am anything but a high water mark for restraint.

[A less presumptuous name.]

This thread makes me sad. I think I've been here for twice as long as Trenton… My posting frequency has been pretty minimal lately. (You're welcome.)

[keira]

Given that the mods are the most frequent offenders, that seems unlikely to help.

sounds like we have a volunteer

 

also obligatory flutteryay

[BMA]

The milestone threads are okay, it's thinking up witty replies to them that's taxing.

Coz u can't just say congratulations - that would be Spam

Besides, it would just be an excuse to move closer to your own milestones.

 

Happy 2500!!! Have a drink.

[Xaiya]

And I'm sure all 2500 posts were full of relevant and interesting content.

[Student of Trinity]

Congratulations, Trenton!

 

This thread is not just a milestone for one member, but a milestone for the boards in general. We can let it mark an era: the rise of a class of members with postcounts in the few thousand. A middle class is emerging. Its brash ways may make Old Money sniff, but it's here, and its post counts are rising. Soon it will wield more and more power. Society will change. More and more, the new immigrants who arrive here, with nothing but dreams in their post counts, will look up not to the few old folks who came over on the Mayflower, but to the bustling middle class members of the community.

 

Upward mobility. It's the Spiderweb Way!

[Trenton.]

The milestone threads are okay, it's thinking up witty replies to them that's taxing.

Coz u can't just say congratulations - that would be Spam

Besides, it would just be an excuse to move closer to your own milestones.

 

Happy 2500!!! Have a drink.

I would, but I'm under the drinking age of consent Huzzah!

[Student of Trinity]

I'm under the drinking age of consent.

So it's okay as long as you don't drink anything older than you are?

 

Actually, trying wine or beer at home, which is legal in a lot of places long before you're allowed to buy the stuff for yourself, is a good idea. In fact, speaking as a parent, I'd recommend deliberately getting drunk at some point, at least a bit, while safe at home with family. Experience doesn't make you less affected by alcohol, but it teaches you to stop drinking it before it feels like you need to stop, because that's when you do need to stop. I'd rather have my children learn that at home than out in some bar. Then they can do what they want, once they're grown, but at least they'll know what's going on.

[Lilith]

Actually, trying wine or beer at home, which is legal in a lot of places long before you're allowed to buy the stuff for yourself, is a good idea. In fact, speaking as a parent, I'd recommend deliberately getting drunk at some point, at least a bit, while safe at home with family. Experience doesn't make you less affected by alcohol, but it teaches you to stop drinking it before it feels like you need to stop, because that's when you do need to stop. I'd rather have my children learn that at home than out in some bar. Then they can do what they want, once they're grown, but at least they'll know what's going on.

 

i'm told Studies Show That giving your kids access to alcohol increases their chance of being problem drinkers later in life, but on the other hand even if that's accurate it's a pretty broad and vague claim on its own. there's a difference between giving your kid a glass of wine on their birthday and not caring if they drink all the beer in your fridge

 

for what it's worth my parents introduced me to alcohol very young, giving me a sip of wine with dinner on special occasions from about age 6 onwards, and for all my jokes i'm a pretty moderate drinker. i might get through a six-pack of beer or one bottle of wine in a typical week. meanwhile my sister basically doesn't drink at all (because she married a recovering alcoholic and so neither of them want to have alcohol in the house) and my brother abused alcohol for years, so i guess all i can say from my personal experience is that my parents' strategy had mixed results

 

also if you are going to get drunk around your parents don't do it when they have guests over, just trust me on this

[adc.]

Although we all know that I'm the awesomest one here currently most of the time forever.

Just like me

------

Congratulations anyway.

------

-Test Subject 8552

[Kelandon]

I don't drink, but I'd like a +1 from this thread. Hey, I just got it!

[Student of Trinity]

I don't know these studies about childhood exposure to alcohol, so maybe they're really great, but in general I'm awfully skeptical of studies that 'show' that something causes something else. Lots of complicated things would be great to understand better, so people have to study these things. It's really hard to study these things, so these fields tend to develop methodological standards that cut people some slack, so that they can actually get out and do something instead of just pounding their heads against the wall. And maybe over many decades the field does get somewhere. In the meantime, though, you tend to get peer-reviewed, state-of-the-art research that simply isn't really very good. The subject is just too hard, so the standards for what counts as progress are too low.

 

The poster child for this disease is the 95% confidence level as a standard for significance. In a lot of fields, that's considered a gold standard. In fact, it's nearly worthless. It means that every twentieth study gets published, with significant results, even if there is absolutely nothing there but noise. There are a lot of important topics on which there are a lot more than twenty studies being done at any given time. And even a single study may comb through its data looking for at least twenty different possible effects. So somebody is always 'showing' something. Then when somebody else tries to replicate the findings, it goes away — 95% of the time, at any rate.

 

Maybe persistence will pay off, and you just have to wait while all the flashes in pans fizzle out. Or maybe no science really gets going until it gets some solid theory under it, so that you can figure out how to control experiments well enough to get far better than 95% confidence. Add a few more sigmas and you do start to approach practical certainty. But there's a threshold of understanding you have to reach, I suspect, before you can start to push those sigmas further.

[Alorael at Large]

From what I know of the studies, they show correlation and not causation, and they tend to show that drinking substantial amounts on a regular basis correlates with more problematic use of alcohol later in life. That's good to know, but it's not surprising, and I don't think the studies addressed what happens when you let your teengare have a glass of wine or a beer with dinner, or encourage them to get drunk once or twice, safely, at home.

 

—Alorael, who thinks the key is to get the kids hooked on high quality alcoholic beverages early. If they've learned to enjoy fine microbrews, expensive wines, and top shelf liquor they'll have trouble chugging the abundant swill in college. They'll be too poor to afford binges. You can protect by inculcating rarified tastes! Think of the children. Buy the good stuff.

[Dikiyoba]

Congratulations, Trenton.

 

Dikiyoba doesn't drink alcoholic beverages. Dikiyoba just can't stand the taste.

[Earth]

S P A M M E R

 

gratz anyways.

[Lauren CW]

This thread makes me sad. I think I've been here for twice as long as Trenton… My posting frequency has been pretty minimal lately. (You're welcome.)

I laugh at this.

Not you, just this statement.

[Dikiyoba]

The poster child for this disease is the 95% confidence level as a standard for significance. In a lot of fields, that's considered a gold standard. In fact, it's nearly worthless. It means that every twentieth study gets published, with significant results, even if there is absolutely nothing there but noise. There are a lot of important topics on which there are a lot more than twenty studies being done at any given time.

You're putting the cart before the horse. The goal of research is not to eliminate the noise but to find as much of the signal as possible. The 95% confidence level is a compromise between reporting false positives and reporting false negatives. Some scientific disciplines apparently do better with a higher confidence interval that produces less false positives but more false negatives. Some scienfitic disciplines do better with a lower confidence interval that produces more false positives but fewer false negatives. (Also, just because 95% is the gold standard for a field doesn't mean that all of the studies in that field will be at that confidence interval. Many studies will be able to reach a higher confidence level, sometimes by a lot. 95% confidence is just the lowest you can go.)

 

Some noise mixed in with the signal is inevitable. That's why research isn't accepted as fact until its results have been duplicated multiple times. When there are 20 studies on the same thing and 19 say one thing but 1 says the other, it suddenly becomes much easier to separate the signal from the noise. Much better to have the process of finding the signal be muddled and confusing than to have standards so impossible to fulfill that no one finds any signal at all.

 

Dikiyoba.

[Ohsky]

ppl, what are you thokin about

[A less presumptuous name.]

http://en.wikipedia.org/wiki/Statistical_significance

[Ohsky]

not a very humorous response

[The Reverend]

The problem is not the 95% confidence in and of itself, it's the fact that many people (including many researchers) do not understand statistical confidence levels. They think that 95% confidence means that there is a 95% chance that a finding is correct - which is a false interpretation. It really just means that there is a 5% chance of a false hypothesis being confirmed giving that level of testing. As SoT pointed out above, this means that if everyone tests their hypothesis to a 95% confidence level, then 1 out of 20 false hypotheses will appear to be true. And if there are a lot more false hypotheses being tested than true hypotheses, then the confirmed false hypotheses may equal, or even outnumber, the confirmed true hypotheses.

 

If more people understood this, they would not accept studies as true until they had been confirmed by multiple independent studies. Unfortunately, most lay people, and even many researchers, aren't familiar enough with these statistical concepts.

[nikki.]

not a very humorous response. Derrailing might be in the line.

 

Please don't. Topic derailment is a beautiful, natural phenomenon. Any attempt to deliberately derail topics leads to posts like this one; husks of posts, completely devoid of merit or humour.

[Alorael at Large]

I'd avoid using confidence when talking about p values. It sounds like you're referring to confidence intervals, which are related but different.

 

To my mind, calling something with p = 0.05 95% likely to be true is reasonable. You're saying that there's a 95% chance that your measurements are not due to chance, and there's a 5% chance that they are. If you make 20 measurements, all of which reject the null hypothesis with p = 0.05, there's on average going to be one of those in which the measurements were a statistical fluke and the null hypothesis is true.

 

In theory, this means that large numbers studies with p values close to the threshold are going to produce large numbers of false results. In practice this is probably true, but I think it's less significant than other sources of bad study data, including misuse of statistics (often by accident), bad data collection, lousy study design, confounding, bias, and all those other lovely human and mathematical flaws introduced into research.

 

—Alorael, who also notes that most scientists take studies with p = 0.04 with a grain of salt. The bigger danger is the fact that amongst thousands of papers published with p = 0.001 there are still going to be some false ones. Human minds aren't good at processing very large numbers of very small things correctly.

[Ohsky]

Please don't. Topic derailment is a beautiful, natural phenomenon. Any attempt to deliberately derail topics leads to posts like this one; husks of posts, completely devoid of merit or humour.

 

ok

[Triumph]

Please don't. Topic derailment is a beautiful, natural phenomenon. Any attempt to deliberately derail topics leads to posts like this one; husks of posts, completely devoid of merit or humour.

 

^ This. You see, here at Spiderweb, we prefer 100% organic, all-natural tangents, produced by free labor without the use of pesticides - none of that artificial, genetically modified derailment you get on some websites. This is the real deal. Granted, being 100% organic may increase our carbon footprint, but that's another story.

[Ohsky]

fine, you´re the law, shapers, hehehe

[Student of Trinity]

The problem is not that 20 studies get published with one contradicting the other 19, and so you can see the noise. It's that the 19 don't get published, because they don't have a finding. They probably don't even get submitted for publication. So the entire field has a strong bias towards over-reporting flukes, even when everybody is actually following their state-of-the-art methodology as well as possible.

 

There are subtler issues involved than just whether your findings are flatly false or correct. For example, often the way you control for irrelevant factors is to make statistical tests. You conclude that your survey about race and education has controlled for income, for example, by showing that income has no significant effect after you've applied your controls. But when significance is a weak standard, this means that a lot of controls are really inadequate. So even if your results are not exactly a fluke, what they seem to mean may be. Statistical noise is a great place to hide your own bias. In effect, you wait until the flukes confirm your assumptions. It's much harder to avoid this than one might think. Humans are instinctively primed to see patterns in noise.

 

Whatcha gonna do, though? You can't just throw up your hands and decide we'll never know. The problem I see is more that there's some inappropriate physics envy going on, with people dressing up research that is really just qualitative or speculative as quantitative and rigorous. Some of the social sciences should back off a bit from trying to be scientific. On important enough issues, qualitative or speculative research is worth having. Fight the battle honestly, instead of with fake numbers.

 

And on the other hand, studies that do manage to achieve really high levels of significance should be trumpeted really loudly, even if their conclusions are boring. They're not just a bit better than the other studies: they're the only real ones. The reason they don't get as much acclaim as they deserve is simple, in my opinion: if you over-advertise your premium product, you'll lose all your bread-and-butter sales. I mean publications.

[Lilith]

and now you can all see why scientists drink so much in the first place

[Rowen]

I grow my own organic derailment in the backyard under a camouflage tarp so the Admins don't see it. It use it to treat my postinitis syndrome, I promise.

[Trenton.]

Math and beer... Are all future milestone topics going to be like this?

[Student of Trinity]

Yes. If you want a picture of the future, imagine some beer spilling on some math, forever.

[BMA]

And don't forget the pokemon and ponies.

 

-CC-

[Sudanna]

As if Lilith would be so pedestrian as to drink beer. Really, Trenton.

[Lilith]

beer isn't my absolute favourite drink but now that i'm living alone it's pretty convenient, because believe it or not i don't tend to want an entire bottle of wine in one sitting

[Ohsky]

wine, agh

[Student of Trinity]

In vino veritagh.

[Alorael at Large]

Since when does wine come in bottles and not boxes?

 

—Alorael, who doesn't drink anything that can't recall the distant nostalgia of the juice box or at the very least the sippy cup.

[Ohsky]

In vino veritagh.

 

in vino black [censored] and diarrheah

[keira]

beer isn't my absolute favourite drink but now that i'm living alone it's pretty convenient, because believe it or not i don't tend to want an entire bottle of wine in one sitting

awww, living alone

 

You should get a cat. Cats are nice

in vino black [censored] and diarrheah

Now now sugarcube, it's okay to not like things but let's be nice about it, maybe provide some constructive criticism about the subject in hand.

 

I expect a fifty-page treatise on the matter on my desk by tomorrow morning, young man.

[Sudanna]

Cats that drink your booze are not nice, and probably won't last very long,

 

You could always join the growing Coloradoan Calref Commune.