Whats The Forums Opinion Of Mad Dog?

Seeing there is quite an argument going on here about the validity of inferences drawn based on a sample size of one, I was wondering just how many pieces of KFC chicken I have to eat before I can say with 99% confidence that I like them.
 
Cobalt, I don't know exactly how far nan's knife bent, or whether it chipped out or not, because I threw it out shortly afterwards (it was in pretty bad shape). Cliff may still have the info in his notebook though.

I did, however, check out steak knives at Wal-mart and you can get a complete set of 11 knives, a pair of scissors and a wooden display block to store them in for under $15. I'd have to give edge retention and handle ergonomics to the Tusk though. The edge will probably chip out easily, but so will a Tusk. The steak knives will resist corrosion much better. And, ofcourse, you can't dig or pry with either.
 
WOW I think This is the most posts i have ever seen. I think they are a great CO.and Geat Knife Builder ...

------------------
Thats a Good ONE
 
Cliff, how could you abuse an old lady's hardware like that? Next time bring your battle mistress will ya.

Puncture, next time keep a set of Busse mean street kitchen knives on hand, they won't let Cliff down.
 
Hoodoo :

But I don't think we are talking about life and death issues unless you are talking about ruining some knifemaker's life [which could be the result of a type I error]. Then perhaps p < 0.01 or p < 0.001.
Click to expand...

Yes, that is exactly why I would always allow an examine and replacement from the maker.

Much of my research is small sample work and I usually work with n between 15 and 30
Click to expand...

What a different perspective, I would call that a huge sample. I can't take data sets with 30 points as I would have trouble keeping the equipment stable for that long (lamps would burn out etc. ).

I can't think of anything that I accept with the 100% confidence that you are espousing based on n=1 or n=2.
Click to expand...

If you shot a 12 guage at a piece of newsprint and saw the results would you not be confident at the 5% level that a sheet of such paper would not be a suitable material for a bulletproof vest? You have to consider the variance against the mean seperation that you are trying to evaluate. In this particular case resistance from one piece of paper to another and the penetration from one shot to the next as compared to the difference in resistance from that one piece of papar to what would be necessary to stop the bullets. If it is not remotely possible for the variance to cover the spread then you don't have to calculate the p value. The only time rigerous statistics are used is when you can't eliminate the correlation at a glance as it is in the critical region.

In these papers are you the primary author? I'd be interested in reading some.
Click to expand...


I would be the one would did all the analysis and interpretation and data collection (except in very rare cases). If you are interested in a month or two I have a paper coming out with is probably going to contain the most rigerous statistical analysis I have done to date. Basically one of the other authors has come up with a new model to use on the effects we study and I have to show that it works "better" than the models currently in use. I would actually appreciate your input on the statistical rigor of the analysis as it is nontrivial. The models are nonlinear and the parameters are highly correlated. If you are interested drop me an email with your contact information and I will discuss with the other authors about putting you in the loop. I can't guarantee that I can include you in an official capacity but I can obviouly show you the work.

Again, I agree--somewhat. There are many scientific claims that can be made and supported by the data without statistical testing.
Click to expand...

You said earlier that my reviews were not science because I can't do rigerous statistical analysis on them. I overlooked it then but you have repeated similar remarks in the above post. You are being condesending and even libelous with comments like these. There are people who have won nobel prizes and who have not done rigerous statistical analysis on the work. Read some journals outside your current field of research and you will find that this is not necessary to get data published - and yet it is still science. As a published scientific author you should very carefully consider when commenting on if someone is "scientific" or not. What you are doing is akin to one knife maker calling another incompetent.

However I agree completely with you that rigerous statistical analysis is always a good thing and that it directly makes better science. The key word being "better" . It is why I always seek out people at conferences to get their perspective and examine their methods in an effort to try and make my work more robust and why I would appreciate you looking at the paper I referenced in the above. The key word being "more" . Statistics does not make science. Look at many of the tests you are using and see when they were developed. Do you really think it was not possible to do scientific work before them?


Now all makers are branded as incompetent if one of their knives fail your test?
Click to expand...

As I have stated a half a dozen times before, I will allow one replacement before reaching a conclusion. I will give a specific example : my RCM rippled on hardwoods, it is possible this was a flawed blade. I have returned it to Newt for examination. If he concludes that it was a flaw and gives me a replacement I will repeat the work that rippled the first blade. If this one again ripples I will conclude that RCM's in general will ripple if the same thing is done to them and I will ask for a refund not another blade. I will also agree that it is quite possible that I could get by chance two bad blade in a row - however this is only probable if the makers QC is extremely low and I have no desire to support someone like that. Either way the same conclusion is reached so it doesn't even matter what his QC% is, the second alternative is to me an even much worse problem than the first.


-Cliff


[This message has been edited by Cliff Stamp (edited 05-01-2000).]
 
There's nothing I'd love better, Cliff, than to collaborate with someone on a project who thinks they can intimidate we with the threat of a lawsuit.
rolleyes.gif
Good luck on that one. If you email me, I will be glad to give you my opinion on the p value for your chances of success.

Instead, if I were you, I might be looking over my shoulder at some of those knifemakers out there that you are calling incompetent. They might have litigious ideas of their own.

------------------
Hoodoo

The low, hoarse purr of the whirling stone—the light-press’d blade,
Diffusing, dropping, sideways-darting, in tiny showers of gold,
Sparkles from the wheel.

Walt Whitman

[This message has been edited by Hoodoo (edited 05-01-2000).]
 
Hoodoo :

who thinks they can intimidate we with the threat of a lawsuit.
Click to expand...

I should have been more clear, you are not going to recieve any such action from me I was simply stating a fact - you are walking on dangerous ground and it really surprised me given your background. Assuming you are a published scientific author, going on a public board and making wide generalizations about people not doing scientific work is libelous as the many references I have given above show that what you are describing is simply false. It is not in any way necessary for work to be done as you have described to be called scientific as there is such work published in scientific journals. You are also not even using a standard defination of statistical by any means that I have seen including the three text references I gave.

I might be looking over my shoulder at some of those knifemakers out there that you are calling incompetent.
Click to expand...

That is an important consideration and why I have avoided making such conclusions even though people keep asking for them. Generally I just state what happened and let people make up their own mind as to what it means about the product. The only conclusion I have stated here is that a maker would be incompetent if his blade failure rate was about 1 in 5. I will stand by that and if someone wants to take me to court and claim that they are competent and that 1 in 5 is an acceptable fault rate for a custom maker they can.

-Cliff


[This message has been edited by Cliff Stamp (edited 05-01-2000).]
 
I should have been more clear, you are not going to recieve any such action from me I was simply stating a fact - you are walking on dangerous ground and it really surprised me given your background. Assuming you are a published scientific author, going on a public board and making wide generalizations about people not doing scientific work is libelous as the many references I have given above show that what you are describing is simply false. It is not in any way necessary for work to be done as you have described to be called scientific as there is such work published in scientific journals. You are also not even using a standard defination of statistical by any means that I have seen including the three text references I gave.
Click to expand...

Thanks for the legal advice Cliff. I didn't realize your expertise extended into the letter of the law too. Usually when I want legal advice I go to a lawyer but now I see I have another source. And it's free too!

As for what constitutes science, the usual standard is that you publish the results of your work in a peer reviewed scientific journal (this is not the same as putting them on a web page
smile.gif
). Is there a mad dog knife paper on the horizon? Enquiring mind wants to know.

And BTW, as to what constitutes scientific criticism. you might be interested in reading David Hull, "Science as a Process" (note that's a real reference, not just some vague reference to unnamed texts]. Our little fracas here is nothing compared to the battles in science that he documents. Funny, I don't recall that he mentioned anything about lawsuits. I didn't realize that's how scientists advance their research. I'm learning new things here.

And as for statistics, I will have to bow to your omniscience. Your the man Cliff. Anyone who can "know" the variance of a population like you with unwavering certainty, clearly is out of my league. I guess duffers like me will have to plod along and do it the old fashioned way, i.e., calculate it. But I'm starting to see what you are talking about. Just today I opened a jar that said peanut butter on it but I wasn't convinced so I tasted it. It only took me one sample to ascertain that it was indeed peanut butter. So my hat's off to you Cliff. Your methods really do have application.

I think this will be my last post on this topic. I don't see any need to continue. So Cliff, you can have the last shot. Maybe you can give me a textbook definition of what a statistical test of significance is. I seem to have gotten it all wrong.

Oh yeah, if I seem condescending Cliff, maybe that's in response to what I perceive as your arrogance. But that's just my opinion.


------------------
Hoodoo

The low, hoarse purr of the whirling stone—the light-press’d blade,
Diffusing, dropping, sideways-darting, in tiny showers of gold,
Sparkles from the wheel.

Walt Whitman
 
Well, what would a Mad Dog thread be without a good pi$$ing match, anyway? Of course, with the size of my prostate, I now refrain from entering any such contests.
wink.gif


Wasn't it Mark Twain (Samuel Clements) who said, 'Lies, damn lies, and statistics!?'

I just got back from a week long visit to the Dog Clan. All is well there, MD is working like, well, a dog, and the blades are getting made. Some weird pricing, however. Did you know that there is a $5 difference in price between the small Lab Rat and an ATAK? I told them to raise the price of the ATAK! (running, ducking, heading for cover)

I remain, (formerly whacko) Walt
 
Originally posted by Walt Welch:

Wasn't it Mark Twain (Samuel Clements) who said, 'Lies, damn lies, and statistics!?'

I've seen it attributed to Benjamin Disraeli, but I don't have a Bartlet's Guide(Quote book?) at my current disposal. Anybody know who the actual author is?

--jay
Click to expand...
 
Edge, so what was your question? I kind of forgot in all the confusion.
 
Here it is, Cobalt.
smile.gif


Originally posted by Edge:
The question says it all, what do you think of Mad Dog knives, are they all hype and can you get a more superior knife for the buck. I have an Attak on order and was wondering what to expect.
Click to expand...

BTTT...

(Back To The Topic)
smile.gif


Ryan



------------------
For the wages of sin is death; but the gift of God is eternal life through Jesus Christ our Lord.

Romans 6:23


 
This topic is 9 pages long!!

ok..

ok..

ok..
.
.
.
wait..

wait..

9 pages long!!!

wait..

wait..

wait..

OK.......

OK..........
NINE PAGES LONG!!!
.
.
.
.
ok.

 
Surly; my question was rather rhetorical. The quote in its' entirety goes, 'There are three kinds of lies: lies, damn lies, and statistics.' It is from S. Clemens' autobiography.

Walt
 
Hoodoo :

As for what constitutes science, the usual standard is that you publish the results of your work in a peer reviewed scientific journal
Click to expand...

Do you consider published reviews (in Blade and such) scientific articles? Why not - let me guess - because they are not scientific journals. How are you making that decision - what defines a scientific journal? Let me guess, it is one that contains scientific articles. How to you determine if something is a scientific article, well it is in a scientific journal. That is a nice circular and therefore meaningless defination.


Or here is another specific example - what about historians? They do a lot of statistics (some of the hardest because of the difficulty of clear data acquisition ). They would call what they are doing scientific research (so would I) however I have seen lots of published papers with no mention of p values at all. In fact I have seen that in journals including those from physics and chemistry as well as engineering. No confidence intervals at all listed with the conclusions, no specific statistic quoted.

In general the accepted method of proving a conclusion, even a very simple one, like for example does a linear correlation exist between two variables, is not constant across fields nor even in different journals in the same field. It is quite possible that an article rejected by one publication could be accepted by another. Check the journals I cited above and compare the methods used to what you would likely do. Here is another simple question, how do you know if a variable determined by your model is significant? Again check the journals and see how different people answer this question. Many different methods, some robust and some not. Statistics does not make science - it is simply a tool used by scientists. The degree to which is it used varies extremely.

Is there a mad dog knife paper on the horizon?
Click to expand...

There is currently a group of people that are working to try an develop a comprehensive series of tests for knife evaluation. These will include materials properties as well as field work and will correlate the two. Assuming this comes to pass will such work on a regular basis be published ? In a knife magazine - there are obvious $$ reasons why this is not overly likely. What about outside? Possible, it is an interesting idea. I never put much thought into primarily because I like the advantages of it being on the web. Mainly because it allows me to update them as my scope of work widens. However the idea of opening up the scope of viewers and presenting knife work in a non-rambo type roll is interesting. It is usually the approach I take when discussing the issue with non-knife people and it often works well.

some vague reference to unnamed texts
Click to expand...

I didn't list the specifics as the defination is common to every references I have ever read. If you want one -> "Statistics for Business and Economics" McClave and Benson. It is a textbook used in Business and I think Nursing. It is a very nice reference for method but there is little or no derivation just the results, as the math level for that is simply not there in those students. It as well does not deal with many complicated issues that often occur with real data - but it is a solid introduction.

As me me being all knowing - since you have once again misrepresented what I have said I will repeat it. I don't claim to know what the makers fault % is, I don't need to. There are only two possibilities. Either it is low enough so that if I got two faults it would be representative of the blades ablities (at your choice of a p value easily) or it is that high that his QC is so bad I would want to stay away from him anyway. Therefore the conclusion is the same regardless of the makers's QC. Now fact is if a maker actually wanted me to have a go at one more blade after two faults, I might even do that and I would probably keep doing it as long as they kept supplying blades thus the p value is as infinitely close to zero as the maker wants it to be. However I don't think that would ever happen.

-Cliff


[This message has been edited by Cliff Stamp (edited 05-02-2000).]
 
A while ago I said that this thread was beginning to remind me a lot of grad school. I take it back. You guys are a lot nicer. If you think this thread is hot you gotta see two academics go at it.
eek.gif


A lot happier since I dropped out of that PhD Statistics program, it's me ...
David
 
You must log in or register to reply here.
Back
Top