Knife Tests By The Experts?????????

[Cobalt]

If you want to perform testing under conditions that will be the same every time, then you cannot, test paper or wood products for the reasons you have mentioned above.

What you can do, however, is test plastic products. High density Poly Ethylene(HDPE or PE), Polystyrene(PS), polypropylene(PP), PET, PVC, etc. These materials all offer very consistent properties from one batch to the next and offer a range of hardness for impact as well as slicing. PVC, for example is very hard. Pet is not as hard but much harder(tougher) that the others. PE and PS are soft and can be used in slicing comparisons. PP is another inbetween depending on the form you get it.

The only problem here is cost. If you can overcome cost, plastics offer an incredible option for testing. Try chopping up a plastic drum. Not Easy.

 

[Joe Talmadge]

My own solution of being able to make sense of results may not be *perfectly* scientific, but it's accurate enough to be extremely meaningful.

Each knife I test, I test against several other knives, including one or two "benchmark knives" that are ALWAYS included in the tests. For example, every 4-inch bladed folder I test, also gets tested against an BM Ascent and Spyderco Endura. You can see the relative performance against the Ascent and Endura, and going back through my reviews, see how other knives did against those two also, and compare the results that way.

The advantage of this approach is that is accounts for an increase in skill. As I've done, say, the poly rope cutting test over and over, I've gotten better at 1. Putting on an edge that cuts well against poly rope, and 2. developing skill in using leverage to get the most from the cut. If I did not test against a benchmark knife, then old results would not be comparable against newer results, because the newer knives would all test relatively better. Testing against benchmark knives, my skill increases on all knives in the test, and the relative positions remain unchanged. So on my tests, don't look to see how many cuts it took the test knife to get through the rope -- rather, look to see how well the test knife cut versus the endura.

I'd also caution you guys about going overboard on accuracy. Being scientific also means understanding the right number of significant digits. I don't strive for any more accuracy in my tests, because I find that my bench tests are an extremely good predictor of real-life performance for me. More accurate tests would take up more of my time, but I can't see how they'd give me any more useful information. By comparing test knives against well-known benchmark knives in well-known test media, I believe I am achieving what I hope to achieve in testing.

And this view that *your* cardboard might be different than *my* cardboard -- again, I think you're going down the wrong road. I can't compare my results to yours anyway, because in, say, a rope-cutting test, I might be much better at rope cutting than you, so even if we had the exact same rope, we couldn't compare our results. But if we both test against the same benchmark knife, our results will be at least moderately comparable.

Joe

 

[Ben E Hana]

Joe that makes a lot of sense and I like Cobalts idea about the plastics too.I still think a Ft/lb number would be good in a destructive test though.But I guess too many numbers might be a drag.
troy

 

[Cougar Allen]

Plastics with the same name vary greatly in properties. For instance climbing rope is much harder, stiffer, and stretchier than boating rope yet they're both "nylon."

If we all used the same brand and model of something for testing, say Rubbermaid wastebaskets of the same size and color (different colors might not cut any different but they might) then we'd have a standard testing material -- until Rubbermaid changes their formula. They're likely to do that at any time and they won't tell us about it.... The formulas used in plastic consumer goods change frequently without notice -- so frequently that often several different ones are available at the same time.

I don't mean to reject the whole idea of using plastics for uniformity, but it's not going to be as easy as everybody testing on 1/8" thick polyethylene ... that standard would be considerably less uniform than white pine 2x4s.


------------------
-Cougar Allen :{)
--------------------------------------
This post is not merely the author's opinions; it is the trrrrrruth. This post is intended to cause dissension and unrest and upset people, and ultimately drive them mad. Please do not misinterpret my intentions in posting this.

 

[Cobalt]

Cougar, I disagree. The basic formula for most plastics made today is PE or PP with some regrind thrown in. regrind is usually PP or PE and it is small amounts only. It is easy enough to get consistent batches of any plastic you want and the properties will vary much less than wood or paper since they are not dependant on the elements of nature. Plastic sheeting can be purchased from any manufacturer quite easily and inexpensively and you are guaranteed the same batch every time.

Oh, by the way Rubbermaid doesn't change their batches as often as you would like people to believe. They may change color addtives but that's about it. Color addtives are a very small part of plastics products.

If you specify that you want sheets of HDPE or LDPE or PVC or PET they will be pretty standard throughout the industry.

 

[Joe Talmadge]

The only problem with standardizing on plastics is that the results will be especially interesting to people who cut a lot of plastic. But different blades have performance in different materials. By being single-minded on trying to have comparable tests, we've forgotten to make sure the tests are relevant to broad real-world usage.

The tests really need to give data that's applicable to real-world usage. To me, that means for a general knife, cutting a variety of different ways (shaving, push-cutting, zipper-cutting, slicing, slashing, chopping, piercing, all as applicable) in a variety of materials (hard wood, soft wood, hard poly rope, soft rope, food items, skin & meat, PVC, cardboard, plastic, etc., as applicable).

It's reasonable to try to standardize as convenient, but some materials will be hard to standardize. And as long as we test versus the same benchmark blades, results on their own will be meaningful, and in addition comparisons to other results should be meaningful, too. If I use ebony as my hardwood and you use kamagong, we're both using hardwoods, and with the same benchmark knife that should be good enough to compare, I think.

My view of the right first step is: figure out good benchmark knives for everyone to use in their tests. They should be inexpensive, ubiquitous, and perform well but not be too specialized. My faves: for 3" folders, test against a delica, for 4" folders, test against an endura. For 7" tactical-style fixed blades, kabar should be the benchmark. Etc.

Anyway, just my view. Trying to standardize on materials is a fine idea, but I think it will be difficult, and you'll still want to use benchmark knives to adjust for the cutter's skill and power/weight. And I'd warn that in focussing on materials that are easy to standardize on, you'll lose sight of other materials the real-world user might be interested in.

Joe

 

[Sergiusz Mitin]

Joe,
I like your idea to test each knife against certain well-known benchmark knife - it can give readers a lot of info.
But what to do with knives intended for basically different use? How we can to compare f. ex. piercing tip of SOG Vision with extremely sharp but relatively weak Endura tip? Flat ground Military always will outperform each hollow ground blade in deep slicking, bread f. ex.
Cutting abilities is not only thing user is interested for. Handling abilities are also very important, if not more important. How can we compare "no slip forward" AFCK handle with "no pull away" Military handle?
Is it possible at all to work out some standard test procedure which would be:

• wide enough to show all futures the reader could be interested for
• accurate enough to show these futures properly
• realistic enough to separate futures important for daily use from commercial gadgets
• reliable enough to compare only comparable things
• readable enough to don't bore readers deadly
  
• etc., etc., etc.

Seems Cobalt is right saying:
"I don't see a problem with people who may not be that knowledgeable about knives doing tests as long as they explain what they did and why."

 

[Cobalt]

Joe, you may be right that it is not a real world test. However, I feel that what is being down now is worthwhile testing and the plastic or Synthetic testing would only give more information into edge holding and impact resistance. Synthetics(ie Cloth) is another fairly consistent media. I don't think tests should replace what everybody does, just add to it.

But hey, it comes down to this. I don't care if your wife tested the knife, using it to cut meat and chop bone. I'M INTERESTED in those results. When you put together 20 people testing a particular knife, both novice and expert you will get better results than if only one expert tested that knife. IMHO

 

[Hopsing]

I'm new to the forum. A friend of mine told me about it. I really like what I see. I have owned knives for a very long time and enjoy using them hard.

Mr. Tex, you appear to be fairly new also, yet you want to judge who is an expert and who is not. I guess this means that I'm not an expert and that you would not value my opinion if I tested a knife that I purchased or owned myself.

So who in here do you consider an expert? I can see several just in this discussion alone; Mr. Cliff, Mr. Talmage and Mr. Cobalt. Mr. Cliff has a great test website and looks like he may be studying physics or something. Mr. Talmage appears to have tons of testing experience and Mr. Cobalt is a Materials Engineer (M.E.?) who obviously knows about testing materials and knows about plastics and other stuff.

Is a test that I do any less worthy than a test they may do simply because I'm not a physicist/ engineer or moderator? I don't think so. Yes my results may not be as good as theirs, but it's still valuable.

I have fallen in love with Stellite knives after seeing a friend of mine, who has owned one for many years. I can't afford one yet, but I'm saving my money. In the meantime, my ontario knives serve me well and I use them hard. My raider bowie is tough and I have not been able to break it.
So if you do not want me to tell people how my testing goes with my knives I'll keep it to myself.

 

[Cliff Stamp]

Cobalt :

If you want to perform testing under conditions that will be the same every time, then you cannot, test paper or wood products

The variance in the cutting stock can be reduced to an amount small enough to be ignored if you take the proper steps. For example, assuming you want to use cardboard as your cutting stock all you do is create a large bin and store all your cardboard there. Once you get a large amount of cardboard you can now draw sample lots from it for cutting tests. The variance from sample to sample will be much lower than the difference in the individual pieces in the bin. The larger your bin the more stable your samples will be. The bin can be refilled as you gather more cardboard. The only important part is to make sure the stock selection is random.

Joe :

Being scientific also means understanding the right number of significant digits.

Precisely, there are a lot of people who seem to think "science = high precision". All you need is precision greater than the magnitude of the effect you are trying to see with an upper limit of the maximum of the largest variance you are ignoring as going beyond that gives you numbers which have no meaning.

Sergiusz :

How we can to compare f. ex. piercing tip of SOG Vision with extremely sharp but relatively weak Endura tip?

You do some work showing the strength and relative weaknesses of each of the designs. Very basically, show that one will penetrate better, and also show what it gives up in order to do so.

Hopsing :

So if you do not want me to tell people how my testing goes with my knives I'll keep it to myself.

I would hope you do not.

-Cliff

 

[Rob Simonich]

Big Tex, you have started a great thread here and one that I think about a lot. Your original question of how do I know this person knows jack about knives is a tough one, especially in a forum like this with 5000+ members. Not everyone is known by everyone else. There is however a wealth of talent in here, and a lot of it in this thread.

Here is my take on testing and testers, subject to change without notice!

The experts I rely on are the end users of the knives I make, I pay special attention to any that have complaints, these guys and gals arent interested in charts and graphs but field performance. My customers range from housewives in the kitchen, (who incedently use knives more than most men day to day) to military personell, professional guides, survival instructors, professional chefs, ranchers etc. Charts and graphs are very usefull to some, but the bottom line is does it work, and does it work very well?

Last year I made a knife for a cattle rancher out of Talonite for castrating bull calves, a special knife with no point at all as there is a lot of work going on around a branding fire and a pointed knife will end up stuck in human flesh fairly regularly due to the the bull calves sometimes weighing 400 lbs deciding to rebel when their scrotum gets cut off to get to the oysters. The ranches down the Boulder Valley in Montana where this rancher is from do all their branding the same couple weeks. They all chip in and help each other and go from ranch to ranch and it is quite a party for a couple weeks there. (The locals call it the testical festival, as there is a big Rocky Mountain Oyster feed and party when its over) The Talonite castrating knife went through (the ranchers guess) 4000 to 5000 scrotums last spring and was still cutting when it was all over, without stoning or steeling the blade. The rancher who had me make him this knife didnt even know what the material was, but said it was the best castrating knife he had used ever, and he had been doing this for 40 years at the least. 40 years times say to be on the safe side 3000 scrotums a year amounts to 120,000 scrotums a year. That makes this guy an expert tester in my humble opinion. When I asked him if he needed me to make any more, he said no, this one was all he needed. This man incedently is named Jack, so he knows "Jack" about knives!

Dont get me wrong, I look at the charts and graphs, I dont understand a lot of them but I look for usefull information in them to help me make good decisions in knifemaking. They certainly have their place, and I respect the work that goes into putting all the information in order. Just my 2 cents worth in a great thread!

------------------
www.simonichknives.com

 

[Cobalt]

Cliff, I agree with the use of wood and paper products, but if you want to be exact to the decimal places, plastics and synthetics offer more consistency. However, I feel this slightly extra consistency is not worth the trouble of changing test media. It's also more unrealistic since you are not going to be chopping on plastic trees unless you are on a movie set, hehe.

Rob, man with a lot of sense and way too much time on his hands.

Hopsing, knives are not a science, there is no true expert in testing. Make your tests and post them, I will definitely read them.

 

[Cliff Stamp]

Cobalt :

plastics and synthetics offer more consistency

No arguement. I just don't feel that it is needed. For example lets assume that there is a 50% deviation from piece to piece of carboard cut in a 20 cm strip. This would mean that you made just one cut you could see a 50% difference between the performance of two blades just because of the difference in cardboard. However, this is not a cause to reject cardboard, while the 50% may seem daunting, it is quickly reduced if the proper method is used.

There seems to be a very large misunderstanding about how variances work and one of the worst explanations I have seen was by Bob Taylor in the following thread :

http://www.bladeforums.com/ubb/Forum1/HTML/001940-2.html

Where he comments (referring to my cardboard cutting comparisions) :

Your testing would be plus or minus 8000%.

Which basically means that I could take a knife and do say 1000 cuts and see a certain level of blunting but the next time need to take 81000 cuts to see the same effect. Needless to say this is not correct. It would make life really interesting if things did work this way however.

Anyway, if you take random 20 cm cuts into cardboard, say 50 in total, and them compare the two blades, the deviation in the average piece of cardboard that the blades saw will be reduced to 7 %. Further if you do multiple runs of the 50 cuts you can even make this smaller. For example do four runs and it will be reduced to under 4 percent. Of course you can make this as small as you want by doing more work and by doing smaller length cuts with a greater amount of strokes.

The same method by the way also eliminates error intruduced by variances in angle, force and speed of cut from stroke to stroke. It is also what allowed me to do the sheath throwing and draw a general conclusion about the relative durabilities without being concerned about the impact differences from hit to hit. I do this kind of analysis for all of the current reviews I do, I don't describe the method in detail as the results are what are of general importance. Of course if people want to discuss method in detail I would have no problems with it. However if people want to make blanket statements about the "unscientific" nature of the work that is a different matter. You are not going to see much in the way of a responce to that, unless it gets really extreme as for example Jeff Randalls comment awhile back that all reviews are biased

-Cliff


[This message has been edited by Cliff Stamp (edited 03-20-2000).]

 

[Cougar Allen]

The words "scientific" and its antonym "unscientific" as they are used by many on these forums ... it's a mystery to me what some of these people think a "scientific" experiment is. All I can figure is they think "scientific" means like the science they see on television ... with expensive-looking equipment ... everything is highly polished and the whole laboratory looks like it's just been cleaned ... with guys in white coats standing around with serious expressions on their faces ... with complicated glassware full of bubbling colored fluids ... with lightning bolts passing between electrodes ... a blend of PBS specials and Star Trek and late-night horror movies....

I keep seeing people start off a post with something like "I did some unscientific testing ..." but I seldom see any bad science in the post ... they used a control, they didn't throw out some of the data because of prejudice ... what was unscientific about it? They didn't use any expensive-looking equipment and they didn't wear a white coat while they did the test, I guess ... and they did it in a garage that could have used sweeping instead of in an immaculately clean laboratory like on tv.

-Cougar :{)
I wrote this post in an immaculately clean air-conditioned computer room and I'm wearing a white coat and have a serious expression on my face ... so you can believe this is a scientific post....

But all the posts I wrote yesterday were written in my living room with dusty old books piled everywhere and I was wearing ratty old jeans and a sweatshirt and I needed a shave and I kept giggling as I wrote, so all those posts are unscientific ... don't believe anything I posted yesterday.

 

[Cliff Stamp]

Cougar :

[reports]

I keep seeing people start off a post with something like "I did some unscientific testing"

I think a lot of people use this for the same reason people say IMHO, it is to prevent a backlash.

As you note there is a lot of very basic assumptions that are very wrong being tossed around. One of the most common is stated by Ben :

I rarely read a review where the number is concrete it is usually just a guesstimate.

There is no such thing as an "exact" measurement, all data is uncertain. For example if you use a protractor calibrated to a degree, at best you can estimate to a tenth of that, which takes a fair degree of skill.

It is no more scientific to say the angle was between 41.1 and 41.2 degrees than it is to say it was between 40 and 50 degrees, it is simply more precise. And in fact it is less scientific to simply say the angle was 41.2 degrees than to say it was between 40 and 50 as the first has no information about the uncertainty or error in the measurement.

As Joe noted, getting more numbers does not always mean better results and sometimes it can in fact be misleading as you can draw conclusions from differences which are not real.

-Cliff

 

[Ben E Hana]

I'm not concerned about characterizing a review as "scientific" or "unscientific" and I have been known to be wrong on basic assumptions before....

But- what is wrong with some more accuracy?If I told you I had a date for you and she was between 30 and 40 years old you might be inclined to ask for a more concrete number

I realize there is no exact measurement,but why not go as accurate as possible?The whole point is moot anyway because I had forgotten Mike Turbers tests.He had jigs and took a quite a few numbers and those tests still left people questioning.This whole thread has made me change the way I think about testing as a whole.I will be more inclined to read a review now and realize the time and thought put into it rather than focusing on things that might be wrong with it.Everyone (including me) with knives should be posting observations,reviews and tests of their knives I think that is the best way to up the quality of the entire reviewing process.
troy

 

[Cobalt]

Ben: "what is wrong with some more accuracy?If I told you I had a date for you and she was between 30 and 40 years old you might be inclined to ask for a more concrete number"

Ben if you had a date for me, all I would care is that she be 36+- 2, 24+-2 and 34+-2. Her age between 30 and 40 wouldn't bother me. I wonder if my wife is reading this? Honey, I was talking about you.

 

[Ben E Hana]

 

[Cliff Stamp]

Ben :

I'm not concerned about characterizing a review as "scientific" or "unscientific"

Unscientific is a very strong term, I wish people would have more care in using it. If you refer to someone in that field in such a manner you are calling them either ignorant or incompetent.

But- what is wrong with some more accuracy?

Nothing, but that was not being discussed. How many digits you record determines precision not accuracy. Anyway, if you take an excess of digits the mildest problem you can have is that you have wasted your time, the worst is that you can deduce properties than are not real (you are measuring beyond the uncertainty limit and calling it real data).

For example lets assume I do some prying and a blade snaps at 46 degrees, I repeat this with another and it snaps at 41. What can you conclude from this? Nothing. There is a large variance in the way I apply force to a blade (velocity and direction) and the way the blade sees it (flex in the stump). This means that if I took 100 blades say and snapped them they would not all break at the same angle but instead the break point would be smeared out based on the conditions of the stress.

Thus when I am doing prying tests the only *significant* difference I will assign is one that I can readily see by eye. If it is smaller than that I don't think that it is a meaningful seperation. Now you could argue that it would be trivial to greatly reduce the variance in the applied force, I could be using a vice and a bar - and you would be correct. This would allow me to get a much more constrained effect. However, I think that this is now starting to get too far removed from actual field stress and could lead to false or biased conclusions. This is related to the point Cougar brought up about the sheath dropping, if you looked at one particular impact vector you could get very misleading data.

Of course if the next time someone wants me to review a blade they are willing to send me 100 or so of them, I will gladly be more precise with the flex during the break work as I can readily get a variance estimate by breaking all of them.

I will be more inclined to read a review now and realize the time and thought put into it rather than focusing on things that might be wrong with it.

There is nothing wrong with dicussing the method but it makes a big difference how you approach it. As an example, awhile ago I think it was Jeff Clark who dropped me an email. He commented that since I do work regarding the toughness / strength and strength of knife edges how about doing some staple cutting. His reasoning was that it was something that people should be able to relate to quickly and of course do it themselves so it would be a good benchmark. I agreed and now it is part of what I do.

Everyone (including me) with knives should be posting observations,reviews and tests of their knives I think that is the best way to up the quality of the entire reviewing process.

I could not agree more with that.

-Cliff