Extreme Judgement : just some spec's and a little splitting

[Cliff Stamp]

If raw data (presented as befits a scientific report) is not sufficient to form a meaningful interpretation...

You didn't look at the raw data, you looked at the summary. The raw data was given in the above and compared the two knives. Essentially what happened was that I sharpened the Buck and compared the work, then noted the difference in angles was likely the critical difference not the actual properties of the steel so I repeated it with different angles to illustrate.

By the way, raw data isn't enough to lead to a conclusion, usually it takes more work to turn the raw data into actual useful numbers than it takes to collect the raw data in the first place. When I started collecting data I was using a 386 to run reduction programs. I could collect the raw data in about an hour, it used to take me a week of solid computer time to take all these numbers and run predictions from them, it isn't a trivial step.

Such a judgment is something that should be done when interpreting collected results not a measurement in itself.

Depends, lots of measurements are hugely uncertain, it doesn't make them biased, or useless, it just makes them uncertain. You can have error bounds which are massive. For example I recently had students calculate densities of fluids using four different methods.

The uncertainties ranged from ~5% to up to almost 90% due to limitations in measureing devices. The goal was to illustrate that science is independent of precision, you can work towards understanding of physical laws even with really crude measurements.

You can for example deduce values for the gravitational constant acceleration with a piece of string of unknown length and counting seconds in your head. I had the entire class do it and then averaged all their results showing them how all the random deviations smoothed out and they got extremely accurate and precise values, within 10% with *NO* measureing devices of any kind.

As for what I do in the reviews, usually I give bounds only on cutting and such so they are standard deviations of a number of trials. The only other times are where people asked me to estimate energy of impacts so I bounded that as well. Similar to poundages, angles and so on for flex tests.

And the fact is, that even the most precise error bounds always in the end lead to a conclusion like "It isn't likely that this correlation is random." That is the most concrete statement you can actually make, being specific about what you mean by likely which is usually <5% for most physical sciences or <1% for biology and medical research.

Which properties are you talking about?

It has an inherently lower impact toughness shows by its chipping out.

-Cliff

 

[OwcA]

You didn't look at the raw data, you looked at the summary. The raw data was given in the above and compared the two knives.

Silly me, than you for telling me what I did.
It's quite simple really. Methodology description implies two test runs at different geometry:

The edge was later modified after hard impacts and repeated tests against other blades

however I can not find two sets of data for Buck 119. And it gets even worse . In CRK Green Baret review you make a reference to two distinct Buck geometries:
1.

Buck 119 used as reference. Edges had a 20 degree microbevel, medium Spyderco Rods

2.

Compared to the resharpened Buck which was 0.025-0.028" thick and ground at 11.6 +/- 0.3 degrees per side.

neither of with corresponds to geometries mentioned in the Buck review:
1.

it was approximately ground at 12-13 degrees per side

2.

the new profile left the edge 0.019-0.022" thick and ground at 9.2 +/- 0.2 degrees per side.

---

By the way, raw data isn't enough to lead to a conclusion

It should be.
What you do with it is a completely different thing. For one, it is a part of your interpretation, someone else with different data mining/analysis methods may come to different conclusions based on the same data. This is quite normal in scientific circles.

Depends, lots of measurements are hugely uncertain, it doesn't make them biased, or useless, it just makes them uncertain. You can have error bounds which are massive. For example I recently had students calculate densities of fluids using four different methods.

What you are saying has nothing to do with my objections.
No matter what the error is, it should be a result of a repeatable experiment. If you experience large fluctuations during an experiment your methods are most likely flawed. Such errors can, if one strives for scientific methodology, only occur when comparing results to actual (accurate) values.
On the other hand nor I, nor anybody else has a faintest idea what do you consider "similar" and what "significant", therefore your experiments are unrepeatable and as questionable as such.

It has an inherently lower impact toughness shows by its chipping out.

I was hoping a for a bit more in-depth explanation. For example in terms of physical phenomena governing solid material deformations.

 

[Cliff Stamp]

It is hyperlinked. In the Buck 119 review there is a link to the Gb review :

http://www.physics.mun.ca/~sstamp/knives/green_beret.html#impacts

There are two sections here, one where the Buck 119 was much thinner and got more readily damaged :

"The Buck had a more acute edge angle, 24-26 degrees included, compared to 35-39 for the Green Beret. "

and another where both were ground at the same angle of around ten degrees per side.

These match the two angle references you listed in the Buck review, 12-13 per side is 24-26 included, and ~10 is 9.2 +/- 0.2.

If you experience large fluctuations during an experiment your methods are most likely flawed.

No, it shows limitations in method, this is not a flaw. Lots of experiments have huge errors mainly because you are often limited due to monetary or time constraints, or simpy due to statistic properties of data, such as for example huge peak widths reducing resolution, or tradeoffs such as signal intensity vs resolution, you often physically can not improve both at the same time.

I can give you published references for people in my work group who do simulations with extreme uncertainties and have to filter and box their data and it is still extremely noisy. High precison simply isn't possible in certain areas. Lots of published data is often just bounded to orders of magnitude and thus there are you are looking for is a power of ten agreement.

Do you actually do experimental research? If so in which field? I am extremely curious as to what manner of research has lead you to believe that a large uncertainty indicates a flaw in the experiment and that you have no experience with box bounding experiments, OoM calculations, and simply highly variable data in general.

For example in terms of physical phenomena governing solid material deformations.

It isn't deformation, it is cracking. The Basic's have a much higher flexibility and impact toughness, even Reeve would not argue that. Are you really trying to claim that a stock Project would fare better than a stock Basic in a head to head test of edge impact toughness? Go over on Reeve's forum and ask him right now if you took those knives and slammed them into something hard with the NIB edges which one would take less damage. if he actually guarantees that the Project will take less damage I will buy both, put both of them on Video smashing into something hard, and show that it is nonsense by a mile.

-Cliff

 

[OwcA]

and ~10 is 9.2 +/- 0.2.

Interesting. So error margins are indeed only trinkets.

No, it shows limitations in method, this is not a flaw.

Do not put words into my mouth. I did not say it is always a flaw, but more often than not it is (granted, it depends on the field of research, but determining fluid densely is not among them). Reaching a limitation of a method could be a result of a wrong selection which is an exterminator's error. In either case it severely limits the scope of experiment and should be emphasised. Such measurements also have considerable less weight, especially if they "prove" something controversial.
Of curse it could well be that no better way exists, that is why new methods are often a byproduct of research.
Do not confuse the need to filter data with poor precision as a result of insufficient resolutions/precisions. If you can meaningfully filter data, it still contains relevant information, while an end result (and I emphasise end result, I am fully aware that 10M data points taken with 20 % precision can quite comfortably produce a very accurate result) with a 90% error does not.

It isn't deformation, it is cracking.

I wasn't seeking your opinion but science behind your reasoning. I would have asked you for some empirical data to back up your statement, but as we have seen, it's not all that reliable, some might even venture as far as deeming it uncredible.

 

[Cliff Stamp]

Interesting. So error margins are indeed only trinkets.

When referencing data, you usually do not copy it in full. ~10 means the angle is approximately 10, this is what the ~ symbol means. If you want more information you look further as to the nature of the approximation and trace it back to the origional measurement to not the extent of approximation.

Generally I quote the error bounds once but in refering to it always round them, so when measuring an angle initially it will be like 15.5 +/- 0.5. but later it is is always ~15-16. Most people reading the reviews find the latter more meaningful as well.

Of course if you want more precise numbers just ask, I can give you the t-values and CI intervals if you need them. My primary field of research is numerical analysis of nonlinear dynamics, with a heavy emphasis on nonlinear numerical modeling.

Reaching a limitation of a method could be a result of a wrong selection which is an exterminator's error.

Again this is nonsense. All methods of experimentation have uncertainties, all you can do it try to reduce them when possible. Some times this reduction isn't overly possible, because of time and/or money. If it costs 100x as much to get 1/4 error reduction you have to ask yourself is it worth it when I could get more data on other systems. if you are working in a standards lab then it is, otherwise, probably not.

The other part which is literally insane is that I do this as a hobby and ti is completly financed out of my own pocket on my own time and I get no monetary compensation for it. Yet you critize methods for not achieving some arbitrary level of precision which you of course apply to no one else in the industry.

The only time that would be relevant is if you were in a research group where you were getting paid for research and not using the funding for proper equipment or slacking off and cutting methods to give you more free time. All of what I do is out of my own free time. Of course the results would be more accurate and precise if I bought three of each knive and duplicated each piece of work with each one. You actually think that is a reasonable critism?

Give me the knives, I don't mind doing the work. If you read other reviews I have done it on duplicate knives when given them to check the repeatability.

Do not confuse the need to filter data with poor precision as a result of insufficient resolutions/precisions. If you can meaningfully filter data, it still contains relevant information, while an end result (and I emphasise end result, I am fully aware that 10M data points taken with 20 % precision can quite comfortably produce a very accurate result) with a 90% error does not.

Again, the percentage error is irrelevant. Calculations and experimentations can be much futher off. My own lab group has done measurements where the results were just order bounded, this means the uncertainty is so large you can just tell the result is 10^n where n is an integer and nothing more than that so the error is *much* larger than 90%, it is like 900%. At times this is enough to generate enough data to allow for testing of a theory and we have several papers along that line.

data

I gave you the experiemental data, twice, and asked you frankly if you were actually making a contradictory assertion. With all of your comments you actually never stated if you disagred with what was noted but danced around it, as noted the maker certaintly will not, and if he has rapidly changed his opinion recently I will gladly demonstrate on video as to the nonsensical nature of said claims. IN fact you are the only person to ever debate that issue, essentially arguing that the thinner edge on a softer Project is more durable than the thicker edge on a harder and tougher Basic. Hell, Hossom disagrees with almost everything I say on principle and I am almost positive he would even disagree with you there. Though he would probably think it was biased of me to point it out.

-Cliff

 

[Jerry Hossom]

...Hell, Hossom disagrees with almost everything I say

Only when you're wrong Cliff, which of course is almost....

 

[jedi_pimp]

fffffffffffffffffffffffffffff

 

[Cliff Stamp]

[Ontario, SOG, Mad Dog, Chris Reeves]

Everytime anyone mentions these knives, you rag on them

Look at the number of posts I have made, now look at the number which deals with the above. Now look at the number of posts made about those guys and the number of times I have responded. In both cases the numbers are a trivial percent.

I'll bet they don'[t flex very much, couldn't hack through knots, etc.

Yeah, he broke one before he sent it to me so I could check the grain. He also doesn't promote them for that type of use, and is quite open about the fact that they would break very easily.

As for the parts that you suggest I remove, lots of people find them useful. Others don't understand them. If you don't or don't find them useful then just ignore them.

What would you do if a student handed in a lab report that used the "scientific" methods you use in your reviews?

That is easy, I would obviously publically berate them, start spreading rumors about them, call them biased, accuse them of being other people, that other people are paying them to do their work for them, say that they have an agenda, make up lies about what they have said, etc. . And then when all that is done hold myself up as an unbiased source of information and say that my actual intention was to help them write a better review.

-Cliff

 

[Steelhed]

Cliff - Are you going to write anything more about the Fehrman? You never answered my question about the handle gaskets, or whatever you call them. Did you perceive any benefit, any cushioning effect. Apparently one can take the slabs off and add weight to balance out the knife. I have not tried it though.

 

[Cliff Stamp]

Missed it, no I didn't notice any effect. The handle it too thin for me and thus it got uncomfortable quickly with heavy batoning. I tried one night just to see how bad it would get if I really pushed hard and my hand was severely reduced in function the next day. If I had to rely on it, I would wrap the handle to thicken it, then wrap it with grip tape.

It wasn't as bad as the Buck/Strider I tried yesterday though. The Solution. The first knot I hit was impossible to cut through. I gave it a few test knocks which made little progress, then slammed it decently hard and almost broke my hand. The grip is just way to thin, and is too may pressure points. Nice cutting profile though, not sure what to make of it as it though, it is like the older Buck Zipper with a "tactical" wrapping, but not really suited to heavy utility use.

As for the EJ, I plan to return it. The edge isn't useable with the ripple, further impacts would cause it to distort more and then tear. I would have done it sooner, just been really busy and have not been able to get to the post. The profile is really nice for a light brush knife. It was in fact what I settled on quite some time ago as about the thinnest I could go for chopping, I was thinking on 0.020", which may be fine for just chopping, no splitting.

For a heavier blade of this type I would want a thicker edge, it really surprised me with all the people using these that no one had a problem with heavy work with an edge this thin. It is not even like the wood I cut here is what you would call harder than average, most places in the US have woods which make what I cut soft by comparison and it isn't even that cold yet, it only got down to -20 C once.

Anyway, 0.035" should be fine, probably even 0.030" for most work, but I would bump up to 0.035" to handle high stress, and idiotic friends with too much ass and too little sense like my brother and his silverback friends. Plus I would also want a steel which dented not fractured on heavy impacts this is critical.

That actually kind of surprised me given the spec's on 3V, I was expected it to get smashed in. Plus the handle is too similar to the TOPS grip which lowers versatility and creates too much pressure in heavy impacts. That was kind of stupid of me, I should have asked was there anyone who was using it who didn't like the TOPS grip.

Anyway I play to drop Phil an email about the chipping and see how his steel responds, he usually leaves it at ~58 as he had blowouts when it was harder. He runs his edges on the large blades really thin like this, but they are not baton capable, cutting tools, not emergency/survivial type blades.

Personally for that type of knife I am not really that interested in wear resistance, if you can get it fine, but it is not like the edge wears away cutting wood, it breaks off. You can cut wood for a very long time with a tramontina bolo before the edge doesn't respond to steeling.

Stick to simpler steels with a very high impact resistance at a high hardness level, ~58 HRC. If you can get wear resistance, corrosion resistance, or impregnated GPS systems, cool, but only if you don't lose the critical components.

-Cliff

 

[Steelhed]

So far my experience with the Fehrmans I own differs only in that I use them differently. I have a Peacemaker, a Shadow Scout and a First Strike. Because of the thin edge I would not choose a larger Fehrman for general chopping, but for general camp use and blade lengths of 7 inches or less, I have been impressed with these knives. They come sharp, they stay sharp, and they resharpen fairly easily because of the thin edge. I have had to reprofile the edge on a couple of my smaller Busses because the edge is too thick for kitchen chores and other less punishing activities. Of course this means they lose some of their prybar and heavy work capabilities, but I have others I've left stock if I need a more robust knife. Of the three Fehrmans I own, I probably like the Shadow Scout best. The thinner stock, longer blade and handle geometry work well for a variety of camp chores. It's a keepr as far as I'm concerned. Thanks for the info.

 

[Jeff Clark]

Cliff, I think that it is time to give up trying to dialogue with some people. You need a standard expression equivalent to RTFM or RTFFAQ when questions get pointless. I didn't notice it, but you need an FAQ on your website to direct some of these questions to. It is also probably not worth your while to mention Jerry's name on the forum.

I appreciate your labors. I've spent enough time in physics labs and in engineering labs to appreciate the work and insight you put into your tests. I don't see anyone else doing a fraction of the work. Thanks.
Jeff C.

 

[kbog]

That is easy, I would obviously publically berate them, start spreading rumors about them, call them biased, accuse them of being other people, that other people are paying them to do their work for them, say that they have an agenda, make up lies about what they have said, etc. . And then when all that is done hold myself up as an unbiased source of information and say that my actual intention was to help them write a better review.
-Cliff

Cliff,

I do not think that you HAVE to compare a knife against another knife to get a meaningful review. I think a rating scale would work fine if it is supported by data. This may help alleviate the bias issue that seems to be a black mark.

You know, I'm not sure I agree with all you have to say about things which is fine, but the response above is priceless.

 

[Eric_425]

Cliff,

I do not think that you HAVE to compare a knife against another knife to get a meaningful review. I think a rating scale would work fine if it is supported by data. This may help alleviate the bias issue that seems to be a black mark.

You know, I'm not sure I agree with all you have to say about things which is fine, but the response above is priceless.

How? The rating scale, supported by data, is going to be data from testing other knives. Which is again, comparing one knife against others. All reviews are comparing one thing against another. Performance is relative. Not to mention the variations that occur from test to test, substrate to substrate.

If he just takes a knife, and it takes 15 chops to cut through a 2 x 4, what would that tell you? It's a good chopper? Give it a rating of 9/10, because it took so few chops to cut through?

A 12 year old could take the same knife, and it takes him a 100 chops to cut through. Did the same knife suddenly become a bad chopper?

Or if he finds an exceptionally tough piece of wood. The same knife might take 30 chops this time.

There's too many variables that would need to be taken into account. The only way to count those out would be to use other blades for direct comparison. A review without them would be almost completely worthless. Yeah, Cliff isn't completely objective. nobody is. But he's the most objective reviewer on this board by a landslide. Really. Nobody else comes within miles. More completely numerical and objective data, measureable and repeatable, come from his reviews than anybody else. Cliff should just rest easy knowing none of those criticizing have put out reviews as unbiased as his.

 

[Lukers]

Cliff
You mentioned that the EJ and TOPs handles are not ideal for heavy use. What sort of contour, thickness and geometry would you recommend(in your scope of usefulness) for a heavy use knife? How significant is the handle/blade angle? For example, the Busse E series is supposed to have a 15 degree cant. Did the cant on the EJ affect the chopping ability and handle comfort?
If you could guess, how would 3v hold up(impact resistance, resistance to chipping) if it had a geomtry similar to the BM-E?
Thanks again for another great review.
-Luke

 

[The General]

One thing rings clear on such threads, Busse knives are strong and well suited to those who perhaps lack proper knife using skills and want a knife that compensates for this.

I don't mean that to be a snipe or a nasty comment, but if you have a Busse you get that reasurance that almost nothing you do will land you with a broken knife.

Reasurance sells. Ask Toyota...

 

[Cliff Stamp]

I have had to reprofile the edge on a couple of my smaller Busses because the edge is too thick for kitchen chores ...

Yes, Busse in general doesn't make good kitchen knife replacements. Spyderco however makes really nice kitchen knives.

Cliff, I think that it is time to give up trying to dialogue with some people. You need a standard expression equivalent to RTFM or RTFFAQ when questions get pointless. I.

It is obvious in the above that it would not help because they don't even read the reviews the quote. Plus I don't mind doing it, and quite frankly it amuses me more so than anything else to have someone critize "scientific" method in a manner which is contradictory to the whole approach. It seems about once a year this thread happens, it is always the same arguement.

You are biased because you compare all the knives to Busse (as noted, a complete fabricated lie), or your reviews are not scientific enough, or your reviews are too scientific. The really great thing about the last two as you can always find people arguing for both of them at the exact same time which is just insanely funny. The greatest thing though is "well I would not have a problem with you except you project yourself as an expert"

No where in the reviews does it promote me as anything. In fact they don't even say who writes them. I don't promote them at all on the internet aside from the one link I make in here and the occasion reference on rec.knives. They are not even in my sig line. I also don't attack or degrade other people in an effort to promote myself. Unlike a lot of other people, some of which even have personal attacks on me and my "ignorance" in their sig lines.

The irony is so high it is of nuclear proportions.

One thing rings clear on such threads, Busse knives are strong and well suited to those who perhaps lack proper knife using skills and want a knife that compensates for this.

You were obviously never in a high stress situation, as in those cases your fine motor skills go to crap and your strength goes really high. Last year I was almost washed out to sea when I was hit by a rogue wave. During the impact as I was being hauled out with the wave I dug my hand in a crack and held not only my own bodyweight, which was considerable, but the force of the wave. In the impacts I took severe brusing to my elbows, hips and lost most of the skin from my finger tips and almost the nails. I felt NOTHING, and I did it (held on), easily.

How? The rating scale, supported by data, is going to be data from testing other knives.

Because if it wasn't a direct comparison you could pretend it didn't exist. It is is frankly stated in the review it is harder to ignore. As you noted, it is a pretty irrelevant performance test with no reference. I had my nephew do some chopping with a Murray Carter knife awhile back. Lovely knife, it didn't do too well with him however. This doesn't mean we should petition for Murray's Master status to be rejected.

You mentioned that the EJ and TOPs handles are not ideal for heavy use.

Not for me, Hood and company love them. I find them too thin and the index fingr cutout too restrictive and too high of a pressure point.

What sort of contour, thickness and geometry would you recommend(in your scope of usefulness) for a heavy use knife?

Murray Carter does a nice grip, I'll see if I can't send you some profile shots. Hossom gets a lot of praise for his grips as well and much of what I remember talking about him with it made a lot of sense. See if you can't handle his knives, the Ontario's should be out soon, or drop him an email at the least.

Did the cant on the EJ affect the chopping ability and handle comfort?

No, this is mainly a long term wrist strain issue.

If you could guess, how would 3v hold up(impact resistance, resistance to chipping) if it had a geomtry similar to the BM-E?

At 0.035 it should be near impossible to ripple on wood, the strength increases faster than linear. The chipping resistance would not be increased, if it chips at 15 it will chip at 25. It just takes more force, but the failure mode is the same.



-Cliff

 

[The General]

Yes Cliff, clear knife use situation there. I have been in many iffy situations where my tools did not fail me even though I was cold to the point where fire would stave off hypothermia and death. I survived, my tools held up, I am a very big guy and quite strong. I did not baby my tools, when your life is on the line you can't. Repeatedly my skill and tools held up, but then you don't know my skill level or what tools I use.

Hint, I burnt wood, I chopped wood. Concrete does not burn so well in our parts...

 

[thombrogan]

Jeff Clark,

There's always the standard list of ways to attack Cliff's work. There are more ways to disparage his reviews than you can flick a Sebenza.

 

[thombrogan]

Concrete does not burn so well in our parts

That's the problem! We only use flammable concrete on this continent.