Tip penetration : amusing, actually quantitative as well

[Cliff Stamp]

I was fooling with a few knives the weekend, looking at a way to get tip pentration aside from the phonebook and wood work. If you take a pop bottle you can measure tip penetration effectiveness by poking a hole in the bottle from the side without knocking the bottle over. of course some knives can't do it so you have to add water, the more water the lower the score. You also have to come in straight on, its easier to do downwards. A really slim hunter gets a massive score compared to a heavy tipped tactical. Kind of amusing, you can put it on a stand and see if you can stab the knife into the bottle without knocking it off. It is not as easy as it might sound.

-Cliff

 

[Ritt]

Are you using plastic bottles, or do you have some seriously sharp knives and mad skills?

 

[Cliff Stamp]

Plastic, if you can do it with glass bottles, then I think you should give old Jack Burton a call.

-Cliff

 

[neosporin]

A "heavy tipped" tactical is meant to have alot of strength for prying power, plus to leave a big hole in whatever you're stabbing... doesn't surprise me at all.

 

[Chiro75]

How do you insure that each test is performed the same (i.e. vector of penetration is equal, angle of penetration is equal in all planesl, speed of knife is the same, point of entry on the bottle is the same, etc)? If scores are being given to knives in this test, then it stands to reason that all variables are being controlled in some way. Also, do you correct for various blade shapes? I.e. the Dodo would probably fail miserably at this test, despite a fine point and a thin blade, unless you account for the method in which this blade would be properly used.

On a round surface such as a soda bottle, lining up the vector by hand for each test would be difficult at best, without inducing a large amount of variability in the testing (X and Y axis variables, Z-axis rotation of the knife). Also, the human body moves in coupled motions, rarely in one plane at one time, so limiting all of these factors to perform a scientific study would be impossible unless an apparatus that stabilizes and moves the knives under constant variables is used.

The illustrations below attempt to explain this simply. In the first picture, an unusual blade shape well known for it's ability to penetrate is poked into the soda bottle. Because the shape of the blade is not taken into consideration, though, it would score poorly because the blunt aspect of the spine will contact the plastic of the bottle, rather than the point. If the vector of penetration is corrected to take into account the shape of the blade, the same knife that "performed poorly" will now perform in the top of the test blades. The second illustration shows how a slight variation in just one plane of the vector of motion used to penetrate the bottle could change the score of that knife in such a test.

So, I would conclude that this test is okay for some fun, backyard comparisons, particularly if one were to do 50-100 stabs with each blade and average the results. But, as a head-to-head, realistic comparison of knife performance, I would conclude that this test is meaningless unless an appartus that eliminates human error be used. Otherwise, there is a high degree of potential that "good performance" and "poor performance" has little to do with the knife, but is rather the fault/result of the operator and his inability to control variables. This could potentially imply to a reader who fails to read the method and instead focuses on the results that one knife is better than another in a task without taking into regard the high probability of error from the experimenter.

 

[Steelhed]

I kind of came up with something similar myself, only instead of pop bottles I use beer cans. One word of caution, though, drink the beer first. Stabs into full cans of beer really make a mess and waste good beer too. Of course repeated tests can lead to other safety concerns.

 

[Cliff Stamp]

How do you insure that each test is performed the same (i.e. vector of penetration is equal, angle of penetration is equal in all planesl, speed of knife is the same, point of entry on the bottle is the same, etc)?

All such human random variations are of no significant consequence. Most people can initiate a horizontal or vertical stab with a high degree of consistency and thus even a small sample mean gives highly convergent behavior, this is why the Sharpmaker works so well for example and can micro-bevel rapidly in a few passes. This is one of the fundamantal principles of sampling statistics.

So for example, when I first did it all I did was smack the bottle around, so the results here would have been fairly wild. However after a few minutes to work out the method I could do it readily with the finer pointed knives. When I failed to do it with some of the heavier tipped ones I would go back to the ones I succeeded with to insure I still had the method down.

The actual real problem with any work like this which relies on your ability isn't the random deviations but systematic ones. This is a real problem for long term comparisons when you don't use benchmarks. If I do wood chopping now with a blade I get much better edge retention and durability than I did 10 years ago because of the massive amount of such work I have done in comparing blades in that time.

This is why you should always use references blades in any such work and periodically recalibrate them to make sure that your method hasn't evolved or degraded. If it does you have to rescale or repeat your work to take this into account.

The other factor can be systematic differences because of your physical ability being off in that day, recovering from a workout, sick, etc., this is why for the comparative work I do which is numerical they are always averages over repeated work done on different days, usually at least four, through the space of a few weeks.

...it stands to reason that all variables are being controlled in some way.

No, just that their effects are bounded. You can't always eliminate or control random varibles to such an extent they can be ignored, you try to reduce them of course when possible, as the Sharpmaker does for example by chosing an angle which is very easy to judge by eye.

Also, do you correct for various blade shapes?

Any work done with a knife is influenced by blade shapes, some do some things very well and not others. The Dodo doesn't carve thick wood as well as the Military, but has a nice hooking action. As noted in the Dodo review, it can't stab a phonebook in the normal way so gets pretty much a zero in that regard, however a cut into a phonebook with the point leading scores very high compared to other blades like the K2.

So when you have a blade of drastically different geometry what you would do is investigate its "normal" behavior, not how it could be limiting in that aspect and then attempt to show how its optimal use could be illustrated. This may lead to running the work in a few different ways to get a set of benchmarks instead of just one.

A "heavy tipped" tactical is meant to have alot of strength for prying power, plus to leave a big hole in whatever you're stabbing... doesn't surprise me at all.

I would hope not. Compare the WB to the Temperance and it is obvious that the WB is far behind in terms of penetrating power. However compare the Temperance to the Manix and it isn't so easy to note which one is in the lead and by how much. Work like the above isn't done simply to rank wildly different knives but to look at knives where the performance is similar and thus evaluate geometry considerations such as more taper but wider blade and so on. It is of course one aspect of performance, the counterpart would be tip strength.

-Cliff

 

[Danbo]

Plastic, if you can do it with glass bottles, then I think you should give old Jack Burton a call.

-Cliff

Hey, that's actually quite funny and witty. I love that movie!

 

[donutsrule]

Plastic, if you can do it with glass bottles, then I think you should give old Jack Burton a call.

-Cliff

One of my favorite movies, ever.

 

[Chiro75]

All such human random variations are of no significant consequence.

I disagree. If you miss the bottle and dink it off the side of the blade, or hit it at a 50° angle instead of 90°, I fathom a guess that it will skew the results.

Most people can initiate a horizontal or vertical stab with a high degree of consistency

I disagree with this statement. Perhaps people can initiate a movement with some degree of consistency, but can they carry that motion out over and over again with consistency? And there's more to this type of test than that movement. How feasible is it that you'll repeat the same motion consistently over and over and over again when you stab a bottle, walk to pick it up, place it again, walk back to where you think you were standing before, and repeat? That's a lot of variables. If I'd designed an experiment with that many variables when I was a chem student in quantitative analysis I would've earned a big fat F! Also, if this motion was so consistent and repeatable by humans, there would be no need to train in things like martial arts. Pulling the trigger on a gun is an even simpler motion requiring fewer joints and muscles, and more simple neurology than stabbing forward, and yet that is something that must be practiced ad nauseum to develop proficiency. My knowledge of biomechanics says that the stabbing motion isn't as simple as you say, and there are more variables (distance, velocity, moevement in the X, Y and Z axes, foot position, hand/eye coordination, consistent bottle positioning, etc).

If I do wood chopping now with a blade I get much better edge retention and durability than I did 10 years ago because of the massive amount of such work I have done in comparing blades in that time.

Exactly my point. These tests rely on your skill level, which means they are not "quantitative" tests as the post suggests. In another 10 years with more wood chopping training, the tests will still be different. Same for stabbing a plastic bottle. This is why a quantitative test needs to have as many variables controlled as possible.

You can't always eliminate or control random varibles to such an extent they can be ignored, you try to reduce them of course when possible,

Exactly my point. You cannot claim that a test that controls none of the variables, then, as a quantitative comparison with subsequent ranking of the results. I'm not saying this test is worthless, just that it can't be claimed to be quantitative or anything above subjective when it is done using the method you've described.

If I was having one of my knives investigated in such a way and judged against others in a claim that the tests are "quantitative" with a conclusion pitting it against other products, I would like to see a lot more control over the variables. Otherwise, the bias of the observer comes into great play, not to mention the variables I've mentioned that are not being controlled in any other way than "I think I'm standing about in the right place and using about the same motion..."

A true experiment comparing various knives will eliminate as many variables as possible and it will use a repeatble method with actual measures of the performance of each knife.

 

[Ritt]

I've got to agree with Chiro75 on this one. It may be an interesting test of skill but it's hardly a repeatable, quantitative test of penetration.

 

[Cliff Stamp]

I disagree. If you miss the bottle and dink it off the side of the blade, or hit it at a 50° angle instead of 90°, I fathom a guess that it will skew the results.

You should not do it when you have been drinking. Most people could conclude that if you miss the bottle you could ignore that trial, and if you go from 90 on one stab to 50 on another I suggest that you not play with sharp instruments in general, and make sure your forks have corks on the end of them in case you accidently poke your eye out during supper.

How feasible is it that you'll repeat the same motion consistently over and over and over again when you stab a bottle, walk to pick it up, place it again, walk back to where you think you were standing before, and repeat?

I check these things all the time as noted for example in the above on different days. The mean results are quite consistent, to the precision noted. I have large sample trials of such things as wood chopping for example which go to normal quickly even in such heavily influenced comparisons, it is the wonder of the normal distribution.

Of course if you really can't do this consistently it isn't quantitative for you, that doesn't mean that everyone has such wild behavior. It is something that of course the user would check themselves the same as any other work done to judge knives. As noted if you actually try to go in on a 90 and hit a 50 instead then no, it isn't for you.

If I'd designed an experiment with that many variables when I was a chem student in quantitative analysis I would've earned a big fat F!

In chemistry high precision is usually very important, one of my friends for example worked for years preparing standards which needed four decimal places in the concentration. Of course when comparing knives you don't need to work to this standard and thus you have to adjust methods accordingly. If I am measing flour for bread I don't use the same balance I would use to measure a solute for a standard solution.

If you are looking for performance differences in knives at a minimum then about 10% is sensible, trying to say performance differences exist at below this level really doesn't make much sense because the user will never see them. Now go back and look at how chemistry lab methods will be effected if all you had to do was get to within 10%, titrations could probably be done with beakers and graduated cylinders.

You can't look at method independently of the conclusions. You must first start off and ask what do I want to measure and now what tolerance do I want. Ok, now look at the random deviations and noting the square root N behavior of such influences, see what steps have to be taken to get the desired results. Your arguement is shown to be trivially false because if people were that uncoordinated they for example the Sharpmaker would never function.

Also, if this motion was so consistent and repeatable by humans, there would be no need to train in things like martial arts.

Again the goal is different, here you want extreme precision as well as automatic focus and pattern involvement. Watch old tapes of Mike Tyson boxing for example, he would routinely go into long chains of punches and body movement which you knew were almost instinct. You are not looking at the average effect here, you can't ask your opponent to stand still while you fire off a half a dozen punches because the average effect will be decent, if you get really sloppy once and drop your guard you probably won't get a second punch.

These tests rely on your skill level, which means they are not "quantitative" tests as the post suggests. In another 10 years with more wood chopping training, the tests will still be different.

Yes, this is why I noted you use a standard and recalibrate. You have to do this by the way with gauges and such in the lab, this doesn't mean they are not quantitative. Lots of things can cause calibrations to be off besides human interaction so this always has to be done periodically, even if you are just using machines.

Quantitative by the way just means being able to assign a number, it doesn't imply a particular precision. You can have quantitative measurements which just record order, this means you know the answer is closer to 10 than 100. papers are published with such precision.

-Cliff

 

[Gossman Knives]

I definately have to agree about NOT doing this while drinking alcohol. Because in fact you may be bending over picking up your wee wee instead of the bottle.
Scott

 

[R.W.Clark]

You guys should know not to bother pointing out the high level of nonsense in what Cliff does. After all he is the end all be all master of the knife universe (just ask him). He might actually be able to answer in under 6000 words if he tries really hard.

Hey Cliff, get a gold membership, start posting photos and videos as proof for what you claim or STFU. Still waiting on that video you claimed you were going to send me over a year ago. Fact is, there is no video. Never was.

Time to put up or shut up Cliff. Provide real proof (in action photos of YOU ACTUALLY DOING THE TESTS or video mpegs of YOU ACTUALLY DOING THE TESTS) or shut up and go away.

 

[Chiro75]

You should not do it when you have been drinking. Most people could conclude that if you miss the bottle you could ignore that trial, and if you go from 90 on one stab to 50 on another I suggest that you not play with sharp instruments in general, and make sure your forks have corks on the end of them in case you accidently poke your eye out during supper.

Who said anything about drinking???? I didn't say miss the bottle. I said if your vector is different from one test to another, the results can be different.

I check these things all the time as noted for example in the above on different days.

I don't think that matters. My contention is that biomechanically your testing will be a bit different each sample. Different enough that no conclusions can be drawn. It'd be like me concluding stuff in my chem lab by eyeballing what I wanted to put in my test tube rather than weighing things.

Of course if you really can't do this consistently it isn't quantitative for you, that doesn't mean that everyone has such wild behavior.

I'm not talking about wild behavior. I'm talking about kinesthetic sense of the body, biomechanics of human movement and the neurology of movement and muscle activation. I'm telling you that you can't repeat movements like that over and over in a way that you can draw conclusions in a test that has the method you described. It's not arguable. I've trained chiropractic students to adjust patients. It takes constant work, and the difference of a few degrees and a few pounds of pressure makes or breaks the procedure. The movements involved in what you're talking about are much grosser, involve more variables, more joints, more neurology... In short, it isn't repeatable enough to claim "quantitative" results.

Of course when comparing knives you don't need to work to this standard

I agree, but then the test should not be declared "quantitative" because it can't be repeated from examiner to examiner. Also, when you are ranking knives based on performance (people ostensibly purchase the knife of best performance in most cases) I think it is very important to have the variables controlled properly, particularly when a knifemaker's liveliehood or the reputation of the a company's product is involved.

Your arguement is shown to be trivially false because if people were that uncoordinated they for example the Sharpmaker would never function.

The sharpmaker is a different animal. You can brace yourself, limit joint movement, get into a fixed position, and the rods are flat and at a fixed angle. I have seen people who can't swing their arms normally when they walk because their neurological function is all jacked up, and yet they wander through life "perfectly fine." This guy can grab a knife and do all sorts of tests, but are they accurate? It's arguable. If you set up motion sensors on all of your moving parts when you did this experiment and watched it later you'd be surprised.

Yes, this is why I noted you use a standard and recalibrate.

How do you calibrate yourself? Based on my clinical experience with relatively normal people, my training in human biomechanics, my knowledge of scientific method in general and my common sense, I still say this is not a measure that can conclude anything. Is it testing how good a skill someone has at penetrating a plastic bottle with a knife, or is is testing the knife? Unless you take the human skill out of the test, you can't state which.

Actually, you did state that it is testing the human skill because you gave the example of your wood chopping tests and how they have changed as your skill has changed. Same way someone who practices rope cutting in a competition setting will be a better rope cutter.

 

[Murky Depth]

You guys should know not to bother pointing out the high level of nonsense in what Cliff does. After all he is the end all be all master of the knife universe (just ask him). He might actually be able to answer in under 6000 words if he tries really hard.

Hey Cliff, get a gold membership, start posting photos and videos as proof for what you claim or STFU. Still waiting on that video you claimed you were going to send me over a year ago. Fact is, there is no video. Never was.

Time to put up or shut up Cliff. Provide real proof (in action photos of YOU ACTUALLY DOING THE TESTS or video mpegs of YOU ACTUALLY DOING THE TESTS) or shut up and go away.

Is this a serious post? Sometimes its hard to tell whether members who know eachother are being sarcastic/ironic or not.

I for one have read & seen pics of cliff testing a CRK Green Beret, & have found the review to be very useful & objective - along with many many of his other posts (on not just testing, but general knife info).

So, on the contrary, I think this place needs more folks like Cliff around.

 

[howiesatwork]

I was fooling with a few knives the weekend, looking at a way to get tip pentration aside from the phonebook and wood work. If you take a pop bottle you can measure tip penetration effectiveness by poking a hole in the bottle from the side without knocking the bottle over. of course some knives can't do it so you have to add water, the more water the lower the score. You also have to come in straight on, its easier to do downwards. A really slim hunter gets a massive score compared to a heavy tipped tactical. Kind of amusing, you can put it on a stand and see if you can stab the knife into the bottle without knocking it off. It is not as easy as it might sound.

-Cliff

A lot of uncertainties there.
The method is flawed, as has been stated by others here.
Angle of attack: The point 90 deg side to side, the penetration angle is measured through the blade how?
A stab made by hand horizontally is not a repeatabe action. A slash is more repeatable in this orientation, but a stab is more natural in an almost vertcal plane.
The blade approach is always in an arc, so handle/blade length have an effect on contacting speed if everything else is constant.
The bottle: The curved surface must be met at exactly the right point to make the penetration. Now, if the point isn't symmetrical, the point will skew the cut upon penetration, so an offset would have to be made, which would have to be predetermined to minimize the surface jumping to the side.
The curved surface is just too much uncertain, compared to a flat surface penetration.
Where are the pics?
Amusing? yes.
Quantitative? Not by a long stretch of the variables, followed by selective collection of "data".
Pseudorandom nonsense would be a good description...
Howard
Sr. Metrology Engr.

 

[thombrogan]

You guys should know not to bother pointing out the high level of nonsense in what Cliff does. After all he is the end all be all master of the knife universe (just ask him).

He didn't think he was until he accidentally huffed some beryllium while resharpening an LM1 Model 10.

 

[R.W.Clark]

This is a very serious post. Stick around longer and you will learn what the vast majority of long term members have learned, Cliff is a blow hard wind bag at best. In 99% of reviews all you will ever see are a few before photos and maybe an after photo showing some form of damage.

I have not personally seen any footage of him actually DOING a test. Anyone can slam a blade into the side of a steel bar and then type two pages about how it chipped out while chopping 2X4s. Or break a knife in a vise and say that it broke while carving a bowl. So until he starts providing real proof and not just windbag text, I am calling his reviews exactly as I see them and that is as a stinking pile of BS.

 

[R.W.Clark]

He didn't think he was until he accidentally huffed some beryllium while resharpening an LM1 Model 10.

Yah, that whole thing proves how little Cliff actually does now about physics when it relates to atomic bonding.

Watch out for that table salt guys! It contains sodium and clorine.