What makes a steel 'aggressive'?

[Traditional.Sharpening]

What do you mean that a carbide rich steel 'cuts better'? This is rather vague and I don't know if you mean cuts longer, gets sharper, etc.

I think that almost all steels can become very sharp. A BESS tester would prove that, but what a BESS tester cannot do is tell you how many carbides are in the blade and how long it will cut.

Given available test results, especially Larrin's repeatable tests of total cards cut by a machine, I think you would agree higher carbide steels cut longer.

I do not agree as Larrin's model has not been proven to actually correlate to real world cutting. Larrin himself makes no such claims that his tests predict real world results. He simply puts forth the data from his tests, the problem is too many conclusions are being drawn from this data which has not been proven to correlate highly. I cannot understand personally why you'd go to all the trouble of modeling with these tests to not even bother to see if the model translates to reality. Cliff Stamp had directly asked Larrin about this on his forum and Larrin's response was that it was simply assumed that the CATRA correlates to reality. That is not science, that is nonsense to me.

We do not cut like machines and CATRA certainly DOES NOT put any lateral loads on the knives which is a common thing in real world cutting. I will agree that if you generally cut to extreme low levels of sharpness then you may find higher carbide steels preferable. I would generally sharpen much more frequently than your average person therefore I find little benefit to high carbide volumes. Actually I'd argue that makes them far less preferable as you lose too many other qualities by gaining the extra carbides. Why would I want a steel that makes it harder for me to resharpen? Fancy stones, more time, etc.

So what you really mean by 'cut better' means that they will cut longer. This isn't necessarily true either because knives do not just blunt by slow wear and there are other factors that influence blunting. Some steels in fact do behave better or worse as far as taking a high sharpness, any steel can get sharp but it depends on how you define 'sharp'. For many simply cutting printer paper equals sharpness, for me I would not really call that knife sharp in a real sense. It's approaching being truly sharp but if it's not popping hairs or cutting newsprint then it's a very low standard.

I don't think I ever said that CATRA tests modeled real world numbers, but they do. The results may be a multiplier, such as *1.5 to get the actual numbers, and they are repeatable, so they are very scientific.

You do not agree that Larrin's model has been proven to actually correlate to real world cutting, but there is a guy who has been doing real world cutting tests on sisal on YouTube for years and they match up with Larrin's CATRA results. I forget the guys name, but his account is "CEDRIC & ADA GEAR AND OUTDOORS" and his tests results are available for FREE.

How much better are new steels vs old ones
https://youtu.be/mo30nNQcHhc

https://youtu.be/mo30nNQcHhc

Full steel testing chart of Cedric & Ada's cut test videos:
https://docs.google.com/spreadsheets/d/ ... d=43566811

Yes, I've seen him and you referencing that testing may in face be proving my point that from what I've seen he mainly cuts clean rope. Has he ever cut dirty materials?

[Traditional.Sharpening]

I think that almost all steels can become very sharp. A BESS tester would prove that, but what a BESS tester cannot do is tell you how many carbides are in the blade and how long it will cut.

Given available test results, especially Larrin's repeatable tests of total cards cut by a machine, I think you would agree higher carbide steels cut longer.

I do not agree as Larrin's model has not been proven to actually correlate to real world cutting. Larrin himself makes no such claims that his tests predict real world results. He simply puts forth the data from his tests, the problem is too many conclusions are being drawn from this data which has not been proven to correlate highly. I cannot understand personally why you'd go to all the trouble of modeling with these tests to not even bother to see if the model translates to reality. Cliff Stamp had directly asked Larrin about this on his forum and Larrin's response was that it was simply assumed that the CATRA correlates to reality. That is not science, that is nonsense to me.

We do not cut like machines and CATRA certainly DOES NOT put any lateral loads on the knives which is a common thing in real world cutting. I will agree that if you generally cut to extreme low levels of sharpness then you may find higher carbide steels preferable. I would generally sharpen much more frequently than your average person therefore I find little benefit to high carbide volumes. Actually I'd argue that makes them far less preferable as you lose too many other qualities by gaining the extra carbides. Why would I want a steel that makes it harder for me to resharpen? Fancy stones, more time, etc.

So what you really mean by 'cut better' means that they will cut longer. This isn't necessarily true either because knives do not just blunt by slow wear and there are other factors that influence blunting. Some steels in fact do behave better or worse as far as taking a high sharpness, any steel can get sharp but it depends on how you define 'sharp'. For many simply cutting printer paper equals sharpness, for me I would not really call that knife sharp in a real sense. It's approaching being truly sharp but if it's not popping hairs or cutting newsprint then it's a very low standard.

I don't think I ever said that CATRA tests modeled real world numbers, but they do. The results may be a multiplier, such as *1.5 to get the actual numbers, and they are repeatable, so they are very scientific.

You do not agree that Larrin's model has been proven to actually correlate to real world cutting, but there is a guy who has been doing real world cutting tests on sisal on YouTube for years and they match up with Larrin's CATRA results. I forget the guys name, but his account is "CEDRIC & ADA GEAR AND OUTDOORS" and his tests results are available for FREE.

How much better are new steels vs old ones
https://youtu.be/mo30nNQcHhc

https://youtu.be/mo30nNQcHhc

Full steel testing chart of Cedric & Ada's cut test videos:
https://docs.google.com/spreadsheets/d/ ... d=43566811

Yes, I've seen him and you referencing that testing may in face be proving my point that from what I've seen he mainly cuts clean rope. Has he ever cut dirty materials? Cliff Stamp did a video with his data showing that the difference between an L6 type steel (lower RC, lower carbide) and 10V type steel (higher RC, higher carbide) were very small cutting well used/aged/dirty polypropylene rope.

If you want to see the effect for yourself of dirty materials then simply go outside and find some dirt.... then start digging. Have a look at how well your high carbide steel held that edge compared to $2 chinese kitchen knife. That is worst case scenario. In this example, high grind-ability steel is preferred and high carbide means in general low grindability relative to all other steels.

As I said I really don't have time to go further with this so please don't take this the wrong way if I stop replying to your quoting me here.

[Traditional.Sharpening]

Now hang on, I don’t think the criticism several posts above is fair. If the criticism was meant to be trollish, I shouldn’t respond. But it’s also possible the post in question represents a gross misunderstanding of the scientific process, in which case, someone needs to respond.

The often-heard and well-worn argument that “scientific tests don’t replicate the real world” is a favorite objection—but it doesn’t slow science down any more than a dog chasing a car. If it did, we’d still be living the lifestyle of the late 1600s or earlier.

Scientists refer to the issue raised as external validity. You could make this objection against most scientific studies, maybe all scientific studies, because a good study attempts to eliminate extraneous variables to more precisely determine cause. However, this allows detractors to claim the extraneous variables, being excluded, invalidate the study because it’s not realistic—because it hasn’t included all possible variables. Claiming lack of external validity is kind of the standard, fall-back objection to any study you don’t like.

Of course, if a study didn’t control for extraneous variables, it would be invalid. That would be the problem of including confounds in the study—which is an even worse problem. So basically you can criticize most any single scientific study on the basis of either internal or external validity. It’s a ‘darned if you do and darned if you don’t’ situation. But it’s harder to criticize an entire field of studies—which is still emerging and developing into what might be called ‘knife steel science.’ But it’s silly to criticize an emerging field for not studying every variable; that comes with time and more researchers in the field.

Science proceeds by establishing a beach-head, and then expanding from there. It doesn’t claim to prove anything (and especially not everything), which is why a good scientist “makes no such claims.” It’s not a weakness, as implied; it is how science proceeds. As a science expands, more variables are examined. In the critique above, the variable of lateral loads is mentioned. That’s a variable that Thomas hasn’t methodically tested (to my knowledge) and yes, it would be nice if someone developed a methodology to test it. Typically, the person who objects to a model because it doesn’t incorporate a particular variable, is the person who takes on the project of testing that variable—provided they have the training to do so. I think testing lateral load is a good idea. I’m looking forward to seeing results. But I see no reason to toss previous research because it didn’t include one particular variable. (I doubt anyone else does either.)

Now we need to address the criticism of “not been proven to actually correlate to real world cutting.” (1) As an aside, it’s laymen who say “prove,” whereas scientists are much more conservative, only saying whether a hypothesis is supported or not. (2) More importantly, “real world cutting” is not defined. What is “real world cutting”? Does it vary from person to person? From medium to medium? From environment to environment? “Real world cutting” would need to be made concrete and precise before it could become a testable variable. CATRA testing is one widely accepted partial measure. Nobody says it’s the only measure or even the best measure, it’s just one measure that happens to be admirably objective and precise and widely acknowledged. (3) Why ask for a correlation? If you’re doing good science you want more than a correlation, which is a low standard. You want a formula. Larrin published a regression formula…a startlingly good one, which obtained an R-squared of .79 with a linear association and .83 with a quadratic! A steel’s response to lateral force could be added as a predictive variable to this (or other) equation and it might refine it further (or it might not—results are a gamble). That’s how science progresses, by adding information to the existing. Not by calling previous work “nonsense.”

I think it’s incumbent for critics of science, or of particular scientists, to have a basic understanding of what science is attempting to do and how it proceeds.

CATRA is widely accepted but it's based on an ASSUMPTION that it models performance in general. It always has from the very beginning been based on an assumption that this is how cutting works but we are NOT robots/machines. The human error will always dominate real world use. The problem is really centered on the fact that the assumption is so strong that many including Larrin go so far as to label CATRA rankings as 'edge retention' as though that were the only factor.

I'm not going to get into it in general but as you can hopefully see the CATRA model has serious issues and hasn't even begun to be tested to see if the model is even remotely tied to reality. I'd generally prefer human generated data, not machine cutting data as we are not machines. Human cutting data is not perfect but if done in large enough volume it helps a lot. I'm not going to continue this discussion further here folks, I've spent my time and energy quota here for the week here.

What is your ideal cut test?

Cutting up my daily 10 quarts of vegetables, lol. I don't really bother with any of that stuff anymore, sharpening skill eliminates the crutch of 'needing' a steel that holds an edge 'forever and a day'. I use White #1 in the kitchen at high RC, similar to 1095, I touch it up twice a day between batches with a few passes per side on a regularly conditioned Spyderco Medium Ceramic.

That's good enough for me to keep it cutting at a high sharpness and cutting ability. If I need more edge retention then I simply alter the geometry of the knife bringing it thinner in key areas and changing the micro bevel angles, etc. I don't have time for doing cutting testing in my life anymore so I'm of no help there to others but perhaps there are some here who can do that sort of thing publicly.

[Traditional.Sharpening]

I do not agree as Larrin's model has not been proven to actually correlate to real world cutting. Larrin himself makes no such claims that his tests predict real world results. He simply puts forth the data from his tests, the problem is too many conclusions are being drawn from this data which has not been proven to correlate highly. I cannot understand personally why you'd go to all the trouble of modeling with these tests to not even bother to see if the model translates to reality. Cliff Stamp had directly asked Larrin about this on his forum and Larrin's response was that it was simply assumed that the CATRA correlates to reality. That is not science, that is nonsense to me.

We do not cut like machines and CATRA certainly DOES NOT put any lateral loads on the knives which is a common thing in real world cutting. I will agree that if you generally cut to extreme low levels of sharpness then you may find higher carbide steels preferable. I would generally sharpen much more frequently than your average person therefore I find little benefit to high carbide volumes. Actually I'd argue that makes them far less preferable as you lose too many other qualities by gaining the extra carbides. Why would I want a steel that makes it harder for me to resharpen? Fancy stones, more time, etc.

So what you really mean by 'cut better' means that they will cut longer. This isn't necessarily true either because knives do not just blunt by slow wear and there are other factors that influence blunting. Some steels in fact do behave better or worse as far as taking a high sharpness, any steel can get sharp but it depends on how you define 'sharp'. For many simply cutting printer paper equals sharpness, for me I would not really call that knife sharp in a real sense. It's approaching being truly sharp but if it's not popping hairs or cutting newsprint then it's a very low standard.

I don't think I ever said that CATRA tests modeled real world numbers, but they do. The results may be a multiplier, such as *1.5 to get the actual numbers, and they are repeatable, so they are very scientific.

You do not agree that Larrin's model has been proven to actually correlate to real world cutting, but there is a guy who has been doing real world cutting tests on sisal on YouTube for years and they match up with Larrin's CATRA results. I forget the guys name, but his account is "CEDRIC & ADA GEAR AND OUTDOORS" and his tests results are available for FREE.

How much better are new steels vs old ones
https://youtu.be/mo30nNQcHhc

https://youtu.be/mo30nNQcHhc

Full steel testing chart of Cedric & Ada's cut test videos:
https://docs.google.com/spreadsheets/d/ ... d=43566811

Yes, I've seen him and you referencing that testing may in face be proving my point that from what I've seen he mainly cuts clean rope. Has he ever cut dirty materials? Cliff Stamp did a video with his data showing that the difference between an L6 type steel (lower RC, lower carbide) and 10V type steel (higher RC, higher carbide) were very small cutting well used/aged/dirty polypropylene rope.

If you want to see the effect for yourself of dirty materials then simply go outside and find some dirt.... then start digging. Have a look at how well your high carbide steel held that edge compared to $2 chinese kitchen knife. That is worst case scenario. In this example, high grind-ability steel is preferred and high carbide means in general low grindability relative to all other steels.

As I said I really don't have time to go further with this so please don't take this the wrong way if I stop replying to your quoting me here.

[Traditional.Sharpening]

**double tap**

[Bemo]

I had started a reply last night and thought better of it. I would agree with Bolster that Dr. Thomas seems to address lateral stresses pretty well in Knife Engineering with the descriptions and testing of toughness and fracturing of various steels. The charpy tests sure seems to test lateral toughness to me. Seems we've strayed away from discussing what makes a steel aggressive and how you would possibly test that.

[Deadboxhero]

Cutting up my daily 10 quarts of vegetables, lol.

Be more specific, What vegetables? What cutting board? What edge angle? What was used to sharpen?


If we follow what you do in your vegetable cut test, we will see what you see?

[Bolster]

...Seems we've strayed away from discussing what makes a steel aggressive and how you would possibly test that.

You're right, and I had a hand in that. Sorry. I should stay more topical in a thread I started.

[kennbr34]

I am no steel or sharpening expert. I read what Larrin produces and make good use of the charts. I just know what I like and stick to buying carbide rich steels for knives under 6 inches if at all possible. Carbide rich steels are chosen for a reason, they cut better than the older non carbide steels, not just that they are newer.

As to BESS testers, they ONLY test what is under less than one millimeter of the knife blade and the remainder of the blade is ignored. I think Larrin has a better steel tester that slices cardboard test pieces, and that would be a better tester overall of how a knife edge is going to perform. I just don't know, will a BESS tester show that a high carbide steel cuts better? I do not think that is what it is measuring.

I could be wrong about Larrin's steel testing rig too, don't take my word for it.

What testing rig tests the quality of a cut, employing the length of a knife blade?

Dunno if you got to see this yet...
viewtopic.php?f=2&t=95803

It's an interesting use of a BESS tester, because typically it doesn't measure a slicing cut. But it's a good showcase of how it's possible to use it for such. Otherwise, the other facet of BESS testing is that you don't necessarily have to stick with only one section of the blade. You can measure at different points, and average the results out if wanted.

But BESS tests and CATRA tests differ more in the fact that a BESS test is supposed to quantify a starting sharpness point by measuring the force required to start a cut. Adversely, CATRA uses a static downward force, and tries to measure the amount of cards the edge can cut under that force; they're really not fair to compare since they have different purpose.

There's no reason why you couldn't use the CATRA method to measure push cutting, but the protocol was simply developed to measure slicing cuts. Likewise, BESS was developed to measure the force of a push cut, but as Shawn showed you can adapt it to measure the force of a slicing cut too. I think the real benefit to using the BESS system is simply that it's more feasible to set up a test protocol with it for home users compared to a CATRA machine. So while it might not really be the most optimal thing to use to test wear resistance compared to CATRA, it's still a useful way to generate empirical data in a standardized way. Otherwise we end up limited to tests like Cedric and Ada's rope slicing test, which relies on a subjective evaluation of paper cutting to determine when a blade is dull. However, just as Cedric and Ada's tests seem to reflect Larrin's CATRA tests, I have also found BESS testing to mirror a lot of the same characteristics that CATRA and rope testing tend to demonstrate, such as edge geometry making a larger difference than hardness level.

Take this testing I did of 10V and K294: viewtopic.php?f=15&t=95675

I used a Kizer 10V edge at 27 degrees inclusive, and a K294 blade at 32 degrees inclusive. While I'm not sure of the hardness levels of each, I can make a reasonable assumption that the K294 is at 63-64 HRC based on what Spyderco had ran 10V and A11 at previously. On the other hand, with Kizer being a budget brand, it's unlikely that they hardened the 10V to that level, so one might expect the Spyderco in K294 to greatly outperform the Kizer in 10V. Except that as Larrin's regression formula shows us, a more acute edge geometry will benefit wear resistance to a greater extent than HRC level will, and the results of my testing bare that out with each knife being practically equal in performance.

In addition to corroborating CATRA data and Larrin's regression formula, the data also shows a clear and significant drop off of initial sharpness, that then tends to fall into a sort of "plateau". This mirrors the experience of many people who say carbide rich steels tend not to hold a very fine edge for much longer than their counterparts, but hold a "working" edge almost indefinitely. Prior to this, I only knew that as anecdotal wisdom, but by virtue of the BESS system testing initial sharpness, I now know it empirically as well. A CATRA test on the other hand wouldn't be able to plot that same sudden drop in the level of sharpeness, but instead would only test until it no longer cut the test media under the static downward force.

Going back to Shawn's testing of slicing cuts on tomatoes, I think it shows that a static downward force such as in CATRA could actually be a good way to measure "aggressiveness". For example, since both edges in Shawn's testing cut into the tomato at 198 and 138 grams of force respectively, you could use a static downward force that is enough for both edges to sever the test media, a static length of draw, and then use the depth of the cut as the metric of aggression. Edge Aggression = Depth of Cut / Length of Draw. In that respect, CATRA is already ideally set up to measure aggressiveness, as Edge Aggression = TCC / Draw Cut, as long as the downward force and length of the draw remains static.

[Naperville]

I am no steel or sharpening expert. I read what Larrin produces and make good use of the charts. I just know what I like and stick to buying carbide rich steels for knives under 6 inches if at all possible. Carbide rich steels are chosen for a reason, they cut better than the older non carbide steels, not just that they are newer.

As to BESS testers, they ONLY test what is under less than one millimeter of the knife blade and the remainder of the blade is ignored. I think Larrin has a better steel tester that slices cardboard test pieces, and that would be a better tester overall of how a knife edge is going to perform. I just don't know, will a BESS tester show that a high carbide steel cuts better? I do not think that is what it is measuring.

I could be wrong about Larrin's steel testing rig too, don't take my word for it.

What testing rig tests the quality of a cut, employing the length of a knife blade?

Dunno if you got to see this yet...
viewtopic.php?f=2&t=95803

It's an interesting use of a BESS tester, because typically it doesn't measure a slicing cut. But it's a good showcase of how it's possible to use it for such. Otherwise, the other facet of BESS testing is that you don't necessarily have to stick with only one section of the blade. You can measure at different points, and average the results out if wanted.

But BESS tests and CATRA tests differ more in the fact that a BESS test is supposed to quantify a starting sharpness point by measuring the force required to start a cut. Adversely, CATRA uses a static downward force, and tries to measure the amount of cards the edge can cut under that force; they're really not fair to compare since they have different purpose.

There's no reason why you couldn't use the CATRA method to measure push cutting, but the protocol was simply developed to measure slicing cuts. Likewise, BESS was developed to measure the force of a push cut, but as Shawn showed you can adapt it to measure the force of a slicing cut too. I think the real benefit to using the BESS system is simply that it's more feasible to set up a test protocol with it for home users compared to a CATRA machine. So while it might not really be the most optimal thing to use to test wear resistance compared to CATRA, it's still a useful way to generate empirical data in a standardized way. Otherwise we end up limited to tests like Cedric and Ada's rope slicing test, which relies on a subjective evaluation of paper cutting to determine when a blade is dull. However, just as Cedric and Ada's tests seem to reflect Larrin's CATRA tests, I have also found BESS testing to mirror a lot of the same characteristics that CATRA and rope testing tend to demonstrate, such as edge geometry making a larger difference than hardness level.

Take this testing I did of 10V and K294: viewtopic.php?f=15&t=95675

I used a Kizer 10V edge at 27 degrees inclusive, and a K294 blade at 32 degrees inclusive. While I'm not sure of the hardness levels of each, I can make a reasonable assumption that the K294 is at 63-64 HRC based on what Spyderco had ran 10V and A11 at previously. On the other hand, with Kizer being a budget brand, it's unlikely that they hardened the 10V to that level, so one might expect the Spyderco in K294 to greatly outperform the Kizer in 10V. Except that as Larrin's regression formula shows us, a more acute edge geometry will benefit wear resistance to a greater extent than HRC level will, and the results of my testing bare that out with each knife being practically equal in performance.

In addition to corroborating CATRA data and Larrin's regression formula, the data also shows a clear and significant drop off of initial sharpness, that then tends to fall into a sort of "plateau". This mirrors the experience of many people who say carbide rich steels tend not to hold a very fine edge for much longer than their counterparts, but hold a "working" edge almost indefinitely. Prior to this, I only knew that as anecdotal wisdom, but by virtue of the BESS system testing initial sharpness, I now know it empirically as well. A CATRA test on the other hand wouldn't be able to plot that same sudden drop in the level of sharpeness, but instead would only test until it no longer cut the test media under the static downward force.

Going back to Shawn's testing of slicing cuts on tomatoes, I think it shows that a static downward force such as in CATRA could actually be a good way to measure "aggressiveness". For example, since both edges in Shawn's testing cut into the tomato at 198 and 138 grams of force respectively, you could use a static downward force that is enough for both edges to sever the test media, a static length of draw, and then use the depth of the cut as the metric of aggression. Edge Aggression = Depth of Cut / Length of Draw. In that respect, CATRA is already ideally set up to measure aggressiveness, as Edge Aggression = TCC / Draw Cut, as long as the downward force and length of the draw remains static.

I like the detailed write-up! If BESS can measure a slicing cut, it can tell the difference in steels, not just edges.

[Deadboxhero]

I am no steel or sharpening expert. I read what Larrin produces and make good use of the charts. I just know what I like and stick to buying carbide rich steels for knives under 6 inches if at all possible. Carbide rich steels are chosen for a reason, they cut better than the older non carbide steels, not just that they are newer.

As to BESS testers, they ONLY test what is under less than one millimeter of the knife blade and the remainder of the blade is ignored. I think Larrin has a better steel tester that slices cardboard test pieces, and that would be a better tester overall of how a knife edge is going to perform. I just don't know, will a BESS tester show that a high carbide steel cuts better? I do not think that is what it is measuring.

I could be wrong about Larrin's steel testing rig too, don't take my word for it.

What testing rig tests the quality of a cut, employing the length of a knife blade?

Dunno if you got to see this yet...
viewtopic.php?f=2&t=95803

It's an interesting use of a BESS tester, because typically it doesn't measure a slicing cut. But it's a good showcase of how it's possible to use it for such. Otherwise, the other facet of BESS testing is that you don't necessarily have to stick with only one section of the blade. You can measure at different points, and average the results out if wanted.

But BESS tests and CATRA tests differ more in the fact that a BESS test is supposed to quantify a starting sharpness point by measuring the force required to start a cut. Adversely, CATRA uses a static downward force, and tries to measure the amount of cards the edge can cut under that force; they're really not fair to compare since they have different purpose.

There's no reason why you couldn't use the CATRA method to measure push cutting, but the protocol was simply developed to measure slicing cuts. Likewise, BESS was developed to measure the force of a push cut, but as Shawn showed you can adapt it to measure the force of a slicing cut too. I think the real benefit to using the BESS system is simply that it's more feasible to set up a test protocol with it for home users compared to a CATRA machine. So while it might not really be the most optimal thing to use to test wear resistance compared to CATRA, it's still a useful way to generate empirical data in a standardized way. Otherwise we end up limited to tests like Cedric and Ada's rope slicing test, which relies on a subjective evaluation of paper cutting to determine when a blade is dull. However, just as Cedric and Ada's tests seem to reflect Larrin's CATRA tests, I have also found BESS testing to mirror a lot of the same characteristics that CATRA and rope testing tend to demonstrate, such as edge geometry making a larger difference than hardness level.

Take this testing I did of 10V and K294: viewtopic.php?f=15&t=95675

I used a Kizer 10V edge at 27 degrees inclusive, and a K294 blade at 32 degrees inclusive. While I'm not sure of the hardness levels of each, I can make a reasonable assumption that the K294 is at 63-64 HRC based on what Spyderco had ran 10V and A11 at previously. On the other hand, with Kizer being a budget brand, it's unlikely that they hardened the 10V to that level, so one might expect the Spyderco in K294 to greatly outperform the Kizer in 10V. Except that as Larrin's regression formula shows us, a more acute edge geometry will benefit wear resistance to a greater extent than HRC level will, and the results of my testing bare that out with each knife being practically equal in performance.

In addition to corroborating CATRA data and Larrin's regression formula, the data also shows a clear and significant drop off of initial sharpness, that then tends to fall into a sort of "plateau". This mirrors the experience of many people who say carbide rich steels tend not to hold a very fine edge for much longer than their counterparts, but hold a "working" edge almost indefinitely. Prior to this, I only knew that as anecdotal wisdom, but by virtue of the BESS system testing initial sharpness, I now know it empirically as well. A CATRA test on the other hand wouldn't be able to plot that same sudden drop in the level of sharpeness, but instead would only test until it no longer cut the test media under the static downward force.

Going back to Shawn's testing of slicing cuts on tomatoes, I think it shows that a static downward force such as in CATRA could actually be a good way to measure "aggressiveness". For example, since both edges in Shawn's testing cut into the tomato at 198 and 138 grams of force respectively, you could use a static downward force that is enough for both edges to sever the test media, a static length of draw, and then use the depth of the cut as the metric of aggression. Edge Aggression = Depth of Cut / Length of Draw. In that respect, CATRA is already ideally set up to measure aggressiveness, as Edge Aggression = TCC / Draw Cut, as long as the downward force and length of the draw remains static.

Good job on the 10v testing.


I did a similar test using paramilitary 2s with different steels.

You might have missed it since it was posted before you joined this forum.
viewtopic.php?f=2&t=91530&hilit=Pm2+edge+retention

It does show much like CATRA that higher hardness with more carbide volume and harder carbides was the winner in rope cut testing, and that 52100 despite having high hardness came in last due to having low carbide volume and the softest carbides.

[Bolster]

Awesome, depthy post. Thanks Kennbr! Well reasoned and informative.

If you don't mind taking a dive into nerd-ville with me to further illuminate just one small point:

...as Larrin's regression formula shows us, a more acute edge geometry will benefit wear resistance to a greater extent than HRC level will...

^ This may have come from examining the coefficients in Larrin's regression formula, where we see 15.8*Hardness (Rc) and -17.8*EdgeAngle(°). What Larrin published here are called "B"s or unstandardized coefficients. They're what are used to create a formula and to calculate slope. "B"s are in the unit of their metric, Rc and degrees and percents. The rub is that Rc goes up to ... whatever it goes up to (70? dunno) and there are 360 degrees and the scale for percent is 100. So these coefficients are on different scales, making comparison troublesome for different-unit Bs in terms of relative strength/size/power/importance. (You could compare all the Bs for element percentages against each other since they're in the same metric, percent.)

A regression analysis also gives coefficients in a different form called "Betas." These are standardized, for the very purpose of being able to compare them to each other. (1 SD change in X for the corresponding SD change in Y.) The Betas are useless for creating a formula; they only serve one purpose, and that is to compare them to each other to determine relative strength/size/power/importance. Since Larrin didn't publish Betas, we really don't know the amount by which one variable compares to (or "beats") another variable in affecting wear resistance. It would be interesting to see the Betas. It may be that the relative power of the variables is larger than expected by examining Bs.

Given the volley of rotten tomatoes and cabbages that are certain to follow the above nerdiness, I will now exit the stage. Again, great post, thanks for taking the time to write it out.

[kennbr34]

Awesome, depthy post. Thanks Kennbr! Well reasoned and informative.

If you don't mind taking a dive into nerd-ville with me on just one small point:

...as Larrin's regression formula shows us, a more acute edge geometry will benefit wear resistance to a greater extent than HRC level will...

^ This may have come from examining the coefficients in Larrin's regression formula, where we see 15.8*Hardness (Rc) and -17.8*EdgeAngle(°). What Larrin published here are called "B"s or unstandardized coefficients. They're what are used to create a formula and to calculate slope. "B"s are in the unit of their metric, Rc and degrees and percents. The rub is that Rc goes up to ... whatever it goes up to (70? dunno) and there are 360 degrees and the scale for percent is 100. So these coefficients are on different scales, making comparison troublesome for different-unit Bs in terms of relative strength/size/power/importance. (You could compare all the Bs for element percentages against each other since they're in the same metric, percent.)

A regression analysis also gives coefficients in a different form called "Betas." These are standardized, for the very purpose of being able to compare them to each other. (1 SD change in X for the corresponding SD change in Y.) The Betas are useless for creating a formula; they only serve one purpose, and that is to compare them to each other to determine relative strength/size/power/importance. Since Larrin didn't publish Betas, we really don't know the amount by which one variable compares to another variable in affecting wear resistance. It would be interesting to see the Betas.

Given the volley of rotten tomatoes and cabbages that are certain to follow the above nerdiness, I will now exit the stage. Again, great post, thanks for taking the time to write it out.

I would be dishonest if I pretended to understand what you just said Math, and statistics especially, are not my forte.

Perhaps my error was in my wording... Am I still incorrect by saying that his formula shows that edge geometry affects wear resistance more dramatically than the HRC on number-to-number basis? In other words, a difference of 5 degrees seems to have a much greater affect on the outcome than a difference of 5 HRC points.

[Bolster]

Perhaps my error was in my wording... Am I still incorrect by saying that his formula shows that edge geometry affects wear resistance more dramatically than the HRC on number-to-number basis? In other words, a difference of 5 degrees seems to have a much greater affect on the outcome than a difference of 5 HRC points.

I think you are correct. The units of degrees are "smaller" than units of Rc. I don't think you made any error, just a refinement that the two coefficients aren't directly comparable in their B form. But they could be, if we had their Beta form. My comment is more of a nudge for Larrin to publish his Betas.

Hey, since we are in nerd-ville already, one other observation. Might the relationship for degrees be curvilinear, not linear? Do we really expect edge efficacy to continually increase, the more acute the angle? Wouldn't there be an asymptote somewhere? Dunno. If it were my dataset, I would check that.

[kennbr34]

Perhaps my error was in my wording... Am I still incorrect by saying that his formula shows that edge geometry affects wear resistance more dramatically than the HRC on number-to-number basis? In other words, a difference of 5 degrees seems to have a much greater affect on the outcome than a difference of 5 HRC points.

I think you are correct. The units of degrees are "smaller" than units of Rc. I don't think you made any error, just a refinement that the two coefficients aren't directly comparable in their B form. But they could be, if we had their Beta form. My comment is more of a nudge for Larrin to publish his Betas.

Ahh, gotcha. I'll leave the math talk to you gentlemen

Though... Just wondering. Is it just coincidence that the effective range of HRC values in cutlery steel is about 55-71 (i.e. 16 points) HRC, and that the "beta" for that in his formula is 15.8?

Hey, since we are in nerd-ville already, one other observation. Might the relationship for degrees be curvilinear, not linear? Do we really expect edge efficacy to continually increase, the more acute the angle? Wouldn't there be an asymptote somewhere? Dunno. If it were my dataset, I would check that.

I can't answer this mathematically, but I can say for sure that edge there's definitely a point where edge stability falls off significantly the more acute you get. I've observed this with trying to do BESS testing on straight razors, for example. Their edges are so thin, the BESS media just dents them rather than being cut. I've also seen things like pocket knife steels chipping out merely from cutting paper if the edge is too thin.

Anyway, I think we're getting pretty far away from the subject of aggressiveness again, but interesting discussion nontheless.

[Bolster]

^ Agreed, I'm wandering OT again (sorry), so I'll keep it short (ish). (1) I'd say mere coincidence that the range of Rc is similar to the B of 15.8. One is an arbitrary range, and the other is an amount of change in Y for a one unit change in X. I can't see how they'd be related. Recall the B doesn't stand alone, it relies on the intercept (-157) and all the other Bs in the equation. If you remove or add a variable, all the Bs (and the intercept) will change somewhat. (2) The fall-off in ability to cut that you observed on very acute edges would indicate a curvilinear relationship at the extremes of the range. What Larrin checked was a linear association (I think), which is probably perfectly acceptable for a "reasonable" range of edge angles. So no criticism of Larrin for testing & publishing linear, but folks need to understand the danger of extrapolating beyond the data swarm. For this one variable of edge angle, I'd plot it out, to see where the linear trend petered out (provided Larrin even tested any extreme edge angles). You don't want to give folks the impression they can keep acute-ing their edges indefinitely, if there's an asymptote, ie, a bend in the line (which could result in a bend in your edge! ha ha).

[barnaclesonaboat]

I did a similar test using paramilitary 2s with different steels.

You might have missed it since it was posted before you joined this forum.
viewtopic.php?f=2&t=91530&hilit=Pm2+edge+retention

It does show much like CATRA that higher hardness with more carbide volume and harder carbides was the winner in rope cut testing, and that 52100 despite having high hardness came in last due to having low carbide volume and the softest carbides.

Wow! Big thanks for linking up this thread, as I also joined after and had not stumbled upon it. And thanks for your dedication to pro bono testing like that/this. This was exactly the type of repeated-measures testing I was wishing for more of after checking out kennbr34's cool dataset from a few weeks ago. It really seems like there is a lot we can learn from repeated BESS testing over the course of experiments that evaluate edge retention, etc., towards developing a richer body of empirical evidence for understanding knife steels. Regardless, cool where this and related discussions have gone in the last few weeks.

[Deadboxhero]

I did a similar test using paramilitary 2s with different steels.

You might have missed it since it was posted before you joined this forum.
viewtopic.php?f=2&t=91530&hilit=Pm2+edge+retention

It does show much like CATRA that higher hardness with more carbide volume and harder carbides was the winner in rope cut testing, and that 52100 despite having high hardness came in last due to having low carbide volume and the softest carbides.

Wow! Big thanks for linking up this thread, as I also joined after and had not stumbled upon it. And thanks for your dedication to pro bono testing like that/this. This was exactly the type of repeated-measures testing I was wishing for more of after checking out kennbr34's cool dataset from a few weeks ago. It really seems like there is a lot we can learn from repeated BESS testing over the course of experiments that evaluate edge retention, etc., towards developing a richer body of empirical evidence for understanding knife steels. Regardless, cool where this and related discussions have gone in the last few weeks.

Dr Larrin made his analysis on the cut testing with BESS recordings back in February 2022.



This is a clip from his video "What is the best hardness for MagnaCut?"