The Best Steel - Cut Tests and Opinions

[marcinek]

So what your saying is cliffs stamp and ankerson tests are also useless. Good to know.

In any case its data. I use all the tests including those in Russia that I can't understand but can read the results. They are all pretty similar and align well with catra in some cases.

1) Who said that?
2) Cliff Stamp is a wanker. It's been proven by science.
3) If you accept data where you cant understand where it came from, or don't care where it came from, is entirely your prerogative. But as a knife user, and one who thinks he is, at east moderately sophisticated, I can't sit idly by when anecdotal observations are passed off as data here...that does a great disservice to new folks who want their data to be based on science.

 

[Tommy-Chi]

This.

And even at the same hardness identical steel still can behave very diffidently due to the internal structure... the size and the place of steel grain, carbide, characteristic of martensite formed etc. are all have so much influence.

I agree! Hear Ye - Hear Ye! Let no man do steel cut tests until the subatomic structure of the material is exactly the same for any particular production model, or people will be sold ... Therefore I suggest mankind set aside all of these silly tests until which time the Star Trek Replicator has been produced.

That's just insulting and rude. We are having a measured discussion about the scientific methods behind the results presented here, and your contribution is "Haters gonna hate."

So you think everyone who has raised their objections to the methods presented has done so for the sole reason that they are mindless "haters" who can't help themselves?

Nobody knows what a controlled experiment is? Nobody objects to inaccurate data and unsubstantiated conclusions?

Of course you are welcome to accept the conclusions given without thinking about what they are based on, but please dont insult the people here who do choose to think about what they are based on.

I've been using Blade Forums for years, far longer than I have been a member and there is ZERO evidence of any scientific methods on Blade Forums that is accurate.

Remember THIS ONE? The Blade Forum members were improperly using the device according to the manufacturer...so there are no scientific results we can look at. Everything has to be taken with a grain of salt and in it's entirety, the body of work reveals TRENDS.

 

[marcinek]

I agree! Hear Ye - Hear Ye! Let no man do steel cut tests until the subatomic structure of the material is exactly the same for any particular production model, or people will be sold ... Therefore I suggest mankind set aside all of these silly tests until which time the Star Trek Replicator has been produced.

Shame a knife manufacturer doesn't take the same knife and make it in different steels so that the steels can be tested more scientifically.

 

[Danke42]

I don't have a dog in this fight, I just have a question. When you mention this youtuber as a social media influencer are you including YouTube as one such form of social media? Not a trick question I'm just wondering if you're referring to something else.
I've watched some of his knife comparisons and reviews, but I too personally stay away from the testing results I'm not sure for the same reasons but regardless I hold scientific objectivity dear to my heart. At the same time though, I'm not so harshly critical to a well-meaning goofy guy who doesn't claim to be definitively testing anything just giving his own personal test. I'm sorry I don't mean to argue too much I just think you can say what you have to say without being so scornful but that's just my humble opinion only.

I don't have a dog in this fight, I just have a question. When you mention this youtuber as a social media influencer are you including YouTube as one such form of social media? Not a trick question I'm just wondering if you're referring to something else.
I've watched some of his knife comparisons and reviews, but I too personally stay away from the testing results I'm not sure for the same reasons but regardless I hold scientific objectivity dear to my heart. At the same time though, I'm not so harshly critical to a well-meaning goofy guy who doesn't claim to be definitively testing anything just giving his own personal test. I'm sorry I don't mean to argue too much I just think you can say what you have to say without being so scornful but that's just my humble opinion only.

Yes YouTube is a social media channel. A huge one.

No scorn; just my "opinion". Scorn is saved for when posting in the more emotion based spots on the site.

 

[DeadboxHero]

Use the scientific method to setup a controlled experiment.

The tester won't know what the steel type or knife brand is. The tests will be preformed by a method that eliminates human error and fatigue.

The tests will be repeatable and the results will be subject to peer review.

I don't disagree, get to it. I'll peer review.

Also where's that knife?

The Amsterdam met is over.

Any progress?

 

[WilliamEdward]

Yes YouTube is a social media channel. A huge one.

No scorn; just my "opinion". Scorn is saved for when posting in the more emotion based spots on the site.

Yep just asking if that's the social media you were referring too. And no worries! Just giving my own opinion as well.

 

[bikerector]

Use the scientific method to setup a controlled experiment.

The tester won't know what the steel type or knife brand is. The tests will be preformed by a method that eliminates human error and fatigue.

The tests will be repeatable and the results will be subject to peer review.

Apparently none of my college chemistry, math, or physics professors understood science. We never used blind studies and made plenty of scientific, analytical based assertions about the found data. I've never used a blinded study unless people were the test subjects and that was more in psychology courses but I never got far in the soft sciences until more recently and I'm not actually doing any studies with it, just learning concepts.

Standardizing a test is certainly desirable, and pretty much required, but a blinded study is not a requirement of scientific discovery. Bias allows for the possibility of introducing error but does not guarantee it and that amount of error does not immediately dismiss the findings of value.

We could argue that the test method is flawed and has error but it's a judgment call on if the expected error is high enough to hold some level of scientific value. I will argue that there are enough constraints to the experiment to hold some scientific evidence of a trend even if the resolution is poor. Shoot, the testers could simply perform the exact same study and simply repeat it over a larger number of iterations and the resolution becomes more clear even if the studies framework remains the same. Then you get your reproducibility and with some statistics we get to see how accurate and precise the tests really are.

You could re-order the blades used each iteration so the same sequence is altered to see if fatigue is a meaningful contributor to error.

P.S. Technically, your study design isn't double-blind because there is no test subject/participant to put forth bias, only a tester/researcher.

 

[Twindog]

I think the problem is that people underestimate how difficult and how complex it is to compare the wear resistance of different knife steels. Hardness plays a big factor. Edge angle plays a big factor. Edge width plays a huge factor. Micro-structure of the steel plays a big factor. Sharpening technique plays a big role. And the testing technique plays a big factor.

Given all the possible combinations of error from testing things that are not consistent on all factors or that suffer from poor technique creates an almost infinite number of outcomes.

There are scientific ways to measure steels for wear resistance that are consistent, but they are valid only for the variables tested -- not the infinite number of other outcomes if different variables were chosen.

It's pretty cool of the OP to take the time to winnow through the key data of one particular person's testing. But without more consistent controls, the data are just all over the place.

K390 is better than M4 for wear resistance -- not likely.
M4 is better than M390 on wear resistance -- not likely.
M390 is equal to S90V -- not likely.
M390 is better than 10V -- not likely.
S30V is better than K390 -- not in your dreams.
3V is better than S35VN -- not likely.
A2 is better than S35VN -- neva happen GI.
A2 has more wear resistance than S30V -- nope.

So with all the obvious problems and with various steels being put into so many different categories, the results are more confusing than useful.

 

[Danke42]

Apparently none of my college chemistry, math, or physics professors understood science. We never used blind studies and made plenty of scientific, analytical based assertions about the found data. I've never used a blinded study unless people were the test subjects and that was more in psychology courses but I never got far in the soft sciences until more recently and I'm not actually doing any studies with it, just learning concepts.

Standardizing a test is certainly desirable, and pretty much required, but a blinded study is not a requirement of scientific discovery. Bias allows for the possibility of introducing error but does not guarantee it and that amount of error does not immediately dismiss the findings of value.

We could argue that the test method is flawed and has error but it's a judgment call on if the expected error is high enough to hold some level of scientific value. I will argue that there are enough constraints to the experiment to hold some scientific evidence of a trend even if the resolution is poor. Shoot, the testers could simply perform the exact same study and simply repeat it over a larger number of iterations and the resolution becomes more clear even if the studies framework remains the same. Then you get your reproducibility and with some statistics we get to see how accurate and precise the tests really are.

You could re-order the blades used each iteration so the same sequence is altered to see if fatigue is a meaningful contributor to error.

P.S. Technically, your study design isn't double-blind because there is no test subject/participant to put forth bias, only a tester/researcher.

Seeing as I did not got into the minutia of setting up every step of the "experiment" It seems like you have read a lot into how it would work. What you've done is commonly called assuming.

To recap, what I want out of test is very simple.

I don't want to see the cuts performed my a human hand because that is inconsistent over then number of cuts used int this type of test.

I don't want the tester to know what they are testing. That remove the chance of "wow I really love brand X knives and Y steel is my all time favorite.

These requirements are not onerous but I"m pretty sure they'd be dull to watch on YouTube.

Make an observation
Conduct research
Form Hypothesis
Test Hypothesis
Record Data
Draw conclusion
Replicate.

Unfortunately draw conclusion seems to be the start and end of the social media method. If that entertains people that's fine but it's a major mistake to pretend that the results of that entertainment are facts.

 

[Danke42]

I don't disagree, get to it. I'll peer review.

Also where's that knife?

The Amsterdam met is over.

Any progress?

You might be asking the wrong person this. I was never in Amsterdam. I was in the Billiard Room with a knife playing billiards as I recall.

 

[Tommy-Chi]

K390 is better than M4 for wear resistance -- not likely.
M4 is better than M390 on wear resistance -- not likely.
M390 is equal to S90V -- not likely.
M390 is better than 10V -- not likely.
S30V is better than K390 -- not in your dreams.
3V is better than S35VN -- not likely.
A2 is better than S35VN -- neva happen GI.
A2 has more wear resistance than S30V -- nope.

At the outset I did claim that due to Pete's own errors the list needs to be taken less than literally with at least a plus/minus of ten cuts or more. Also, don't forget that not every manufacturer optimizes heat treatment and tempering of steels.

Are you pulling your "list" from the list that I posted? I think you need to go back and read how he arrived at the data. I only looked at the K390 and M4 and it is correct according to your own statement.

Pete is comparing TWO variables simultaneously. I'm not his production manager, and I'd suggest against that, but that is what the data represents.

K390 and M4 are presented in the list twice. The first time they are listed with a 20 degree Worksharp edges, and then they are both listed with a 17 degrees Worksharp edge.

 

[Tommy-Chi]

Shame a knife manufacturer doesn't take the same knife and make it in different steels so that the steels can be tested more scientifically.

Spyderco should never have stopped producing Mule Team blanks for any of their steels. The only people that have the full collection are steel collectors, not edge test scientists.

 

[DeadboxHero]

You might be asking the wrong person this. I was never in Amsterdam. I was in the Billiard Room with a knife playing billiards as I recall.

Regardless, any progress on your knife?

 

[Twindog]

I read your list again, and the conclusions don't change. In your "unleashed" data, you have

M4 at 489 cuts.
K390 at 480
Elmax at 320

Those results are far outside published, scientific CATRA testing.

Bohler Uddeholm published these CATRA (scientific) results:
M390 at 61+ Rc got 958 cuts
Elmax at 62 Rc got 930 cuts
M4 at 61 Rc got 899 cuts
D2 at 61 Rc got 674 cuts
http://www.bucorp.com/media/CATRA_Test2.pdf

They don't list K390 in that testing data, but K390 is an A11-class steel. It's substantially tougher than 10V, another A11-class steel, but the wear resistance is said by BU to be basically equal.

So comparing an actual scientific test to just your "unleashed" numbers show major differences.

I appreciate your effort, but when I saw how poorly your data comport with far more rigorous data, I lost interest.

At the outset I did claim that due to Pete's own errors the list needs to be taken less than literally with at least a plus/minus of ten cuts or more. Also, don't forget that not every manufacturer optimizes heat treatment and tempering of steels.

Are you pulling your "list" from the list that I posted? I think you need to go back and read how he arrived at the data. I only looked at the K390 and M4 and it is correct according to your own statement.

Pete is comparing TWO variables simultaneously. I'm not his production manager, and I'd suggest against that, but that is what the data represents.

K390 and M4 are presented in the list twice. The first time they are listed with a 20 degree Worksharp edges, and then they are both listed with a 17 degrees Worksharp edge.

 

[Danke42]

Regardless, any progress on your knife?

Closer every day.

 

[Yonose]

Exactly right, I agree with everything you said. Pete's tests are not scientific. But he doesn't claim that they are.

His tests are useful though as long as you don't read too much into them. They do help with everyday knife buying decisions.

For example, how much more should I spend to buy a knife with a premium steel like M390? Everyone would probably agree M390 is better than S30V. But how much better? If M390 is just a little bit better than S30V, then it probably isn't worth the extra money (all other factors being equal). But Pete's tests show that M390 is significantly better in edge retention, and this helps me make a more informed buying decision.

An upgrade from 420HC to M390 is about $40 for Kershaw’s Link or Dividend. Wouldn’t consider the blade in 420HC at $40. From s30v to M390 the most significant advantages are ease of sharpening and toughness, from what I understand, and probably not worth paying double for. But I see the pm2 in m390 being sold on the exchange for $240, about double the cost of the pm2 in s30v. This I don’t understand.

 

[Yonose]

1) Who said that?
2) Cliff Stamp is a wanker. It's been proven by science.
3) If you accept data where you cant understand where it came from, or don't care where it came from, is entirely your prerogative. But as a knife user, and one who thinks he is, at east moderately sophisticated, I can't sit idly by when anecdotal observations are passed off as data here...that does a great disservice to new folks who want their data to be based on science.

I can’t argue with point#2 since I don’t know the man, but his tests (and Ankerson’s)are at least as rigorous as what you were demanding from the average tester. If their tests aren’t “scientific” then whose tests are?

 

[me2]

Why would anyone base steel choice decisions on what was referred to as "bro science"? Publishing results means the tester had a certain level of confidence they are correct and the methods are sound. That is science. Saying published results aren't scientific means the tester doesn't want criticism and won't/can't defend/support the conclusions/methods. Or, it means the tester doesn't think the conclusions are reliable.

On another point, testing of hand held knives should include hand held testing. Testing of ceramic knives in CATRA shows huge advantages of edge holding over steel. However, ceramics cannot take the side loads and twisting common in hand held use. Not including hand held testing on a knife intended fort hand held use is an example of a poorly designed experiment that can lead to erroneous conclusions.

 

[marcinek]

I can’t argue with point#2 since I don’t know the man, but his tests (and Ankerson’s)are at least as rigorous as what you were demanding from the average tester. If their tests aren’t “scientific” then whose tests are?

Ankerson's testing methods have not been described in this thread, but I do know they have been subjected to years of peer review here on an entirely open forum and are highly respected.

You and the rest of the "nonhaters" really should relax though and be patient, it certainly does appear we are nearing the point where the scientific method is replaced with the number of Likes your YouTube video gets.

 

[Mr.SATism]

TLDR: Testing is great for letting us know the value of a knife

CATRA, Ankerson, Creely, Knifeinformer, C&A, and the one guy who tests edge retention on kitchen knives on YT (I kid you not, and it's kind of cool TBH). These are the sources that I use when I evaluate knife purchases (Most of these tests have the edge retention tests going from sharp to dull, and I do appreciate seeing a dull blade at the end of the tests)

To me, all of these results give me a GENERAL IMAGE of what blade steels can and cannot do in terms of EDGE RETENTION. Now does this mean that every one of these tests are the words of G-d? To me, no. Not because something is wrong with the tests, but because 99% of the time, I don't know what happened to the blade from bar of steel, to fully finished product.

This goes for all the above testers, but there are too many variables to give me a 100% guarantee that their tests will be accurate for the knife that I am holding in my hand. I can only hope that what I hold will perform the same as these tests.

With the world being the way it is, I can get an ultra-premium, super nitro volcano heat quenched, kissed by the maker blade and still have it not last as long as people say it does. Or the other way around, I can get a low-end steel that might perform above and beyond what I expected. Or I can just not sharpen it in a way that would give me maximum edge retention. But there's a 99% chance that I will never even come close to having a super steel go dull on me with light->medium use.

So does that mean that all these tests are irrelevant? Nope. It means that I have multiple sources of data that I can use. It also lets me see the value of each steel in a knife.

For example: let's say that I want to purchase a 4" production auto knife as a heavy use EDC, I can look at the market and see that I have some options in stainless steel and some options in CPM154 (yes we have more options, but let's just focus on these two for now). I can hop on all of these sources, and see how the steels performed so I know which one would last me longer (let's say I don't strop as well). So for edge retention exclusively, I would pick CPM154.

Bear with me for just two more paragraphs lol.

So now that I know what the price range for a desirable steel is (CPM154 in this case), I would know that for steel exclusively I would not pay the same price as I would for a CPM154 blade.

Yes, this example includes some extremes, ignoring: production quality, company value, sharpening job from the factory, etc... But what makes this interesting are knives of the same model in different steels. I assume that most, if not all of us are aware of Spyderco's sprint/mule practices so I am going to cut this off short.