Lock Picking Spectrum of Difficulty

[xeo]

The difference would be subjective vs scientifically calculated. In which case the difference is always huge.

[GringoLocksmith]

Difficulty is usually subjective.

[Josephus]

Subjective and objective measurements are not what they appear to be prima facie. It is somewhat a modern myth that whatever objective is both outside human experience and better than subjective measurements. All that objectivity brings is a standard to measure against and subjectivity does not. How that external objective measure is created becomes troublesome.

Scientific measurements, when not attempting to derive a categorical fact, are not at all objective. How do you define 'difficulty'? What makes you think that any specific feature always increases difficulty for every person? Does a combination of metrics that individually increase difficulty, when combined, actually decrease it overall? Some features that clearly increase difficulty in one model clearly decreases it in another. Further, for some people different features or metrics that appear more secure actually make locks easier for them. How do you know that what metric or weighting you chose is correct? What methodology will it take to create such things? It becomes a 'turtles all the way down' scenario when trying to find an objective way to create an objective measure. You must then account not only for the metrics to be used and how those will be measured, but how those metrics will be chosen, why they will be chosen, how to prove that those measures really are objective (they aren't), and how to prove that they are a viable measurement in the first place. At some point you will hit a limitation of knowledge. At some point some feature is more difficult because it feels more difficult and for no other reason. There is nothing scientifically empirical about that, or is there?

There is no such dichotomy between subjective measurements vs empirical means. A person that really likes cake can test each type of cake available and decide on what sorts of cake he likes best. Then the conclusions found from that subjective testing can be used for future predictions over what new types of cake he would most likely enjoy. I know I do not like things with mushrooms in them. It is entirely subjective on my part. I can use that single data point, 'I do not like mushrooms', to predict and test new things. If something has mushrooms in it, I probably wont like it. Unless they are made of brass.

Even if the limitations of knowledge, subjectivity in methodology, and acceptance of empirical measures based on personal taste, there is the problem of rumination or iteration. So you have a good methodology of choice down, and you have a methodology of testing, a way to store the results, and a way to utilize them in some practical fashion. That still leaves a lack of bounds problem. When do you stop adding metrics? When is the weighting good enough? Where is it accurate enough? How do you measure that accuracy? With metric standards there is always a forced choice in limitation of bounds. There will always be some point where you must say 'that is accurate enough'. A database with 10,000 options on how a lock can be more or less secure isn't at all useful, but it does have the possibility of being more accurate than a metric of one option. Secondly, with changes in an environment things that once were difficult may no longer be. Twenty years ago Medeco would have been ranked higher than other brands that are considered its superior today. Suddenly all those metrics and methodologies have to be thrown out or reworked on an ongoing basis to maintain even remotely the same accuracy. Instead of increasing in accuracy over time (as ranking systems do) metric systems diverge from the mean in unusual, unpredictable ways. This is one of the more significant issues with technical stock analysis. The validity of metrics are moving nearly as fast as the data.

The alternative to this is the basis of the thread already. An evolutionary ranking system. It bypasses the difficulties of what measure should be chosen, why is it chosen, how is it chosen, and the limitation of knowledge problem. It maintains the ability to predict outcomes based on past data. Being able to know when the job is done is the prime factor in completion. With metrics, there is no done. There is no cessation of work. You could continue on from here to the end of time with the project without ever completing it. Ranking systems even get more accurate over time instead of less. Of course a 'hot or not' lock page certainly would be way more fun to use than downloading a relational database, but that is just my subjectivity showing.

[xeo]

This isn't cake testing where taste is subjective. This has an objective, literally, to open the lock with tools. The metrics that go into determining how difficult it is to complete the objective ARE objective. The interpretations of the metrics are not subjective, they are stringent, objective and clear. The actions required to open the lock are also not subjective. You must satisfy X requirements using tools and you are bound by the security parameters and metrics of the lock and those requirements vary in difficulty simply by nature of existing or not existing and by how many of them exist. Everything about a lock is bound within an objective framework and boundary therefore I think an objective grading system is the way to go. Does this always equal reality 100% No. No calculation is perfect when there are so many variables. People are still trying to perfect formulas for ranking and grading that have been in development for years. I think going the evolutionary/subjective route opens you up to skewed data from inexperienced pickers, confirmation bias and a whole host of other nasty things that will end up giving you a poor quality final product.

All of that aside, using a relational approach like I suggested allows you to query the data in unique ways and will allow you to generate metadata that may surprise you.

[elbowmacaroni]

I think someone may have read "Zen and the Art of Motorcycle Maintenance" one too many times and is transposing quality into difficulty.

[flywheel]

-Asking for help from any computer programmers-

Here is to LockMash™, or whatever name it ends up with!
Thanks!

Good day folks. I haven't had any responses yet to my request for help. I did a cursory search and found that facemash type programs exist out there but need to be created offsite. I'm thinking the whole framework could be copied to this site and tweaked for our needs. I can't help with that but am willing to begin collecting, resizing, and labeling lock pictures for a database.

I'll just dip my toes in the murky waters and say if 100-200 lock pickers (let's call them KP members, coincidentally the only lock pickers I know) told me an American is harder to pick than a 4-pin "made in china" lock I would be inclined to listen. The same goes for 10 super experienced and skilled pickers telling me to pick a 7-pin BEST WK before taking on a TrioVing DP10. This is not meant to be a dissertation or mathematical proof on lockpicking. This is a list created with enough input from enough pickers to have significant value. Up through level 5 we'll have a great sample size and pretty much cement the difficulty rating for locks through that level. As the levels increase the sample size gets smaller but since only experienced pickers votes are counted in this range a general roadmap is created for upcoming pickers.

I believe such a list is worthwhile, does have value, and I would like to make it happen. However, I need help to get this going. Just a little bit here and there over time. Creating something does take a little work but is always worth the effort after it is completed.

Take it easy!

[Oldfast]

Great stuff! And QUITE an undertaking. Even with a collective effort, it'll happen oooover time.

I dunno if I'll be all too much help unfortunately. Definitely not with the computer portion of it. lol
In regards to picking- though I've been at it for some time now, when it comes to high security,
I'm still at a fairly low level really. Maybe I can help with pictures or something along the way.

When I first started I would've loved a list to work my way up!
Appreciate the effort and enthusiasm. Thanks Flywheel
Here's to a worth while project

[MrAnybody]

Subjective and objective measurements are not what they appear to be prima facie. It is somewhat a modern myth that whatever objective is both outside human experience and better than subjective measurements. All that objectivity brings is a standard to measure against and subjectivity does not. How that external objective measure is created becomes troublesome.

Scientific measurements, when not attempting to derive a categorical fact, are not at all objective. How do you define 'difficulty'? What makes you think that any specific feature always increases difficulty for every person? Does a combination of metrics that individually increase difficulty, when combined, actually decrease it overall? Some features that clearly increase difficulty in one model clearly decreases it in another. Further, for some people different features or metrics that appear more secure actually make locks easier for them. How do you know that what metric or weighting you chose is correct? What methodology will it take to create such things? It becomes a 'turtles all the way down' scenario when trying to find an objective way to create an objective measure. You must then account not only for the metrics to be used and how those will be measured, but how those metrics will be chosen, why they will be chosen, how to prove that those measures really are objective (they aren't), and how to prove that they are a viable measurement in the first place. At some point you will hit a limitation of knowledge. At some point some feature is more difficult because it feels more difficult and for no other reason. There is nothing scientifically empirical about that, or is there?

There is no such dichotomy between subjective measurements vs empirical means. A person that really likes cake can test each type of cake available and decide on what sorts of cake he likes best. Then the conclusions found from that subjective testing can be used for future predictions over what new types of cake he would most likely enjoy. I know I do not like things with mushrooms in them. It is entirely subjective on my part. I can use that single data point, 'I do not like mushrooms', to predict and test new things. If something has mushrooms in it, I probably wont like it. Unless they are made of brass.

Even if the limitations of knowledge, subjectivity in methodology, and acceptance of empirical measures based on personal taste, there is the problem of rumination or iteration. So you have a good methodology of choice down, and you have a methodology of testing, a way to store the results, and a way to utilize them in some practical fashion. That still leaves a lack of bounds problem. When do you stop adding metrics? When is the weighting good enough? Where is it accurate enough? How do you measure that accuracy? With metric standards there is always a forced choice in limitation of bounds. There will always be some point where you must say 'that is accurate enough'. A database with 10,000 options on how a lock can be more or less secure isn't at all useful, but it does have the possibility of being more accurate than a metric of one option. Secondly, with changes in an environment things that once were difficult may no longer be. Twenty years ago Medeco would have been ranked higher than other brands that are considered its superior today. Suddenly all those metrics and methodologies have to be thrown out or reworked on an ongoing basis to maintain even remotely the same accuracy. Instead of increasing in accuracy over time (as ranking systems do) metric systems diverge from the mean in unusual, unpredictable ways. This is one of the more significant issues with technical stock analysis. The validity of metrics are moving nearly as fast as the data.

The alternative to this is the basis of the thread already. An evolutionary ranking system. It bypasses the difficulties of what measure should be chosen, why is it chosen, how is it chosen, and the limitation of knowledge problem. It maintains the ability to predict outcomes based on past data. Being able to know when the job is done is the prime factor in completion. With metrics, there is no done. There is no cessation of work. You could continue on from here to the end of time with the project without ever completing it. Ranking systems even get more accurate over time instead of less. Of course a 'hot or not' lock page certainly would be way more fun to use than downloading a relational database, but that is just my subjectivity showing.

I think my brain just melted

[rerun12]

I think someone may have read "Zen and the Art of Motorcycle Maintenance" one too many times and is transposing quality into difficulty.

was thinking the same thing heh, was disappointed phaedrus had not yet joined the forum

[Deadlock]

Whereabouts would something like a Squire "Old Fashioned" lever padlock go?

Correcting mistake here. I meant the Squire "Old English" type of padlock.