Spec Ops Leaders Want Return to Fundamentals

[Ken White]

very broad indicator of trends, good units neither count nor note the number killed. Individuals may, depending mostly on how close they were or whether they had to pick up the body or parts or not but units don't count 'em because it generally has little relevance to whether the mission is getting accomplished or not. That's pretty well true at any level of conflict.

Unless, of course, someone in the chain is over enamored of 'metrics'...

[Bill Moore]

They are not "either/or" or mutually exclusive - they can be mutually supporting and reinforcing if applied correctly.

Max161

I agree with Max on this comment that direct and indirect, DA and CMO, and any other ying/yang combination we can come up with can be mutually supporting, but what we haven't mastered in many cases is the "if applied correctly". There should be one battlespace owner that dictates the supported/supporting relationships between units and their missions to avoid a situation where unit A conducts an operation setting back unit B's operations.

For example, if unit B (any conventional combat arms unit) is focused on providing local security and building rapport with the locals using civil-military operations, etc., then all the sudden Unit A come into their battlespace unannounced to whack to some bad guys they were following creating undesired secondary effects in unit's B's AO that set them back several weeks. This is an example of an unsupported and unsupporting command relationship. Crap happens on occassion, but many times this crap could be avoided if we used our doctrine, and all operations were required to be approved by the battlespace commander. Special Forces and Conventional Combat Arms units (after years of COIN experience and training) are quite capable of doing both the Ying and Yang approach simultaneously, while other units are strictly Ying, no Yang. As a community we need to figure out how to use their Ying more productively, because Ying is good when used right.

As for SOCOM, I think ADM Olsen's comments speak for themselves (see bourbon's post), and I also think SOCOM is now on the right track. SOCOM needs a world class direct action capability, and they have it. They also need a world class indirect capability, and they have that also relative to other nations, but we can get better in this area with more command emphasis and we will.

and it ain't looking good for the home team barring some reform in Congress...

Ken

As for the interagency process and Congressional reforms, I won't hold my breath. We need to figure out how to win with the real world political environment we live in until enough good people run and get elected to make substantial change in the system.

[Ken White]

As for the interagency process and Congressional reforms, I won't hold my breath. We need to figure out how to win with the real world political environment we live in until enough good people run and get elected to make substantial change in the system.

and isn't going to get fixed any time soon. We have have been dealing reasonably well with that disconnect for years and will have to today and for a while.

Simple old unity of command is never as easy as it should be but it is important. Right now it's too personality dependent and one fool in the wrong place can mess things up for a bunch of folks. I agree with you that while we're better than we used to be in many respects, we could still get better -- and we all need to realize we have to work it out ourselves.

Er, uh, make that yourselves. I'm retarded...

[Surferbeetle]

Unless, of course, someone in the chain is over enamored of 'metrics'...

For fun (after work), I just spent the last couple of months putting together a simple excel spreadsheet, tied in with some google earth and census data, which used monte carlo analysis and some simple home-made algorithms to quantify risk to capital, analyze a population, predict actions, and compare and analyze various 'accepted' business ratios and metrics (capital budgeting) for a specific industry. I am thinking about how (and gathering my strength) to apply bayes theorem, generate some fancy genetic algorithms and use the black-scholes model to examine some stock picks (or I might get a monkey and a dart board). I am also thinking about how to use some of these techniques to quantify a work problem involving water and earth pressures more accurately. Spreadsheet Modeling & Decision Analysis (ISBN 978-0-324-31256-0) is an interesting 'how to' book.

The point of this metrics ramble is that their are many folks way ahead of me who are using simulations/metrics to predict things. I am just 'baby-stepping along' trying to figure things out and during my journey I often think about one of my heroes, Robert Manning. Long story short this guy with no formal engineering education started out as a draftsman, worked his way up to a senior public works engineer position, and was able to survey, select, combine and simplify several hundred years worth of hydraulics work into some workable equations (he was in his 70's when he presented these equations) that we still use today to predict hydraulic phenomena which were previously unquantifiable/unpredictable at an acceptable level (he died in 1897). Sometimes I just walk down the hall at work and talk to an engineer in his 70's who is doing some amazing things with some hydraulic computer models and get inspired that way.

I guess one could say that despite being aware of the many variables I have hope in the metrics department when it comes to predicting outcomes in warfare. Fortunately I also know that I have a metrics bias and the folks I work with help me keep it in check

[Bill Moore]

I just spent the last couple of months putting together a simple excel spreadsheet, tied in with some google earth and census data, which used monte carlo analysis and some simple home-made algorithms to quantify risk to capital, analyze a population, predict actions, and compare and analyze various 'accepted' business ratios and metrics (capital budgeting) for a specific industry. I am thinking about how (and gathering my strength) to apply bayes theorem, generate some fancy genetic algorithms and use the black-scholes model to examine some stock picks (or I might get a monkey and a dart board).

Surferbettle, it is 0742hrs or 7:42 A.M and after reading your last post I think I'm going to have a beer.

Tomorrow I'll go to the library and order, "Spreadsheet Modeling & Decision Analysis (ISBN 978-0-324-31256-0)" and leave it laying around the office to see if I can impress anyone. Probably not, they know I'm the monkey and dart board type, so that book is obviously way above my level of technical education (lol).

By all means please continue to pursue

despite being aware of the many variables I have hope in the metrics department when it comes to predicting outcomes in warfare.

I personally am not convinced yet that metrics will lead us to cause and effect, rather they'll show some gross level correlations at best, but I remain open to the idea.

[Ken White]

Metrics are imperative for many things, most hard science and most fields of engineering being but two. I suspect sales and marketing are another. The insurance business certainly requires 'em.

OTOH, been my observation that in warfare -- as opposed to the logistical or a very few support functions thereof -- metrics generally serve little purpose other than to give a false picture and that such an outcome is particularly probable if the improper metrics are chosen for illumination...

Though heaven knows these guys (LINK) agree with you on the potential...

[selil]

Metrics (like statistics) is the crutch that people who can not think use to excuse bad decision making in dynamic environments where cognition would normally be rewarded.

Put another way:

Numbers absolve failure where common sense has been ignored.

[Surferbeetle]

Bill,

Don't forget the sausage and pretzel for first breakfast to help smooth things out! When I go for just the beer, first thing, I often tend to fall down.

Bill, Ken, and Sam,

I can't speak to George Soros' politics however I am impressed with his results in quantitative modeling in finance. The downside in quantitative financial modeling is pretty scary. I can't begin to imagine the pain associated with losing the amounts of cash that some of the quant funds have. It appears that they trusted their models too much; a point which you guys consistently point out.

I have suffered through enough heat, cold, bugs, nasty cases of the runs, and having people try to physically ruin my day to forget that models are just a representation of reality. The power of the models I use in engineering and business however have me hooked on the possibilities of adapting them for war.

Now where is that beer...

Cheers,

Steve

[Old Eagle]

I'm with Max on this. ADM Olsen is making all the right noises about FID and SFA. With upcoming SFA proponency, he'll have ample opportunity to make his mark.

In fact, he may be able to teach some of his Army bretheren something -- like leading a re-juvenation of FID in the Army!

I am somewhat embarrassed that it looks like it's going to take a SEAL to lead the way where the Army once held sway. (Dude! That rhymes!!)

[120mm]

Bill,
Don't forget the sausage and pretzel for first breakfast to help smooth things out! When I go for just the beer, first thing, I often tend to fall down.

It's been almost a month since I had that last....

Living in Bayern just rocks, sometime!!!

[Schmedlap]

I am thinking about how (and gathering my strength) to apply bayes theorem, generate some fancy genetic algorithms and use the black-scholes model to examine some stock picks (or I might get a monkey and a dart board).

Can you actually use Excel to do calculus? I'm brushing up on my calculus in order to try my hand with Black-Scholes (I'm taking a graduate course in options trading). In the past, I simply bought put options when I expected a stock to tank and call options when I thought it was grossly overvalued and volatile. But, I'll see how much money I can lose with the assistance of mathematical models.

Long story short this guy with no formal engineering education started out as a draftsman, worked his way up to a senior public works engineer position, and was able to survey, select, combine and simplify several hundred years worth of hydraulics work into some workable equations (he was in his 70's when he presented these equations) that we still use today to predict hydraulic phenomena which were previously unquantifiable/unpredictable at an acceptable level (he died in 1897).

That deals with phenomena that can be recreated and tested to confirm its validity beyond a shadow of a doubt. That, to me, implies greater simplicity of the subject matter and greater certainty along the way as to whether you're on the right track. I don't see any parallel to social science. But I'm also not attempting to discourage your work - it would be neat if you came up with something useful.

[Rank amateur]

I'll see how much money I can lose with the assistance of mathematical models.

A lot and I don't need a spreadsheet to figure it out. I work on the theory that the more money the stock broker makes on the transaction, the worse the investment is for me.