Dr Mann - I copied and pasted your 3:53 post into a .odt document and emailed them to both Sasho MacKenzie and Dave Tutelman.
Here is the response from Sasho MacKenzie
"I think everything in the document is pretty accurate. I don't know what type of grip style the participants in Koike's studies used. Presumably a split hand baseball style grip so as to maximize internal validity (obviously at the expense of external validity). I don't think there would be any 'error' in the measurements if a grip with some overlap was used; you just couldn't attribute the top and bottom grip measurement to isolated hands.
I don't think the fundamental force/torques patterns would change much if the same golfer used a vardon vs a split grip...likely just some changes in peaks and timings...could be wrong though."
------------------------------------------------------
Here is Dave Tutelmans reply:
"I got a chance to see Sasho's answer. We are in agreement. I just use more words to explain it.
Let me start by responding to something that was in a more detailed email you sent me the other day asking essentially the same question. When I wrote my introductory article on modeling the swing in 2012 (http://tutelman.com/golf/swing/models.php), I explicitly excluded Homer Kelley's TGM. Here's a quote from the very first page of that article. "It is certainly a detailed discussion of the bones, muscles, joints, and club during the golf swing. But it falls short of a model in that it is all qualitative. There is nothing there that tells you how to analyze it mathematically, nor has that analysis ever been attempted (to my knowledge, anyway)."
That is my view of Dr Mann's work as well: very interesting video-supported discussions of fine aspects of the golf swing, but completely qualitative. There is no attempt to put measurement on the forces and torques, so it is not a kinetic model. Sometimes he is working a problem that biomechanics has modeled, but often not so.
On to the question at hand. By pure serendipity, I ran across a 100% relevant quote just this morning:
"All models are wrong, but some are useful."
In a discussion of that quote (usually attributed to the statistician George Box), another statistician (David Cox) writes, "The very word model implies simplification and idealization... The construction of idealized representations that capture important stable aspects of such systems is ... a vital part of general scientific analysis and statistical models"
I'll gladly go with that. The real trick is not to abstract away those features important to the question you are trying to answer. The question I was trying to answer was, "At impact, is the hand couple positive or negative or dependent on the golfer?" There were three sources I looked at (so far) to get that information:
(1) The work of established swing modelers. Sasho MacKenzie and Young-Hoo Kwon say it is always negative. Steven Nesbit says some golfers are positive and some are negative.
(2) Shaft bend should be a reliable indicator of hand couple. Shaft bend says the couple is always negative at impact.
(3) Direct measurement of hand forces by instrumenting the grip. Koike's work says it is always negative.
Your email is about #3, and it questions the validity of the graphs if the golfer uses a Vardon grip. The issue you raise boils down to how to ALLOCATE THE FORCE TO A HAND where the overlapping right hand thumb pad and small fingers (I don't think this was mentioned, but should have been) are pressing on the left hand, which in turn is pressing on the grip.
If the question I want to answer is really about allocating the force, then it is a difficult one, and one that Koike did not address. Dr Mann is absolutely right in that regard. But if the question is evaluating the hand couple -- which is my question, whether or not it was Koike's -- then it is not a problem. The simplified, idealized model takes care of it. I don't care which hand's effort was actually RESPONSIBLE for the force, just where it was applied on the handle and how it contributed to the couple. The sensors give me that, regardless of whether we resolve which hand's muscles were actually responsible for it.
The Koike paper allocates the forces based on simple assumptions about which hand is responsible for which sensor's readings. Those assumptions may be right, wrong, or ambiguous. But there is nothing ambiguous in how they contribute to the hand couple. And that is all I am interested in. The model works for me. It may or may not work (likely not) for an instructor telling the student which arm to push in which direction.
Hope this answers your (Dr Mann's) question."
--------------------------------------------------------------------------------
DG