Here's what you need to know...
Rigid body analysis - where the shape and dimensions of a body are considered to be constant and undeformable regardless of the forces applied to it - is useful in the analysis of complex human movement, like the deadlift.
The Standard Pulling Position is where the bar travels up the legs in a vertical line over the middle of the foot. The shoulders stay just forward of the bar until the top of the pull.
The most efficient way to pull a barbell is straight up. Always.
A correct deadlift will start with the bar about one inch from the shin, placed with the shin in a perfectly vertical position.
The heavier the weight the more efficient the pull must be. Warm-ups can be done wrong, as can rows, cleans, snatches, and anything else with a light weight.
Looking up when deadlifting is a common mistake.
The alternating grip allows you to pull more, but it can cause injuries.
Warning: Dry technical article. Very little cursing, no bashing, no actual controversy. Just logic and analysis. Caffeine will be helpful.
Rigid Body Analysis
The human skeleton is the system of levers that we use to interact with our physical environment.
It's operated by a system of motors - little tension engines called muscles that operate the levers of the skeleton by generating a contractile force.
The details of the operation of this physical system can therefore be understood through analysis, and these details can be as complex or as general as the level of the analysis.
With more than 600 muscles and 200 bones, things can get very weird. Or they can be rendered understandable by pulling back a little and viewing the system as a chain of segments that transmit the force generated by the groups of muscles that operate them.
Running, jumping, throwing, swimming, and hitting a ball are fantastically complex activities if all of the constituent muscles and bones must be enumerated, each of their contributions to the functioning kinetic chain considered, and their individual force-production/transmission roles calculated.
It's not only impossible with our current instrumentation, it's rather pointless, because enumerating the individual contributions of the forearm muscles and bones of the wrist to a racquetball serve don't help us serve the ball, or learn how to do it better.
Movements become analyzable if you consider them at the macro level - the functional segments of the body operating against the resistance of the implement or the ground, and the primary muscles responsible for their operation.
The concept of rigid body analysis - where the shape and dimensions of a body are considered to be constant and undeformable regardless of the forces applied to it - is useful in the analysis of complex human movement.
A stride in a sprint is a hip/knee extension with a contralateral hip/knee flexion. The kinetic relationships of the feet, shank, thigh, trunk, and arms are complex enough without having to enumerate the roles of their constituent components.
In coaching the sprint we're primarily concerned with the force production and transmission relationships of the segments themselves, and we're less concerned with their internal components.
It's important to understand what the calf muscles, quads, hamstrings, glutes and adductors, spinal erectors, and shoulder girdle muscles are doing, because this understanding informs us about force production within the functional segments.
It isn't as important to detail the mechanical function of the obturator, the peroneus longus, or the long head of the triceps brachii in the sprinting motion, or their respective percentages of force contribution at any given point of their participation in the movement.
Their functions become important to consider in the event of an injury, and an informed coach knows enough anatomy to understand these functions, but we don't single them out for training because they're worked along with the primary muscles as they fulfill their anatomically-determined role in the correctly-performed gross motor pattern.
More important are the feet, shank, thigh, trunk and arms during the development of the acceleration as revealed by the analysis of their relative angles and their relationships to the ground.
The individual muscles that make this happen are not of concern during the execution of a record performance, unless one of them fails due to injury.
Strength training has long been the victim of a lack of focus on the movement patterns of the segments of body itself, in lieu of the great deal of attention being paid to the constituent components - the muscle groups of bodybuilding-think.
Let's examine your favorite and mine, the deadlift, from the perspective of rigid-body analysis, and see if we can't come to a better understanding of what actually happens when a bar is pulled from the floor.
The Mechanics of the Deadlift
An examination of men's division deadlift records on video will show you what you need to know about the relationship between the loaded barbell and the lifter: the barbell travels up the legs in an approximately vertical line over the middle of the foot, while the shoulders stay in a position just forward of the bar until the top of the pull.
And this is true no matter what the lifter attempts to with the bar; you may start somewhere else, but you're going to end up doing the pull this way.
We call this position The Standard Pulling Position, because, as we'll see, it results in the most efficient pull for deadlifts, cleans, and even snatches.
(The reason why women's deadlifts don't always obey this rule has to do with the same reason women can perform a much higher percentage of their 1RM for reps, but that's a subject for another article.)
It's important to know why this amazing lack of variation exists in pulling a heavy barbell from the floor, because if we know what's going to happen when we pull a heavy deadlift, we can make better plans for making it happen with the greatest level of reproducible efficiency.
If you know that the barbell is coming up in a straight line over the mid-foot, you can practice placing it in that position and keeping it there.
And if you know that the shoulders are going to be a little forward of the barbell when it leaves the floor, you can make plans to have them in this position when the pull starts.
Knowing why is good, because it convinces you that correct technique is important.
It's not nearly as important to know what role the individual forearm muscles play in the effort, or the obturator, the peroneus longus, or the long head of the triceps. They're working, and good deadlift technique is designed to make them contribute to the pull.
But more important is the overall relationships of the major segments of the kinetic chain to the load on the bar.
Gravity is the force that makes mass weigh what it does. It's the attraction between two masses.
In the case of the earth, it's really really big relative to the barbell, so the attraction between the two masses is heavily skewed.
Since the attraction is between the two Centers of Mass, the center of the barbell and the center of the earth have a very predictable relationship: gravity operates on the bar, and it does so as a downward pull, in the precise direction of the center of the earth's mass. Always.
We call this direction vertical.
As a result, work done against gravity is always performed vertically, upward. Always. All work done against the gravity vector will be done parallel to that vector in the opposite direction. Therefore, the most efficient way to pull a barbell is straight up. Always.
Why the Mid-foot?
A correct deadlift will start with the bar about one inch (2.5cm) from your shin, placed with the shin in a perfectly vertical position. This will place the bar directly over the mid-foot, for everybody.
Human feet are quite thoroughly proportional, front to back, and we've seen people with size 3 to size 17 shoes position the bar over the mid-foot with it placed an inch forward of the shins.
Once the bar is over the mid-foot, turn your toes out a little, maybe 15 degrees.
When this position is taken and the knees are dropped forward and out a little, so that the shins touch the bar without moving it and the knees stay parallel to the toes, the shin angle will be perhaps 7 or 8 degrees forward of vertical.
The knees-out position allows you to use the hip external rotators more effectively when you pull, and it also makes more room between your thighs so you can set your lumbar extension more effectively.
Knees out also involves the adductors in the pull as well, by stretching them into a position where they can more effectively participate as hip extensors.
And the slight shin angle allows the quads to perform enough knee extension to help with the start of the pull. Knees any more forward than this will cause the shins to push the bar too far forward to pull a deadlift. Here's why.
The foot is your point of contact with the ground. If the bottom of your foot was an arc, as though you were standing on a 12-inch piece of pipe that was split down the middle into halves, the mass of the system would be in balance directly over the center of the arc - your mid-foot.
If you were in good balance, you could squat down and stand back up by keeping your center of mass directly over the middle of the arc. Even though your feet are essentially flat against the ground, the balance relationship is clearly the same.
The surface area of the feet is the area across which the load of the combined mass of the barbell and your body is distributed against the floor.
A man with a size 11 weightlifting shoe will have a surface area of about 90 square inches (580 cm²), or an average of about 5.5 pounds per square inch with a 500-pound deadlift (3.84N/cm²).
At 800 pounds, the load is closer to 8.9 PSI. If the barbell is placed in a position halfway between the toe and the heel, the average load across the contact surface will be fairly evenly distributed - balanced - between forward and aft sections of the foot.
If the most even pressure distribution occurs with the load at mid-foot, positions either forward or behind this point will display an unevenly loaded surface area.
Forward of the mid-foot, the load distribution will skew forward, reducing the surface area of force distribution - likewise if the bar is behind the balance position.
The most important effect of this uneven loading is that it sets up uneven moment forces between the balance point and the lifter/barbell Center of Mass measured vertically from the floor.
Moment force, or leverage, is the turning force your hand transmits to a bolt through a wrench. If we measured the force of the spin along the shaft of the bolt, that quantity would be torque, but we aren't concerned with torque since nothing in a human joint spins under normal circumstances.
We're merely concerned with the force transmitted along the shank, thigh, and trunk segments between the barbell and the floor - the force that moves the load.
A wrench multiplies the force you can generate with your hand by trading a longer arc of motion at a higher velocity for a much shorter arc and a higher amount of force. A crowbar works the same way, distance and velocity exchanged for force - mechanical advantage.
The moment arm is the distance between the point of force application (your hand on the crowbar) and the point of rotation (the bend in the short end of the crowbar), measured at 90 degrees to the force application, and it's the way the multiplication of force is calculated.
The longer the moment arm, the greater the leverage at the point of rotation.
A crowbar has a long segment, a bend, and a short segment, and is meant to be used to generate high force when prying a nail out of a board. Here's an example of the same long/short segment configuration used the opposite way:
The trebuchet is a siege engine that uses a very high amount of force on the short segment to increase the velocity of the light load on the end of the long segment.
In this system, the multiplication takes place in the opposite way: like stepping on a rake, high force operating the short segment rotating around a fulcrum produces a high velocity at the end of the long segment, if the force is high enough on the short end.
The angular velocity on both sides of the fulcrum is the same, and the long moment arm multiplies the velocity along the arc at the end of the long segment.
Now, think of the horizontal distance between the mid-foot and the 600-pound bar as a lever against your feet, with the vertical distance up the legs as the segment. If the bar is above the knees and you shift forward 2 inches, the distribution of the weight against the floor shifts forward.
Along the vertical distance between the floor and the barbell, leverage is being applied to the system that wouldn't be there if the bar was directly vertical to the mid-foot.
The horizontal distance between your hips and the bar constitute a moment arm, one that you use to generate enough force to pull the bar up. The additional 2 inches of moment arm your technical mistake has created gives the load on the bar some additional leverage against you.
The total amount of moment force that must be generated is optimal when the bar is over mid-foot, but now the force of the additional 2 inches must be added to the amount of force you have to produce to keep the 600 pounds moving up.
Moving the bar back behind the mid-foot to shorten this moment arm doesn't work, because you're off-balance backwards against the floor, and because moving the hips back lowers them in relation to the knees.
Since the bar moved forward, the countering force that must be added to the pull comes from behind.
At the bottom of the pull, the posterior tug comes from your calf muscles - the gastrocs and soleus. You can feel this effect for yourself if you stand up straight and let your bodyweight drift onto your toes.
It's easy enough to counter this off-balance shift without a 600-pound bar to deal with, but adding enough backward pull from a bar position above the knee can't really be done, since calves aren't designed or positioned to generate force against a barbell above the knee.
Those of us who have ruptured an Achilles tendon know how this works with a load on our backs or on the floor in front of us - the effect is subtle, until you can't produce it effectively, and then you miss a rep or fall forward.
If you start with the bar forward of the balance point, it will roll back to the mid-foot before it leaves the ground, thus correcting the problem while the vertical distance between the bar and the foot is manageably short.
If the weight is light enough, as with a clean, the bar can be pulled off the ground in this forward position, but it will be pulled back toward the balance point as it rises, thus creating a curve in the bar path off the floor.
If it's light enough to pull very inefficiently, as with a snatch, a strong lifter can throw it all over the room and still catch it overhead - not efficient, but quite commonly performed.
If you get forward higher in the pull, the way you might try to save the lift is by making your back angle more vertical earlier in the pull than you would if the bar was in balance - you look up, and you lean back, or at least try to lean back.
This doesn't work very well, as some of the recreational federations have admitted by placing a spotter behind their less-than-technically-competent deadlifters at their meets.
Why the Shoulders Forward of the Bar?
This is a very interesting phenomenon, because it doesn't seem logical that the arms would not just hang straight down when loaded, like a plumb line mirroring the gravity vector.
The point of rotation is the shoulder, the barbell is in the hand, the arm connects the two points, and the damn thing ought to behave properly.
But it doesn't.
Every heavy deadlift hangs from the shoulders with the arms at a slight angle, perhaps 7 to maybe 11 degrees, with the shoulders just forward of the bar, and with the bar over the middle of the foot.
The back angle adjusts to place the shoulders in this position; if the hips are too low, with the shoulders back behind the bar, the back angle will adjust horizontally to place the shoulders in the Standard Pulling Position.
Furthermore, the following generalization is born out by repeated observation: the heavier the pull, the more likely it is to conform to this alignment, and the lighter it is the greater the amount of deviation it's likely to display.
Watch Brad Gillingham's 881/400kg deadlift. Brad comes to the bar and carefully places his mid-foot directly under the bar - shins about an inch from the bar - takes his grip without moving the bar, drops his knees forward until shins touch the bar, squeezes his chest up, and pulls the bar in a vertical line to lockout.
His back angle at the very start of the pull doesn't change to more horizontal, and in fact immediately becomes more vertical as the bar is pulled up.
Notice the shoulders just forward of the bar (ignore the black shirt, look at his shoulders and hands), and notice that the arms don't become vertical until just before lockout.
Perfect pulling efficiency is displayed with the bar over the mid-foot and the shoulders just in front of the bar, the heavier the weight the more efficient the pull must be. You can watch videos for hours on end that show this pattern.
Warm-ups can be done wrong, as can rows, cleans, snatches, and anything else with a light weight. But when the last bits of your limit capacity are approached, this alignment must be displayed or you miss the pull.
When the bar is forward of the mid-foot during the reps of this set, it drifts back during the pull, but notice the position of the shoulders over the bar, evidenced by the arm angle.
Mike is strong enough to pull this weight with a little horizontal slop in the floor pull, and he does this with heavy attempts as well. But the bar always drifts back to the mid-foot.
More importantly, why are his shoulders in this persistent forward position? This is best explained with an illustration.
Moment force - leverage, like your hand generates when it pulls on a wrench - is most efficiently applied at 90 degrees to the tool being pulled on, be it a wrench, a crowbar, or a hammer.
The latissimus dorsi muscles attach the lower back to the upper arm; specifically, the lats originate on the back from T8 all the way down to the sacrum and across the top of the pelvis, and have an insertion point on the medial anterior proximal humerus, essentially in your armpit.
The shoulders just in front of the barbell place the lats in the best position to apply moment force to the humerus - to best pull the bar back to keep it over the mid-foot balance point.
The lats at 90 degrees is the alignment generated by the shoulders-in-front-of-the-bar position with the arms hanging at an angle, as the figure shows.
You can see the effect of this position quite clearly: control of the tendency of the bar to drift forward is best accomplished with the backward tug provided by the lats.
And the back angle - the angle between the plane of the trunk and the horizontal floor - is what controls the position of the shoulders relative to the bar, and thus the angle of attack between the lats and the humerus.
The actual back angle that occurs in a properly positioned pull is dependent on anthropometry.
A short back and long legs will obviously generate a more horizontal back angle in the proper position than a long back and short legs. In either case, the proper position for the lats against the humerus will result from a shoulder position just in front of the bar that hangs the arms at 7-11 degrees, any variation being due to anthropometry.
But the lats generate tension between both origin and insertion, so they're pulling on the lower back as well.
Part of the back angle settling into position with shoulders forward of the bar is probably a function of the hips being pulled up into that position by the lat's anterior tension on the low back, as the muscle belly settles into its optimum force production length in isometric contraction.
This effect also rotates the mass of your body up into a position more above the bar. As your own mass moves into a position so that more of it is forward of the bar, the center of mass of your body lines up better with the center of mass of the barbell.
Pulling it upward is then easier, since you obviously can't pull upward on something in front of you. A deadlift will always be in front of your legs, and that's why you lock it out with a slight backward lean - to balance your mass back against the bar in front.
But during the pull, some of your mass - including your head, if your neck is positioned correctly - remains forward of the bar, and the arms don't become completely vertical until you finish the pull.
As for the minutia of the system, the teres major and the long head of the triceps also cross the shoulder joint between the scapula and the humerus, and therefore they're also involved in this moment force relationship.
But the triceps long head is at a very poor angle for generating moment force across this gap, and the teres is a short small muscle; neither are terribly important to the consideration of the larger mechanism at work.
The lats hold the bar back over the mid-foot, and they hold the back angle in a position to keep your mass more over the bar.
Two other back angle configurations are possible.
This is the most common start position in competitive powerlifting, and it's inefficient for reasons that are obvious in light of this analysis.
The most compelling argument against it is the fact that everyone who starts a heavy deadlift here exhibits a shift in back angle towards the Standard Pulling Position.
You can try to pull the bar with vertical arms and hips down too low - squat with the bar in your hands - but it just doesn't work that way.
The fact that the extraneous movement occurs before the bar leaves the floor, and that it can be pulled without this extraneous movement demonstrates that a higher-hips position is a more efficient place to start a deadlift.
Note that the greatest lifter in the history of the sport of powerlifting demonstrates this shift in back angle.
The pull starts when the bar actually leaves the floor, and during the time between when his upward motion started and the bar leaving the floor his back angle changed.
In fact, there's a slight pause in the motion as the system comes into stability, just before the bar leaves the floor, as the lats grab the bar. This is a very common start technique, and there are countless examples of it, to the extent that it's considered by some to be the best way to pull.
If the hips are low, the lats are not optimally engaged with the arms, and the lower back isn't optimally anchored to the bar - the lat generates tension against both the origin and the insertion.
As the lat becomes optimally aligned to hold the bar back over the mid-foot, it also pulls the hips up into their characteristic position, and the resulting back angle is stable just as the bar leaves the floor.
Again, the actual angle depends on anthropometry, but the alignment itself produces the back angle.
In this position, the lat's angle of attack has opened beyond 90 degrees. This is recognizable as the classic stiff-legged deadlift position, with knees extended and back angle more horizontal.
Those of us (who even lift) know that you can't pull as much in this position as you can with an optimal back angle, and that if you start here you'll almost certainly miss the attempt unless you can lower the hips and pull the bar back in.
The ability to stabilize the bar over the balance point is critical to the ability to apply enough force to pull the deadlift. But there is another reason...
What About The Hamstrings?
The hamstrings are really a poorly understood component of the squat and the deadlift, the subject of much unnecessary stretching and mobility wodding.
Remember the basic rule of hamstring kinematics in the pull and the squat, (and most everything else too): As the hips flex, the knees flex.
If the angle of hip flexion remains roughly equal to the angle of knee flexion, the hamstrings haven't changed muscle belly length very much. Thus, hamstring flexibility isn't the determining factor in squat depth.
If the hamstrings don't change length - this is very hard to measure, and is why it hasn't been done quantitatively - hamstrings function primarily as isometric stabilizers of the back angle in both the squat and the pull.
In both movements, their primary contribution to the kinetic chain is support of the back angle. They anchor the ischial tuberosity to the medial and lateral attachments on the tibia, at the knee.
This pelvis-to-knee connection functions as a bridge between the extending knees and the extending hips, enabling the force generated by the quads, glutes, and adductors to move the load without the back angle collapsing horizontal.
Along with the spinal erectors and the lats, the hamstrings are the isometric glue that holds the pull together.
This, coupled with the fact that a muscle generates its greatest contractile force isometrically, at its resting length, means that whatever position generates the best semblance of resting muscle belly length will be the strongest position to use the hamstrings isometrically.
Our analysis shows that in a squat, the low-bar position holds the hamstrings at resting length (or possibly slightly stretched, depending on the lifter's anthropometry) throughout the movement.
In a high-bar squat or front squat, the knee angle is more closed and the hip angle is more open, shortening the hamstrings from their resting length and therefore diminishing their isometric capacity to anchor the back angle.
In a deadlift, the correct start position finds the hip angle more closed and the knee angle more open than in the low-bar squat bottom position, with the hamstrings stretched slightly beyond their resting length, allowing them to contribute primarily isometrically as they hold the back angle in place, as well as contributing to hip extension at the top of the pull as they return to their resting length.
The position which generates the most mechanically-efficient angle of attack on the tibia and the pelvis will be the position that can generate the greatest back angle stability during the first part of the pull.
Look at these three positions and decide for yourself which works best for hamstring function.
What About The Back Angle?
As the bar travels up the shins in a deadlift, the back angle becomes more vertical. This begins immediately after the bar leaves the floor.
In a clean or a snatch - sub-maximal accelerated pulls by definition - the back angle should (and usually does) stay more constant and more horizontal until the bar gets much higher. Staying out over the bar is a familiar concept to some Olympic lifters, their having heard the cue many times.
What does it mean, and why?
The concept of the moment arm is quite important throughout barbell training, and nowhere more important than when considering the role of the back in the clean and snatch versus the deadlift.
The primary difference between the Olympic lifts and the deadlift is acceleration - the first derivative of velocity, the rate at which velocity increases.
A clean is a pull that's accelerated enough to catch on the shoulders.
In order to impart sufficient momentum to the loaded barbell that it continues on upward between the end of the pull and the catch on the shoulders - that portion of the bar path during which no force is transmitted to the bar by reacting against the ground - it must be sufficiently accelerated so that it floats long enough to shift the feet and the arms into the catch position from the pulling position.
The snatch is caught overhead, and is thus a longer pull with an obviously lighter weight.
As such, there are no slow cleans or snatches. They must be accelerated or they don't rack, and this is why they're better measures of power than dynamic effort deadlifts. How precisely do you miss a dynamic-effort deadlift?
In stark contrast to a clean, a deadlift can be slow. Some world records have taken 8 seconds to pull, as opposed to less than a second for a clean or snatch.
This critical distinction is the reason for the difference in the behavior of the back angle between the fast and slow pulls: a deadlift shows a back angle that becomes more vertical as it leaves the floor, while a clean and especially a snatch preserve the horizontality of the back angle as long as possible.
Dimas was the undisputed master of staying out over the bar, preserving his back angle until the bar was above his knees.
This is important because of acceleration, and the way it's produced in a fast pull. A clean is a pull in which the back angle/hip moment arm can be maintained long enough to accelerate the barbell - to increase the linear velocity of the barbell by using the mechanical advantage of the longer moment arm.
The back segment between the hip and the barbell is the tool the lifter uses to accelerate the bar.
Like the trebuchet, the clean uses the short moment arm loaded to a high force level and operating over a short arc around the fulcrum to accelerate a much longer moment arm with a lighter load on the end - if you're strong enough.
If the weight is too heavy, this long moment arm can't be maintained; the back angle becomes more vertical immediately because the longer moment arm can't be operated by the posterior chain musculature, the weight therefore can't be accelerated, and it's a deadlift instead of a clean.
And this is why a deadlift can be done with some spinal flexion, and a clean is much more dependent on lumbar and thoracic extension.
Many heavy deadlifts have been pulled with a round upper back, but Olympic lifters who fail to keep their backs tight and flat will have problems reproducing their technique accurately - a rounding back is a power leak, and the deadlift doesn't rely on power and acceleration like the Olympic lifts do.
The Other Details
Sometimes, but not very often, the fine anatomical details of origin/insertion/action of the individual muscles are critical, so forget all that shit I said about this earlier.
Here are a couple of examples where the details about things that can't be analyzed as a rigid-body problem are important.
Eye gaze direction is critical in all barbell exercises, and is the most commonly misunderstood aspect of squatting and deadlifting.
The chest follows the eyes, and the back angle is the chest. It's very important to understand the role of the trunk segment in the mechanical execution of these two lifts in particular, and how the eye gaze direction can affect the physics of the lifts.
Since this is a deadlift article, we'll stay with pulling. If you look up, the vast majority of you will have overextended your cervical spine - your neck.
This seems probably counterproductive, considering the fact that the spine is best loaded in normal anatomical position, the way it's best configured to transmit force between the vertebral bodies spaced apart with intervertebral discs, who like to be loaded in compression the way they're designed to transmit force.
Your head actually weighs 5-8% of your bodyweight, and keeping your neck in normal extension maintains the position of more of your body's mass in front of the bar. It actually produces a longer back segment, some of which is then positioned forward of the bar.
We've already observed that you can't pick up a heavy object that's too far in front of you, and that more of your mass forward of the bar balances the alignment of the bar's center of mass and that of your body.
Keeping your head in line with your back maintains this mass relationship much better than an overextended cervical spine. This effect is obviously more important in a lighter pull, like a clean or a snatch.
Now, observe the anatomy of the traps. Note that it has an origin from the base of the skull all the way down to T12, thus making it the longest muscle origin in the human body.
Note that almost all these fibers insert on the spine of the scapula. The traps therefore support the scapula, which articulates with the arm, which holds the bar. The bar therefore hangs from the traps, which hang from the spine, from the base of your skull down to