At one point in middle school, I was part of a swim team. While I managed to keep up with the others for the most part, I was never leading the race. Over the years that followed, I attempted to re-learn the motions of swimming on my own schedule, and from a more physically informed perspective. This article outlines aspects I have learned which are often overlooked in swim training, approaching the subject with consideration of fluid dynamics, drag, and energy conservation. A lot of the ideas presented here were inspired by Terry Laughlin's "Total Immersion" and "Freestyle Made Easy", which present the athletic side of the subject.
Swimming is heavily influenced by technique, and effects of slight changes in body position are both subtle and additive, or in other words, good swimmers optimize their efficiency through doing lots of small things right. Because any one "small thing" does not display an obvious advantage, it is important to establish a method that will allow for clear comparisons between different swim techniques. I would encourage swimmers to approach this like the metaphorical scientist (lab coat, clip board, and all) and set aside sessions not for speed or strength training, but for just experimenting with different techniques and improving strokes. One way to determine the efficiency of a swim style, is to measure the number of strokes (not time!) required to swim one length of the pool, with the goal of using the fewest strokes possible. For a swimmer that has not focused on stroke efficiency, the improvement in this number could be tremendous (reduction by a factor of 2 or more). Thus I would recommend establishing a baseline of how many strokes it takes you to swim a set length with your current swim style, and then see how this improves as you make adjustments for efficiency.
Much like the airplane is interpreted as being affected by the forces of thrust, drag, lift, and weight, the swimmer in the water is affected by the same forces. A major difference is that water is about 1000 times more dense than air, so buoyancy makes a substantial contribution to lift, and all motions are slower. However, there is a similarity in that small changes in the shape presented to the fluid flow can have major implications on movement efficiency. Since the velocity of swimming is relatively slow, and there is not much tactile feedback from water flowing along the surface of the skin, it is difficult to intuitively understand what parts of the body are causing drag. In fact, I recommend that aspiring swimmers try indoor skydiving - floating on air in a vertical wind tunnel - because that experience makes it easy to observe how even tiny changes in body shape (such as holding the fingers together or apart) can change one's motion in the wind stream. The same thing is true in the water, just on a slower time scale. Because the slowing effect of drag appears easier to interpret, I will begin with the more overlooked vertical axis - lift, buoyancy, and weight.
The vertical level of the swimmer in the water has a significant influence on the efficiency of swimming, namely in causing drag if the swimmer's body is low under the surface. Even one leg, or one foot, hanging down below the vertical surface of the rest of the body, like a sideways rudder under a boat, will substantially slow down the swim. High efficiency comes from being very close to the water surface, because less water is moved, and because the water that is moved is closer to the water-air interface which may shear freely. For the same cross-sectional area, having that area oriented so most of it is near the water-air inteface will lead to lower drag than having most of it deep below the water. Thus, it is in any swimmer's interest to control his body position and immersion depth, so as to stay as high as possible in the water. This is achieved both by dynamic lift (from forward velocity and an upward angle of attack) and by buoyancy (from the body displacing a larger volume of water).
Buoyancy is important to all lifeforms (and objects) in the water. With a constant body weight, displacing a larger volume of water will cause an effective upwards force on the body. Since the human body is mostly water by mass, humans are close to neutrally buoyant in water. Fat is less dense than muscle, so body build has a significant effect on buoyancy. Swimmers should try to gain some experience with scuba diving, because buoyancy control is exemplified in dive training. But without going that far, it is possible to get a good idea of the effect of buoyancy with an experiment in a deep pool (deeper than standing height). Start by jumping in, holding on to the edge, and inhaling as much air as you can in one breath. Hold your breath (with lots of air in the lungs) and let go of the edge, relaxing your muscles and allowing the body to float freely. After a few seconds, make a mental note of the body position in the water. In my case, due to low body fat, my head is just barely above the water surface, and the rest of my body is vertical with legs pointing straight down. Hence my build is not well suited for swimming, and I have never been particularly fast.
Next, release a bit of air from your lungs, and note how your body immediately sinks lower into the pool, as well as the rates at which your arms and legs sink lower. Keep doing this, exhaling small volumes of air, until you begin to sink towards the bottom. Buoyancy is an unstable situation (more precisely, this is determined by compressibility), meaning that for a given displacement and weight, you will either float towards the surface, or sink towards the bottom. In between these two edge cases, there is an ephemeral region of neutral buoyancy - where you float without being pushed up or pulled down. Find how much air you have to hold in your lungs to reach this region, get a feel for how much of your lung capacity is used, and make a mental note of your body's 3D position in the water relative to the surface when you do so. Finally, repeat this exercise applying the typical breath pattern you take while swimming (but not actually swimming at this point - just staying still and floating up or down). This will give an idea of your buoyant depth in the water and how much your breathing affects it. (Most likely your body position while swimming is quite different from the position you take on in this exercise with muscles relaxed, and this difference is due to the effects of dynamic lift and propulsion.) In my case, I begin sinking at about 0.8 x lung capacity, and my body is vertical in the water.
Here an important connection is made - between breathing and buoyancy. In fact, breath control is an important contributor to swimming efficiency, specifically, one should practice a breathing style that keeps the lungs filled with air for as much time as possible. This results in high buoyancy overall. I identified this as a major factor behind my poor performance on the swim team - I would take a quick breath in and then breathe out at a constant rate until the next inhale, "blowing bubbles" all along the trail. While doing this, my buoyancy was dropping and my body was steadily sinking into the water, substantially increasing drag and making me slow. Presently, I use a different technique - exhale rapidly, then immediately inhale rapidly and hold the breath all the way until the next exhale. Except for the fractions of a second when I exhale and inhale, my lungs are filled with air, enhancing my buoyancy, raising my body position, and reducing drag. One way to summarize this technique is that swimmers should never "blow bubbles" under water - do a fast exchange of air with the head held above the water, then hold your breath until the next stroke.
Even with this improved breathing technique, because of my low buoyancy to begin with, I sink by tens of centimeters when taking a breath. My body position is then restored by dynamic lift, but this absorbs forward velocity and converts it into upward velocity. There is another optimization which may be applied if your overall buoyancy is low, taking advantage of the buoyant lift: after the body sinks due to exhaling, and you have filled your lungs with air, take a "mini dive" under the water, such that the buoyant force which now seeks to lift your body to the surface has to act against the water overhead. Then by angling your body upwards to "push the water back" as you float up, you are allowing the buoyant force to propel you forward.
The rest of the lift comes from dynamic effects - angling your body to travel upwards, or using thrust to propel yourself upwards. Both of these require energy expenditure, so it is in your interest to maximize buoyancy through breath control as much as feasible. Slight efficiency improvements may be realized by controlling the timing of your breath relative to the propulsive effects of the stroke, because likely the body position (and hence the angle of attack) changes when taking the breath. In my case, again because my overall buoyancy is low, I time this such that I obtain some lift from leg propulsion coincident with the sinking due to exhaling, and this reduces the peak depth below the surface reached by my body.
Note another aspect of flotation comes from the interplay between the center of lift, the center of mass, and the center of buoyancy, which are not identical. Control of these allows turning the body such that most of it is near the surface, so it is more like an "arrow" (horizontal) and less like a "pole" (vertical). Here the position of the limbs relative to the lungs matters. For me, holding my arms forward and over the head, hands stretched out towards where I am swimming, contributes not only by making a more streamlined cross section (like a long racing scull boat), but by placing weight forward of my center of buoyancy, hence exerting a moment that lifts my legs up towards the water surface, thereby reducing drag much more than the "scull boat" effect alone. Thus, along with keeping my lungs full for a high fraction of the stroke, I also seek to keep both arms held forward for a high fraction of the stroke.
Drag slows down the swimmer and is the means by which energy is dissipated to the water. The best way to reduce drag is to reduce frontal contact with the water. Since some contact is inevitable, it is desirable to minimize the depth and width of contact. It is easy to imagine that rowing a boat that has an extra oar dragging behind it in the water will not be as easy as rowing a boat with a streamlined hull. Similarly, a swimmer who allows a leg or a foot to drop down towards the bottom of the pool is making his body more like the boat with an extra oar, and a swimmer who keeps the feet pointed and legs close to the water surface makes a streamlined shape that has less drag. It may be instructive to spend some time submerged at a pool, wearing swim goggles, looking at other swimmers' body position from the side. Most casual swimmers will have their legs immersed well down in the water during some part of the stroke, and that part will cause major slowing down of forward velocity. Because water is dense, a difference of even a few cm will lead to measurable performance improvements. For the competitive swimmer, streamlined swim suits and swim caps will further reduce drag.
In addition to not letting the feet drop down into the water, the depth and breadth of the hands during the stroke should be controlled. When analyzing my performance in breast stroke, by going through a single stroke piecewise and tracking change in velocity, I found that a major and substantial slowing down occurred when I was bringing my hands forward after the propulsion part of the stroke. In fact, I would lose almost all forward velocity during this move. I trained to reduce drag during this part of the stroke by keeping my hands very close to the body and also turned with palms facing up so that only a small cross-section was presented to the water. Once I learned this non-intuitive motion, I obtained about 0.2 x reduction in number of strokes required to traverse a lane.
On the propulsion side, the equation is reversed: it is now desirable to increase the area and depth so that a high force may be developed. The hand taking the stroke should be extended deep into the pool (with the tradeoff of an extended hand being harder to move because of the mechanical moment, so some intermediate position should be found by the swimmer) and the palm kept wide to push the water back. Possibly opening the fingers slightly will lead to a higher area. However, the hands should not be the only, nor even the primary, means of propulsion, because both the area and the muscles for the hand are limited. The legs have a much larger area and muscles, and the largest area is the body (core) itself. In my estimation, effective swimmers obtain about equal propulsion from these three areas. With the legs and core, water is pushed back tangentially, by a wedge-like body position moving vertically while moving forward. This is how fishes swim - they can get a little bit of propulsion by pushing with their fins, but they get dramatic speed by curving their entire body alternatively left and right, pushing water backwards in the process. This is helped by the distribution of elasticity and mass in their body, in that the tail is more flexible and comes to act like a paddle when the strain wave reaches it. To swim like a fish you would have to keep your head sideways, so it is more common for strokes to use this motion in the up and down directions. From a floating position, push down with your head, then as if "riding your own wave" keep pushing down with the chest, hips, and legs, so that the water that started near your head is pushed back towards your feet. Then repeat this with pushing your head up. An exercise that targets this motion is swimming entirely immersed underwater (holding your breath, no "blowing bubbles"), with arms outstretched in front and body kept long, without moving the arms or legs but solely by curving the body up and down. I am able to swim an entire lane in this manner without exerting much effort.
With the legs it is more difficult to obtain a good angle to "push" the water, and this is often compensated by forceful acceleration (even though splashing water all over the place is no indication of speed or efficiency). However one must be very careful that the powerful kick does not turn into an "extra oar" after the propulsive part is finished, because a leg hanging down in the flow will sap momentum away from the swimmer much more readily than the swimmer can replace it with another kick. I am a fan of "sideways kicks", which are possible in free stroke, or shallow and hip-based kicks such as in breast stroke, because they do not require the leg to be moved down into the pool. The toes should be kept pointed during the kick as appropriate, again to not create extra drag. The exercise of holding a foam board in front of the head and kicking one's way down a lane is common but has not improved my kick at all (other than increasing strength). Starting with an ineffective technique, as I did, this exercise would tire me out and my technique would deteriorate even more towards the end. I would recommend instead trying this exercise with short fins worn on the feet, because this accentuates the "tangential push" effect and helps to develop a feel for what an effective kick does, which may then be extended to the kick without worn fins.
In a competitive setting, an effective start can easily be a determining factor in winning a race. By pushing off the land while in the air (or perhaps on the water surface), the swimmer gets effectively free forward momentum. The goal then is to increase this momentum by pushing effectively, and later to effectively transfer this momentum into forward velocity in the water. The typical start is tens of cm above the water surface, and this represents potential energy which may be converted into forward velocity by an effective diving position, adding to the existing forward velocity component obtained by pushing the platform back. With further practice, the swimmer may add extra initial energy by pushing upwards as well as forward during the takeoff, thus increasing the vertical potential energy at the peak of the dive and the distance of first water contact. After the first contact with the water, the motion taken should be similar to that discussed above - "riding the wave" and pushing the water back tangentially with your entire body, while minimizing drag by not letting any body parts hang down in the water. Matching the boundary conditions implies that the optimal angle to enter the water is slightly under 45 degrees, and the takeoff should be at an angle approximately half-way between that and the horizontal (so not entirely forward, but not at 45 degrees either). Swimmers seem to find this angle intuitively with enough practice, which is really the main takeaway here - the start should be practiced specifically during a session and not just left as a "thing to do" before swimming laps.
One of the easiest segments to lose time in a race is during the turn around from one lane traverse to the next. When coming up towards the end of the lane, it is important to keep the forward momentum going as long as feasible, so not to start the turn too early. The turn itself should be rapid, putting the swimmer in a position that allows for a long push off the wall, and maintaining a streamlined body afterwards will complete the turn. The method of "diving under" and performing a front flip underwater seems efficient, and should be practiced in a dedicated session.