After 19 years of running I can answer a few questions. These are my answers, and other experts will give different ones. Also, my answers are not applicable to each individual runner; different builds and amounts of talent require different approaches. But still, here they are.
At least once a week (because you typically lose shape in two weeks of non-running). More only if possible; the limiting factor is the amount of strain the kinetic apparatus can take before getting injured. In other words, the limiting factor lies in how fast the kinetic apparatus recovers from the strain of running. Muscles need three to four days to recover from serious training, so one to two serious training sessions a week are the natural maximum for many. Talented runners are able to do much more though; you have to find your own limits in this through trial and error.
If you want to train more often than your body seems to allow, this is only possible by drastically reducing the intensity of those extra sessions. For highly talented runners this means running at a very low pace; for most people it will mean including walking breaks or even walking all the way, for even running at the lowest possible pace may be too strenuous in these "recovery training" sessions in between serious training.
The great danger of light recovery sessions is that you overdo it, thus actually hindering recovery, reducing the amount of serious training you can do, and causing injuries. Exactly those same dangers exist when you "cross train" for recovery; cross training means to do an alternative activity like cycling or swimming.
People mainly do these recovery or cross sessions because they are so enthusiastic about running they want to spend time on it every day. It is a psychological thing. It is claimed recovery or cross training speeds up recovery (and medically it should), but to my experience it does not, and often does the opposite. Complete physical inactivity between serious training sessions gives the best result, but is psychologically hard to work up. Especially when preparing for a race, it is almost impossible to constrain yourself to inactivity several days a week.
When preparing for a long distance race, the first necessity is to get used to being in motion for the duration of the race. In the one and a half to two months preceding the race, several (at least 5) long runs are needed that last as long as the projected time for the race. The pace may be well below race pace, so that one covers between 2/3 and 4/5 of the race distance. Apart from getting you used to race duration, long runs make the kinetic apparatus stronger, teach your body to use fat as a fuel (rather than just glycogen, which is used in shorter runs) and may make you lose weight (fat) if needed.
In terms of distance covered in training (training volume), in the last several months before the race this should be at least three times the race distance per month. When the conditions in this and the preceding paragraph are both met (the one in the preceding paragraph being by far the most important), one can safely finish the race. This formula is correct for races that last between half an hour and six hours. I am aware that training schedules in books typically let you do a multiple of this. But this suffices.
When you are not preparing for a race, it is not possible to give a recommendation as to how far to run, except that a training volume of 40 km per month is about the minimum to get the health benefits of running. Also, if one wishes to lose fat, long slow running is best, because that is when the body uses fat for fuel. In shorter faster running, glycogen is used and the fat is left intact (even though you use the same amount of energy). How long (to lose fat)? Two to three hours is best. That may seem long, but is easily possible when the pace is set low enough. Also note energy consumption is determined by the distance you run (and your body mass and the efficiency of your style), and not by the duration of the run. Energy is force times distance; time is not part of the equation.
A question that sometimes comes up is: how long will you be able to keep running at more or less "race pace", based on recent training? My formula to predict this is to take the average length of your last five long runs (provided these took place within the last two months) and multiply that by 1.33. Beyond this distance you will probably have to go very slowly.
With this is traditionally meant: distance covered in training per week. But that is not the smart way to control volume. The limiting factor, for most runners, is the strain that can be put on the kinetic apparatus without causing injury. And while muscles can adapt to a new level of strain or recover from slight damage in 3 to 4 days, the most critical tissues - tendons - need a full month for that because of their slow metabolic rate. The same goes for bones and joints, but tendons are the most notorious for ending running careers.
So to control the strain of training safely, and allow tendons, bones and joints to adapt, training volume should be monitored (and increased when desired) per month, not per week.
How to smear out the volume for a given month over that month depends on the relative size of the volume; in a low-volume month it is typically smeared out equally, in higher-volume months, when one is building up the volume, there is usually a steady increase over the first about 3/4 of the month, followed by a relatively easy remainder of the month. An alternative pattern sometimes used is that of a hard 1/4 of the month followed by an easy 1/4, and then the same, but with an increase, in the second half of the month. A specific preparation for a race usually has one low-volume month followed by one, two or three ever higher-volume months as meant in this paragraph.
The purpose of controlling training volume is to prevent injury and overtraining. It is not to influence race performance directly; race performance is only directly determined by the length of the long training runs and the quality of the speed training sessions.
From the previous paragraph it follows that no particular training volume is sufficient to be successful in a race. One cannot say, "run so many kilometers per month and you will be able to run a marathon". It's the long runs and possible speed training sessions that determine that ability. But there is probably a required (but not sufficient) volume for each race distance, and I estimate that at three times the race distance per month. Note the difference between required and sufficient. There is a required volume, but no volume is sufficient.
A side-effect of high training volume is that it makes you slower than you are capable of. So before an important race, one reduces the volume ("tapering") to obtain one's true speed. This taper lasts between a few days and a few weeks, depending on the race distance. It is said that for very long "ultra" races that last more than a day, it is better not to taper but to stay in a rhythm of daily running very long.
Running starts where walking ends. To find out where that is you may take some long measured timed walks over the course of several days (your walking pace varies per day and is an indicator of running shape of the day). I find my walking pace varies between 6 and 7.5 km/h.
When I run I go at least about 9 km/h; slower is possible but awkward. So there is a gap between highest walking pace and lowest running pace. This gap can be filled by alternating walking with running; the average pace then ends up between 7.5 and 9 km/h. These numbers will vary individually, but important is that you can control the pace in this range by including walking breaks.
The reason to fill this pace gap is that it enables you to control the intensity of the long runs; you need to get the pace right. And "right" means: such that you can do the long runs without injury. And for some, the right pace for long runs may be in this gap.
If you want more than just to finish a race, you need to do something to get faster. The initial limiting factor in long distance running is your anaerobic threshold; that is, the level of effort where you get "out of breath", where running becomes rapidly more uncomfortable and you are forced to slow down or stop.
The amount of oxygen inhaled is then no longer sufficient to burn glucose at the required rate for that pace, and your muscles start processing the glucose without oxygen, resulting in the formation of lactic acid. That causes discomfort. After you stop or slow down, the lactic acid is cleaned up again, which requires oxygen, so you "breathe afterward" so to speak. A popular fallacy is that lactic acid causes muscular pain; it does not. That pain results from tiny ruptures caused by repetitive strain.
The threshold pace depends on talent, shape, and how far you want to hold the pace; for the very best it's around 19 km/h, for others maybe 10 km/h. I find it is about 15 km/h in good shape.
To get faster in races, you need to shift the threshold to a higher pace, which can be done by regularly running close to threshold pace, e.g. just under it. As it is strenuous to do so for a long time at a stretch, one typically does this via aerobic interval training. You then run a relatively short distance (typically 1000 meter, but everything from 100 to say 5000 meter may do) just below (slower than) threshold pace, rest a few minutes (walking), and repeat that a number of times. Initially you repeat it once or twice, later more often. How often depends on the distance of the individual runs; advanced runners go up to a total distance of about 10 km (as in 10 times 1000 meter). After a month of this type of training (done e.g. once a week), your anaerobic threshold will lie significantly higher, so you can run much faster without discomfort. After two or three months the result may be even a bit greater. When you stop doing training where you run around threshold pace, the threshold comes down again.
A more strenuous way to get even faster is to run beyond (faster than) anaerobic threshold pace. That way you get used to the feeling of discomfort of lactic acid, and learn to stand it, which enables you to be a little bit faster in races compared to running below the threshold (also, the threshold itself may be shifted a bit more this way than it has already been with aerobic interval training). This type of training too is done via intervals (anaerobic interval training in this case). Typically you need to run at least 250 to 300 meter at this pace to get significant discomfort. The resting breaks are longer than with aerobic interval training, and the number of repeats is lower, like 2 to 5. When the initial distance (e.g., 300 meter) no longer gives enough discomfort, you go to 400, 500, 600 and so on meter. I think some people go up to over 1000. But beyond 500 or 600 meter, it will not be purely anaerobic anymore, as the aerobic system kicks in after about one and a half minute, and it becomes a combined anaerobic/aerobic effort.
However, anaerobic training is usually not recommended for other than very advanced runners. One says the risk of breaking down, getting injured or overtrained, is high. It is warned against in many books on running.
The typical approach to getting faster is to first get a "basis" via long slow running, then do the aerobic interval type of things, and finally the anaerobic stuff (if one ever gets to that at all). I sometimes doubt this order (but I seem to be the only one), and think the other way around might work better. First short fast running, then extending that, so that you go via anaerobic running to aerobic and finally to long and slow. This seems more natural to me; after all, children run spontaneously exclusively anaerobic. They run as fast as they can for a few seconds and then stand still (when they've run out of phosphates, to be discussed later). Then when the teacher tells them to run to the other side of the field and see who arrives first, they have to extend that a bit and get into anaerobic glucose processing. But still not aerobic. They only go aerobic when specifically instructed to constrain themselves, hold the pace down and so on. Aerobic running is an adult, learnt, cultivated type of thing. The natural order is from short and fast to long and slow.
Using a map one can find and measure one or more courses to run in training. Two types of courses are possible: "out and back" and "round trip". The first type is the most satisfactory because of its flexibility; using (mental) marking points for various distances along the course one can run any distance up to the twice the course's length. Also, you get further away from your house than with "round trip" courses, which do not have this flexibility.
I have read dozens of books on running. Many are interesting, instructive and worth reading, but there are shortcomings to almost all that need mentioning:
When training schedules are provided, these almost always let you train far too often, too far and/or too fast in the long runs. They do not respect individual differences in amount and direction of aptitude. They are aimed at pure endurance types, probably because they have been written by such types. They are crammed full of wishy-washy short or medium long training sessions to arrive at a certain weekly volume and number of training sessions per week. But weekly volume does not make you faster or help you to the finish, to my experience. Long runs help you to the finish, and fast running makes you faster. In between, rest is needed. Most other things are ballast and come forth from a desire to be constantly occupied with running, or from feeling guilty when one is not training. Yes, I know there are many people who successfully train that way. But I am saying that way of training is not correct or even possible for all types of runners.
A typical example of my experience with schedules in books: a schedule says "do this to be able to run 10 km under 50 minutes". Then I look at the schedule and know (through experience) I would get badly injured in days if I followed it. Too often, too long, too fast in the long runs. But I can run 10 km under 50 minutes, even without specific training, and have run it in 42 minutes with a fraction of the training that is recommended in the schedule. Obviously, such schedules are incorrect for my type, and I can only suppose they are correct for pure endurance types, who have a very high percentage of "slow twitch" fibers in their legs.
You sweat when running, so you need to drink to replace that water. To know if you drank enough, it suffices to look at the colour of your urine after running. It should be clear and transparent, almost colourless, and certainly not yellow.
To achieve that it is important to drink much right before running. I drink about half a liter then. And almost the same when I return. That is more than sufficient for distances up to about 17 km. when I run longer I may take a bottle of 300 or 500 ml water with me and drink while running, every 4 km or so. Or I arrange it so that I pass my house once (or twice in a very long run) to drink at home.
It is funny that books on running always give the most ludicrous advice about drinking; for instance, they tell you to hide bottles or cans of liquid along the course. What are they thinking? That I take the car every time before I go running to hide a can in the grass here and there? That I look around to check if no one sees what I am hiding? Are there really people who do that?
Or they advise to have someone accompany you on a bike or by car to hand the water to you. But they don't say where you can get those slaves. Also this of course defeats the purpose of running entirely.
About what to drink: water. Only when you run so long that you exhaust the body's glycogen supplies, which typically occurs beyond 30 km, it makes sense to drink liquids that contain carbohydrates. That way you postpone the moment of glycogen exhaustion. For shorter distances it is nonsensical to drink anything else than water.
It is also good to know the salt concentration in sweat is lower than that in the body fluids, so that the salt concentration in the body actually rises while you sweat. This means it is dangerous to try to replace the lost salt with salt tablets and the like; the salt concentration in the body then rises even more, the blood becomes thick, and people have died that way. Drinking water is far more important and urgent than eating salt.
If the intensity of a training session does not exceed that of a "warming up" effort, as is the case in a long slow run, there is obviously no point in warming up beforehand. The first several kilometers - about four I would say - serve as "warming up". If the intensity is higher, as in interval training or continuous faster running, it is good to warm up with slower running or walking for several kilometers or close to half an hour.
A point of debate is whether one should do stretching exercises after warming up and before the main part of training. Some are hysterical about this and insist you bring each and every muscle in your body "to length" before even lifting a finger. Others say stretching reduces the tension in the muscles, and that that is bad because you need a certain tension to run; without sufficient muscle tension, the bones and joints take all the strain. Statistics show that runners who do stretching exercises get more injuries than those who don't. Some say now that stretching before training or race is a form of mass hysteria through indoctrination.
I cannot give the final verdict on this. I have had periods of years when I did do stretching exercises before the main part of training, and periods of years when I didn't. The incidence of injuries appeared to be greater in the periods when I did do them, but it is hard to be certain of cause and effect. As I got more experienced in running I got less inclined to do stretching exercises after warming up, and I got less injuries.
Stretching after high-intensity training is a different matter. I have always done that, although I cannot determine its effect. The idea is that muscles tend to get "shorter" after hard work, and that you can keep them at their original length by stretching them afterwards. That sounds logically. I do not know if it has been proven though.
Although stretching exercises as part of a running training session are controversial, one may consider stretching in dedicated sessions say once a week, as a separate activity. In such a session of 20 to 30 minutes a wider variety of exercises can be done, with more time for each one. This has much more effect than the brief stretching one typically does after (or before) running.
But even the one has to stay critical; points of concern are:
Is it worth the time and effort put into it? For the result may remain limited, especially if you are a stiff person by nature.
Does the stretching not cause injuries in its own right, which are then mistaken for running injuries? Now and then you have to leave out stretching altogether for a while to verify this. I actually suspect that a significant part of "running injuries" is caused by the obligatory stretching.
Are you not stretching the ligaments rather than the muscles? Many exercises taken from yoga affect the ligaments, making the joints looser. The great flexibility one sees in people who do a lot of yoga is often due to ligaments stretched beyond their normal length. And this may be unwise for runners or even for anyone. One wants the ligaments to stay as they are, giving firmness to the joints so that they won't get dislocated. It is important to know that ligaments are not as elastic as muscles; once you have stretched a ligament beyond its original length, it stays that way and that joint will never return to its original firmness.
It has been said endorphin is produced in the body while running, causing you to get "high", not feel pain, and even causing addiction to running. Researchers have found however the amount of endorphin produced is too low cause these effects.
It is a fact that running can cause a sense of euphoria, and this occurs after running quite long (more than half an hour) at a fairly high pace (somewhat below the anaerobic threshold, but certainly not too far below it; well above the pace where one would use fat for energy production, so in the pace range where glucose is used aerobically). I do not know what causes this euphoria; perhaps it's purely psychological, perhaps it's the high glucose and oxygen levels in the blood, or the fact that the brain works faster (reaction time is shorter when measured immediately after running), or perhaps it's endorphin after all.
Related to this may be the fact that running causes the brain to form new and more connections between neurons, and even to form new neurons. The euphoria and the new formation probably result from same cause, as does the known mild anti-depressive effect of running.
An important dilemma is whether to train on hard surfaces (concrete, asphalt, stones) or soft surfaces (sand, grass, forest paths). Provided one has the luxury of choice.
Hard surfaces are more available and give less risk of acute injury through missteps. But you need shoes with good shock absorption, which - the shock absorption - does not last very long, so you have to buy new shoes often. The soles also wear fast. And the risk of chronic injury is greater because of the pounding on hard ground; you can do less training before RSI problems start. Especially in forms of training where you run fast, hard surface is very undesirable because of this risk; stress fractures in the shin bone are typically caused by fast running on hard surface. Another point of concern is that the presence of shock absorption in the shoes reduces stability and thus promotes deviations in foot development, which may lead to injury.
Soft surfaces are safer with regard to chronic injury, and also leave the muscles more flexible and make them stronger. And you don't need so much shock absorption in the shoes. On the down side, soft roads and paths are hard to find, and the risk of missteps is greater. And you get dirtier with bad weather.
On the whole, soft surface is preferable. Hard surface is acceptable to cover great distances at low speed.
Advantages of taking part in races are that you can run much faster than in training (partly for psychological reasons, partly because the other runners break the wind, and so on), and that you get some official proof of your performance. Races also function as speed training for later races.
A problem with the road races where everyone can participate is that they are focussed on the longer distances; 15 km to the marathon are considered the real stuff. 10 km is seen as a short distance, and 5 km as only for beginners. Very rarely you see a 3 km race. There is no recognition of the fact that some are naturally better at shorter distances. Serious races over short distances are only held in track athletics, but are not open to non-members (of track athletic clubs). So you are forced to run very long distances to be taken seriously in road running. And if you try to run well in the shortest available race (like 5 km), you lose too much time at the start to reach your best performance. It's crowded, so you can't get away well in the first 100 or 200 meter. You lose perhaps half a minute before you get to run your own pace.
The longer the race, the lower the average pace. In books you see formulas for that like, "if the distance doubles, the pace drops by X %", where X is anything between 4 and 10, depending on who wrote the book.
To my experience in races however, pace drops more like 12 to 15 % when distance doubles. Again, this may be a matter of aptitude; maybe the formulas in the books are only correct for endurance types, and the decline is much greater for strength types. Another explanation might be I have never been able to do enough training for the longer races (e.g., for my 2005 half marathon race of 1:45.55 I did not do any speed training at all, only long slow running up to 18 km), being held down by injuries and the need to prevent them. But then, that boils down to same thing in the end.
So my current thoughts are that the speed decline when distance doubles may be around 3% for pure endurance types (95% slow twitch fibers), but probably around 20% for pure strength types (95% fast twitch).
When I speak of slow and fast twitch muscle fibers here I mean the muscle composition in the lower body; it is different in different parts of the body, e.g. it is possible to have mainly fast twitch in the lower body and mainly slow twitch in the upper body. For running, the composition of the lower body is relevant in determining if you are a strength type or an endurance type. Note that a strength type in running may still have a skinny upper body with mainly slow twitch fibers (fast twitch fibers become big when you train, slow twitch fibers stay thin).
Running, combined with walking, has played an important role in human evolution in at least two ways. Humans have been hunter-gatherers from the advent of homo erectus, 1.8 million years BC, to the rise of agriculture from about 8000 BC on (by that time they had become homo sapiens, but only because we call them that). It is estimated daily distances up to about 40 km were covered in hunting and food gathering. Actually there have always remained peoples who live that way.
Also, humans, both erectus and sapiens, have migrated from Africa to various parts of Asia, Europe, Australia and America in that same period. On foot, obviously. Large groups will have walked after scouts found suitable areas to migrate to in long runs of several days or more.
In both hunting-gathering and migration it is relevant that humans are the fastest land animals over long distances.
When dealing with long distance runners it is quite conspicuous they are less aggressive and violent than the general population. You typically do not hear of marathoners being suspended from racing because they hit an official (which is not uncommon for sprinters), and you don't see the scenes of violence and aggression that are inherent to (not American) football and ice hockey. Those who participate in long distance events are clearly a selection from their population, not just by ability to run long but also for peacefulness, constraint and emotional stability.
It may be that those with an innate aptitude for endurance are also peaceful by nature, and it may be that training for endurance sport makes you peaceful. Although the first is more likely than the second, both are probably true. In the light of violent crime and war, this observation may be important.
Just after writing the previous paragraphs I learnt of a Netherlandic female marathon runner in her mid-thirties who has become notorious for her violent outbursts - calling names, throwing water bottles - during and after races, both towards other runners and the audience. She says it comes forth from frustration, and that she is almost 36 and wants to make the most of it while she still can and therefore is greatly annoyed by anything that goes wrong. I have been reading about her, and for the moment I think her behaviour fits Tourette syndrome, the tics and outbursts being brought up by the stress of running at a high level of intensity.
It is hard to run if one is overweighted, and up to a point, lighter is better for long distance runners. Only for sprinters this is different, as good sprinting requires a more muscular build, and muscles are heavy. Good long distance runners though tend to have a BMI (Body Mass Index) between 18 and 20. A BMI of 21 would be exceptional for a good marathon runner.
The BMI is the quotient of one's mass in kilograms and the square of one's height in meters. It is said normal values lie between 18 and 25, but it really depends also on how you are built and how muscular you are. For a naturally slender person, a BMI of 23 means overweight.
Long distance runners are in this sense - body mass - somewhat deviant from the general Western population, where most people eat far too much and probably more than half of the population is overweighted. This causes a shift in perspective; the norms for overweight as used by insurance companies and the medical profession are too lenient (meaning they underestimate the number of overweighted persons). On the other hand, the term "anorexic" is more and more used incorrectly for persons - like fashion models, actresses and some athletes - with a perfectly normal body mass and figure, possibly from a kind of jealousy. Obesity has become the standard, and with a normal healthy weight one is now at risk of being mistaken for skinny or anorexic.
While the many beneficial effects of physical exercise are scientifically proven and undisputed, science does not agree on exactly what constitutes a healthy diet. We do not know what best to eat, and in what proportions; individual experimentation still pays off.
What is known with regard to long distance running is that it tends to exhaust the body's glycogen supplies, and that those are best regenerated through a diet relatively high in carbohydrates and low in fat and protein. Compared to normal people, runners need more carbohydrates, it is thought, but not more fat and protein. Note this does not mean that runners should eat very large quantities of carbohydrates (those would only be stored as fat), but that the percentage of carbohydrates in a runner's diet should be higher than that in a passive person.
On the other hand, humans have been nomadic hunter-gatherers for most of their existence. It makes sense that we would function best on the corresponding diet, which consists of meat, fruit, vegetables, nuts, seeds and berries. Absent therein are our present foods high in carbohydrates, like bread, potatoes and pasta (also absent is milk and anything made therefrom). So the modern long distance runner's diet differs from that of the ancient hunter-gatherers. Yet hunter-gatherers, of all people, cover great distances daily on foot; some estimate this at around 40 km on a typical day, while Australian Aboriginals are said to go up to 70 km a day to reach the next water well. Apparently our modern carbohydrate-foods are not required for such feats. But are they better than the ancient diet? That remains open to experimentation.