While I have been running since 1986, only in recent years have certain elementary rules of training for that activity become clear to me. Had I known these principles decades ago, it would have made a difference. I have by no means discovered all of this myself, but learnt much by reading, the biggest source of inspiration being the Verheul method or Souplessemethode. Additional insights came from sources like ChiRunning, the Pose method, natural running, and barefoot running. Below I interpret these ideas, based on my own experience.
This is done with short moderate-intensity interval training over distances up to 400 metres, with long recovery breaks. Effort is, at most, just below (slower than) the anaerobic threshold, while the breaks are the same distance as the fast run, and are covered by walking and jogging (or fully walked, in case of beginning, slow, or injury-prone runners). In other words, effort is such that one does not pant after each run. Since endurance runs tend to spoil this resilience and may lead to a heavy style, they are avoided in training. "Endurance run" is to be understood here as uninterrupted running for more than about 10 minutes. Another reason to avoid endurance runs is that they cause injuries, to my experience.
The long breaks serve to let muscles and tendons regain their full elasticity for the next fast run. This way, a light-footed running style develops automatically, without the need to do specific technical exercises; a definite advantage compared to technical methods like Pose and ChiRunning, which require much practising and attention to technique.
This is so because of the detrimental effect on style that endurance runs have, and the injuries they may cause. Instead of traditional endurance runs, long moderate-intensity interval training is used, with efforts just below (slower than) the anaerobic threshold, lasting 3 to 6 minutes mostly. The effort is such that one does breathe deeper, but without suffering. The recovery breaks are executed the same as with the short interval training, with the corresponding longer distances.
These long interval sessions are only done after one has first developed a light-footed, resilient, economical running style via short interval training. They are reserved for somewhat advanced runners, therefore. This long aerobic interval training raises the anaerobic threshold, and is to be preferred over actual anaerobic training. Races, which should be run frequently, also improve aerobic capacity, and can already be done before one has advanced to the long interval sessions.
If one is injury-prone, the trouble with running is that when one runs long enough to positively affect the aerobic system, that tends to cause injuries. Interval training with efforts of several minutes is a way to get around that, but still strains the kinetic apparatus to some extent. First developing a light-footed style with short interval training reduces this problem. If one still finds that the long interval sessions are too hard on the kinetic apparatus, one may try replacing (part of) them with other aerobic activities like cycling, walking uphill, walking up stairs, and so on.
In other words, it should be possible to do that amount and intensity of training daily without injury or other symptoms of overtraining. No doubt, according to this rule, most runners train much too hard and/or long. Implied is that the ideal approach to running is frequent (even daily) but light training. Such an approach is common in lower-impact sports like cycling, but runners traditionally train so hard that they habitually damage the kinetic apparatus. Unfortunately, the kinetic apparatus can only take a limited amount of running, depending on individual disposition, and this amount, I am sad to say, has in my case not increased over the past three decades. Since I am not unique, this will be true for many others too. Therefore, frequent but light training is advised especially if one tends to develop injuries when training as a runner.
The principle of getting incredibly strong through frequent but light, soft training is illustrated magnificently in the parable of the Chinese boy who became a Shaolin monk. I myself had an experience like that in 2016 when, after a year and a half of training according to the Verheul/Souplesse method, I ran a 10 km and tried to break my personal record from the 1990s, when I trained harder and was some twenty years younger. During the race there was a hard cold wind from the north, and it was as if I was standing still now and then when it hit me. Runners were almost blown off the road, and I gave up all hope of improving my youthful best time. Still, I saw ever fewer opponents ahead of me, and someone by the road called out I was in third place. When I finally reached the finish line and looked at the clock, I saw to my utter astonishment that I had taken more than three minutes off my personal record and was first in my age category.
Once one has found a training volume that can be sustained without injury, one should stick to it for the time being rather than raise it. The progress will consist of a spontaneous slight increase of speed in training, and of improvement in race performance. Only when no improvement in races occurs any more, training volume may be increased by one step, and that new volume is then again kept constant for some time (provided it causes no injuries).
The systematic increase one sees in many conventional training schemes is dangerous as it inevitably leads one to one's limits, resulting in injury or overtraining.
Anaerobic training teaches the body to work anaerobically and form lactic acid, and that (well-trained) energy system will then be put to use early on in a race situation, causing one to slow down due to the buildup of lactic acid. Therefore, anaerobic training should be mostly avoided, the races being the main anaerobic stimuli. Also, the aerobic interval training described above (done just below the threshold) has the effect of raising the anaerobic threshold. Stimuli just below the threshold raise the threshold, stimuli above the threshold may lower it.
Walking puts less strain on the kinetic apparatus than running, and since injuries of that apparatus are the main limiting factor in running training, it makes sense to walk when possible, and only run when a speed is desired that can not be walked. If one is injury-prone, much can be solved by walking instead of jogging during the warming-up, cooling-down, and recovery phases of interval training. To understand why this is so, one should realize that the total distance run or jogged over a given period (week, month) roughly triples when one goes from walking to running in the warming-up, cooling-down, and recovery phases; a significant increase in strain on the kinetic apparatus! Also, slow running is bad for style, and slow running in a bad style is harder on the joints than is fast, resilient running.
A consequence of walking the warming-up phase is that the first few runs in short interval training (and the first run in long interval training) need to be done extra slowly to ease into it. It will work anyway, worry not.
Factors like excitement, adrenaline, being well-rested, lighter clothes, and the presence of other runners around one make running in races easier than in training. For that reason, races are the time to run continuously (which is better avoided in training) and to exceed the anaerobic threshold. In races, one can do that with less risk of injury. Races are the only endurance runs and tempo runs in this training approach. Injuries are mostly caused by training, rarely by racing. A logical, yet counterintuitive, conclusion that must be drawn is that injury-prone runners should run many races and train lightly; extremely injury-prone runners might consider not running in training sessions at all and only running races frequently.
There is a rough relation between the speed in relaxed interval training and race speed; in a race, one can maintain a particular speed a few dozen times further than in easy interval training. This factor has a wide variation of say 15 to 40, averaging around 25. For instance, if you can run 10 km in 50 minutes in a race, you will run 400 metres (1/25) in about 2 minutes (1/25) in easy interval training. Similarly, 200 metres interval sessions are done around 5 km race pace, 100 metres around 3 km race pace, and 1000 metres a bit slower than half marathon pace.
Regarding how often one should run races, once every two to three weeks may be the optimum for many; after that, the effect of a race wears off. Racing every week is the standard for talented runners. If desired, anyone can run a race each week, but the average runner will have to abandon training altogether then and just rest in between races.
For a good running style, the muscles around the waist, hips, and upper legs are so important that one should do so-called "core stability" exercises at least several times a week (daily is best), either included in running sessions or separately; I prefer the latter. There are many sources that explain such exercises, so details are not given here. As a bonus, these exercises are also the best prevention of lower back pain I know of.
If one has certain recurring injuries that require specific muscles to be exercised permanently, these exercises can be done together with the core stability exercises. The set of core stability and possible other exercises should be so light that it causes no muscle soreness (except when one does a given exercise for the first time).
I recommend a full-body integral flexibility routine of several minutes, which must be done daily. It is important that the exercises are always executed in the same order, in an uninterrupted fluent movement. Thus, one can remember them well and carry out the routine fast without thinking, spending little time. I prefer to do this separate from running, not directly before or after it, and sometimes in one session with the core stability exercises. An example is the flexibility routine developed by Nicholas Romanov, inventor of the Pose method of running.