Two per build as I was following the same plan for each I did. I couldn't really tell you if it was a good idea or not but I did get progressively faster as far as standalone marathons got.

I do around 5-6 20 milers and usually one 21-22. Don’t see any reason to run longer than 20 miles, the biggest impactor of your performance is still gonna be mileage.

---

Terrible Tuesday’s Triathlon

Many first-timer training plans will peak out at one 20 miler, based on the notion that if you can get to the 20 mile mark in decent shape then surely you can 'coast in' the last six. However, in a race where you are trying to do well, you need to hit 20 miles with enough gas in the tank to do more than coast. So it's necessary to practice that a bit. As others said, mileage is key, and mileage that is part of 15 to 20 mile runs counts for more than lots of 5 to 10 mile efforts.

---

Less is more.

This will depend on your goals and total weekly mileage.

In general the more you run and the faster you want to go, the more >20 mile runs you’ll do.

i think there is a whole host variables that come into play, like where are in the season and how many long races have you done. Also, depends how fast you are and what your goals are too. I never go on mileage but on time. Beyond a certain point, (think most research says around 2:30) your chance of injury grows faster than your return on fitness. I would never have a slower athlete run for 3 hrs just to get a 20 miler in, but would work on weekly mileage build.

I agree with previous answers.

Shitty first-timer plans peak at 16
Good first-timer plans have one 20-miler.
Top canned plans like Pfitzinger 18/70 and 18/85 have five-six. I call them 'top', because probably above 85 mpw you don't need pre-canned plans.

I would do one(actual training marathon in race), and then one 17 miler, and nothing over 14 besides those two. But I only would do an 8 week build on very low mileage, most all at race pace. Little cycling to supplement the fat burning, and that seemed to get my best out of me...

No doubt no coach or program would council this approach, glad I didnt have any of that interference when I went for it...

1, maybe 2.

But I like to build up and spend a good amount in the 2 to 2.5 hour range.

To go sub 3 I need at least 3 in one month

It's going to be different for everyone. I did several 16-18 mile runs plus at least one 20 miler and one 21 or 22 miler, and a 24 miler and a 26 miler. Two different marathon coaches prescribed at least one 24 and one 26 miler for me. I did have my best marathons and a PR doing this. It's risky and I did have several running injuries but they usually happened early season/winter running. When I didn't have injuries I usually did pretty well and could handle this volume of long runs. My mileage only peaked at 50-55 miles per week. I was early 40s (female) and PR marathon was a sub 3:08. That all being said I've know successful marathon runners who've done more and some who have done less. I now do Ironmans and longest run is usually 21-22 miles in my late 40s.

I ran these long runs since my coaches believed you have to train the distance you wish to race. Many might dispute this. But it worked for me back then. Could I have done as well running less? Maybe, maybe not. Those long runs did give me great confidence for my marathons and I felt super prepared.

---

Death is easy....peaceful. Life is harder.

My marathon racing days were 1986-1996 when there wasn't as much info (good info and bad info) available about how to train for a marathon. My training wasn't that different from what Monty wrote other than a shorter long run. My keys to good marathon times (good for me anyway) were a lot of hard runs, race pace and faster, and moderate daily mileage. Most years I ran a spring and a fall marathon, twice I did 2 spring marathons about 6 weeks apart. For the marathons that I ran decent times, I topped out my long runs in the 15-18 mile range, mostly at 15-16 but I did a few 18 milers. Never ran a 20 mile training run, other than the 2 springs I did 2 marathons I treated the first one (Boston in both cases) as more of a training run than a race and was focused on the 2nd marathon that spring as my target race.

Looking back at my training logs from those days, when I had good marathons I had done back to back 13 mile days at least twice in those training cycles. Both 13's were run pretty hard, just a couple minutes off my half marathon pr at the time. In today's language those would have been "marathon pace" workouts. The 2nd 13 mile day was always difficult, and I think that's where the value was for me - in going out and running a hard 13 on tired legs and getting through it. My last back to back 13's would be about 2 weeks out to race day (as little as 12/13 days out and as much as 16/17 days out). I was usually pretty solid the last 10k of a marathon even though I didn't have the traditional long runs in the bank. If I went out and ran the first 20k within myself I'd be in a great position to race home the last 15k-20k.

For basic daily mileage I was usually around 60-70 minutes. Once or twice a week if I was strapped for time I'd do 45-50 minutes. Wednesdays, track workout day, I'd usually be 75-80 minutes including the warmup and cooldown. I was running pretty much 7 days a week. Normal marathon buildup, the dedicated training cycle, was usually about 10-12 weeks but could be as little as 6 weeks for Boston. Outside the dedicated training cycle I'd stay fit with a wide range of workouts so I always felt like I could adjust pretty quickly if I decided I wanted to do a marathon or do some bike racing or some XC ski racing.

The main reason I didn't do longer long runs BITD...I just didn't know better. I trained with my college coach a lot and he was a good marathoner (2:34) so I just did what he did which was 1:45-2:00 max for long runs. No regrets, it worked out fine.

I was 21-30 years old at the time. When you're younger I think you can get away with less of an optimal training plan and still produce a decent result. My marathon pr is right around where McMillan's calculator says it should be based on my 8k/10k/15k pr's, although the calculator says my 5k pr should be almost 30 seconds slower and my half marathon pr should be about 90 seconds faster than they actually are.

This exactly. If you want to maximize your marathon performance I would say at least 5 20-milers in a 4 month build. Average weekly mileage in the last 2.5 months should be 50-55. If you can do more then great. I would say one day a week of speed work is probably all you need. Volume is king when it comes to proper marathon training.

---

------------------
http://dontletitdefeatyou.blogspot.com

Zero.

If you're a racer in a marathon, you'll train to a speed bias and survive the miles in the race.

If you're a participant in a marathon, you'll train to mileage bias and finish the race.

Younger people (<30) should be more biased to speed. Into the 30's and up, bias will be towards a balance or mileage especially as top speed changes relative to race pace choices.

If you want to go beyond 18 miles in a day, split the workouts.

After 18 miles (or so) of running you're burning muscle. Longer recoveries required.

Race to taper and recover with fast speed and long distance on race day. That saves the body over the years.

Triathletes can substitute the long runs with bike and swim for aerobic stuff, and weights for strength.

https://www.palmtreesahead.com/
https://www.palmtreesahead.com/tactics2faster-new

I’m of the belief that 16-18 is enough for the long run if done consistently. I’ve always felt that it beats you up past the benefit that you would see to go beyond that.
It’s also individual. I am someone who never does more than 13 miles at once, or 2.5 hours on the bike at once. I carry a ton of endurance naturally and also race 3-6 IMs per year. Those in themselves count as long days.

---

------------------------------------------------------------------------------------------------
New Training/Racing Log - http://www.earthdaykid.com/blog --- Old Training/Racing Log - http://colinlaughery.blogspot.com

I did it all bass akwards, and set my all time marathon PR back in the 90's, a month after running 88 miles in a 24 hour race. Then 20 years later, I ran my masters marathon PR in early 2016, after training for and racing a 50K, in late 2015, with a handful of 30+ mile training runs. That masters PR was set in the middle of training for another upcoming 50K, and a 12 hour race. In both cases, I had thrown out the canned plans, and was winging it for racing ultra distances. In both cases I had not specifically targeted the marathon as an A race... but entered them because I had the fitness to do ok. And in both cases, the PR's were an unexpected but very nice surprise. On the flip side, in the past, when I tried to PR off canned plans, I frequently blew up in the race, or got injured before the starting gun.

---

Athlinks / Strava

Six

I’m also increasingly of this school of thought, that sustained weekly mileage and tempo pace workouts are more critical for performance than lengthening your longest runs — especially once you’ve raced a few marathons and you know what the last 8 miles will feel like. Going long past a certain point detracts from your ability to do high weekly volume, which can be counterproductive (ymmv of course). I do a lot of runs of around 10 miles during peak training.