For Christmas 2021, I bought myself a big PR at CIM. It’s my 5th in-person marathon where I finished with 3:34:13, taking 40m37s off LA Marathon 2020. It’s a significant improvement that was the result of a lot of things done right, including training, race planning, and execution. Here are my notes.
Training and Nutrition
We all claim that we “train” for marathons, but this is the first race where I seriously adhered to a schedule. I emerged from the pandemic putzing along at what I could get away with which was 2 group runs (with a run club) per week between 3-6 miles, plus a 13+ mile fasted long run. This often added up to 20 miles per week, but was inconsistent.
However, I had much more information going into CIM than in all of my other races. I had an updated metabolic cart that showed a significant improvement, and perhaps for the first time in my running career I truly knew what heart rate to race at. I knew how much fat and muscle I had in my body, and I knew my resting metabolic rate. Zach Bitter Coaching helped me put all the pieces together into something cohesive.
Fortunately, my job is pretty much as sedentary as I want it to be, and I allowed myself plenty of time to train and rest. Having to help organize a run club was much more of a distraction, since I often could not run at target pace and the post-run festivities (beer) were less-than-ideal nutrition. I had picked up some weight over the pandemic and did a few weeks of strict carnivore in the lead-up to two conferences over the summer. The idea was to lose significant body fat before the race, but it was extremely difficult to combine that with proper nutrition and fueling for the training cycle. I hovered around 166-168 lbs, which was about 5 lbs heavier than when I did my first marathon in 2018.
DTLA Running is my local club.
The actual low carb / keto training mechanics were fairly simple. I tried to do most of my runs fasted so I can maximize aerobic capacity. The two relevant heart rate numbers are AeT (aerobic threshold) where the body transitions from burning all fat to mobilizing some carbs (glycogen); AT (anaerobic threshold, sometimes AnT) where the body can no longer clear the amount of lactate it produces via anaerobic respiration, and beyond which the effort is unsustainable. According to the test, my AeT was a humble 135 bpm (56% VO2Max) but my AT is a whopping 173 bpm (92% VO2Max). I decided that 160 bpm would be my race pace, since HR is always a little high on race day. For all of the “race pace” runs, I targeted 160. For a more detailed discussion, see this post.
For short runs, less glycogen is mobilized so being in a fasted state wasn’t so important, but for the long runs I tried to go into those in a very low carbohydrate ketogenic (VLCKD) state 48 hours prior. This is something I have done before, but I did notice that my training runs felt more uncomfortable and “glitchy” even when at the same intensity as prior runs. This was a sign that the fasting was working and that I was getting used to glycogen depletion. I also experimented with a variety of paces during these long runs; instead of running the entire thing at 150 bpm, I accelerated to race pace after a 7-mile “warmup” (at AeT) and ran until I got tired, and then proceeded at a moderate pace. It was because of this that none of my long runs were truly “comfortable,” it was something I hadn’t tried before, and I would not find out how well it worked until race day.
Since I didn’t jump into serious training until about 7 weeks out, I had a limited amount of time to do speed and strength work. I did resistance bands twice a week, and most speed work turned out to be running at close to race pace since it was already so close to anaerobic threshold. I relied on the usual Downtown LA terrain for hills and had a few track sessions and tempo runs with full-body HIIT workouts.
It turned out that the biggest challenge wasn’t speed or long runs, but to fill the miles in with a bunch of AeT runs at 135 bpm which felt like death marches in molasses. At this pace, running is inefficient and the biomechanics were never quite right. There were plenty of times the training felt wrong and I had to put some blind faith into the program.
The 40+ mile weeks took a toll and there were some sporadic aches and pains. In particular, my left knee was rather sore during the high-mileage weeks, but it would only manifest when I had those slow AeT runs. I had no major issues that prevented me from running and fall is a good time to train in Southern California. The taper period felt just right and by race week I had a feeling that I was ready to fly.
Pre-Race
There were lots of options for hotels in Sacramento, but at the end I chose to drive to an Airbnb that was a few miles from the finish because it looked quiet. Just to ensure a stress-free experience, I booked 3 nights so I can get all the kinks worked out prior to race day, and not worry about checking out until the day after.
CIM had made a big deal about signing up for an expo session, but I missed that email and none of it was enforced anyway. I walked around the expo for a few hours, and my initial impression was “wow, everyone here looks really fast!” Based on my years of looking at runners, I had never seen a faster cohort. Nobody here planned to walk this race. It was intimidating to think that there’s a possibility I might embarrass myself somehow in this field.
For food, I chose options that were high protein, medium carbs, and low fiber at chain restaurants that offered consistent and repeatable experiences. This is for a few specific reasons other than no new things on race day. First was to have things I know I tolerate well. Second was to reduce the amount of poop generated and not induce inflammation. For example, fried items that contain seed oils cause acute artery hardening up to 24 hours after ingestion, which is not what you need for maximal blood flow. I’m honestly not convinced of the merits of carbing up before the race but I did it anyway.
Dinner the night before, at 4p. It was a 20 oz. cut of prime rib with rice, green beans, and some bread.
Much of Saturday night was spent amateurishly agonizing over the race kit. I knew it was going to be a cold day so the dumpy outer layers were to be tossed, but it was what to carry on course that made me think and rethink. In hindsight, I probably could have forgone the electrolyte as there was plenty on course that was much more than was needed for the cool temps, and keeping the thin outer layer was a good call.
The race buses leave downtown at 5 am and I found out the night before that light rail didn’t start running on Sunday morning until after 5a which meant I would miss the bus. I thought about driving to downtown but discovered that a Lyft would only cost $10 and I could pre-schedule it, so I did that instead.
I was still not used to the features of the Lululemon shorts, so I ran with the Flipbelt.
Race Day
I barely slept all night, and it was a relief to hear the alarm go off at 3:30. It was the earliest I ever got up for a race because of having to get to the shuttle. My Airbnb host had coffee set, and I read with interest an NYT article about the nexus between coffee and poop and sure enough I had a regular bowel movement after the coffee. I tinkered with the race kit a bit more but ended up having plenty of time before the Lyft showed up at about 4:40 and I was on my way.
It was also the shoe show. I’ve also never seen such a collection of expensive shoes.
Of course our bus driver got lost and it took almost an hour to get to the start. I tried to drift back to sleep but the adrenaline was raging, so I listened to the people talk around me. I had gotten into a more conversational mood and was surrounded by people who talked a fast game so it was nice to pick their brains for a bit as well. After we arrived, we were free to stay on the bus to stay warm, so we continued to talk. One person was an elite runner. I asked to take a photo of her bib and discovered that elites don’t have numbers.
I had a conundrum when I approached the start. To the right was the corral I was supposed to be in, a 3:30-3:44 (Based on best-estimated race pace, I was 95% confident to finish faster than 3:45). But that corral was full all the way through about 4+ hours and with minutes remaining before start I wasn’t sure if I wanted to push through the crowd. To the left was the 3:15-3:29 corral which was faster than I wanted to go, but given my race experience, I knew that people tend to overestimate their corral placements when they are not enforced. The effect of being stuck in a slower corral is that you will never run with the right pace (you are always passing a slower cohort and the people at your pace had a head start and therefore will always be ahead of you). It takes more energy to pass people, and I had high hopes for a significant PR, and so I went left.
Gun time was 7:00a, and I crossed the start a few seconds after 7:05. I ran with the crowd and we went at about a 7:50 pace for the first 6 miles. The first thing I noticed was that the course was significantly hillier than I anticipated. It was advertised as downhill but it is a net downhill while the first half was rolling hills. Given this, I could not depend on instantaneous HR for pacing feedback (I was running to heart rate, not pace). The Fitbit wasn’t really accurate even for that, so it was next to useless. Instead I listened for the per mile average that my tracker was reading back to me and I decided at around Mile 6 to dial back my effort because HR had already pushed to 165. After I did that, I was noticeably passed by a bunch of people.
I really resonated with her shirt, which read: “The No Luck Club; hard work and effort only.” Associating race outcome with luck is an implicit denial of the hard work, experience, and preparation required to get here.
You’d think that people would be focused on running during a race like CIM, but there were still some social moments. One person rolled up and commented that I looked like I knew what I was doing, so we talked for a few minutes where I sold him on the merits of a run club. Another person had a shirt that I liked so I asked to take a photo. At least I have plenty of practice taking photos while on my feet. I also had to make adjustments to take off, fold up, and hang my outer layer around the waistband, which I did without slowing down. It turns out that you have a lot of time during a marathon for uncomplicated tasks.
I was successful at keeping my HR somewhat within planned limits (in the low 160s) all the way to the halfway mark. A huge crowd greeted the runners here. One would never expect this many people lining the course and cheering us on for a “small local race.” I saw quite a few spectators who had Boston Athletic Association shirts or jackets on. It was probably the most Boston race gear that I’ve ever seen. Soon after this, I gave up on HR monitoring, and instead relied on perceived effort. HR crept up to the 170s range. I had raced Pasadena Half at AnT, but not a full marathon, so I was a little apprehensive about getting up there, but I felt fine and continued on.
Nutrition-wise, I had a banana while chatting on the bus which was probably around 6a. I had my first gel soon after the halfway mark, a second gel at about Mile 18, and a third at Mile 22. In each case, I had grabbed one at the previous station and watched for the next water stop. At the first glimpse of one (you can tell by the large blue tanks used by volunteers to fill), I would start on a gel, which I can usually finish by the time I arrived at the water to flush it down. I was careful to not drink to excess, only to quench apparent thirst and perhaps even a little less than that.
Photo credit @jenlabrie
While I was passed by quite a few people earlier on, I saw people walking as early as shortly after halfway; around Mile 18 or so the tables were turned and I began to pass people in earnest. Some were going slow and some paced me but stopped for water and were never seen again. I didn’t need to pee and my legs were on autopilot. After about Mile 22 I got that familiar feeling that I couldn’t do much better than what I’m already doing (also known as the “when will this be over” signal). I had a little bit of sciatica (which I affectionately call a “butt cramp”) and took a few swigs from my electrolyte jar, but the combination of gel and electrolyte was pretty gross, so I kept that to a minimum. The pain would recur one more time but it was not debilitating, nor was it anywhere as bad as my second marathon in Carlsbad 3 years ago.
I finished the last 6 miles of the race at pretty much exactly my anaerobic threshold. I wouldn’t say it was painful, but I literally could not even sprint the last 100 yards and I was glad it was over.
I had little physical symptoms other than trembling and weakness in the knees. I grabbed a bit of food and sat down by the curb for about 15 or 20 minutes. It was a great time to drink about 1L of water and let the body sort itself out. Other than being underdressed for the cold, I didn’t have any major trouble moving about or any digestive distress. I ran into quite a few friends at the finish line and should have been better at taking photos.
Race Analysis
I paced really well. Sure, the PR was nice, but I ran in a faster corral than I intended and passed more people than passed me. Overall, I was 30 runners ahead at the finish than at the 5k timing mat, and 21 ahead of my division. From the chart, I was the most “behind” at about the halfway mark, passed 81 people at 30k and another 198 people by 40k. Had I been in the “right” corral, this effect would have been more dramatic. This is consistent with my on-course observation that even among fairly seasoned marathoners capable of pulling off a 3:30 finish, the late course “bonking,” caused by an over-reliance on carbs, continues to bedevil many.
Did I really need the carbs? This is a question I’m really interested in. Other than LA Marathon 2020 where I had one gel around Mile 18, this is the first time I’ve introduced significant calories on race day (beyond the coffee with cream I have every morning). If you are not familiar with metabolic cart testing and AeT/AT terminology, see here for a discussion.
The conventional advice is to take as much as you can tolerate, or 25-60g of simple carbs per hour. I spent 3.5 hours on the road and consumed about 100g of carbs inclusive of the pre-race banana which would put me on the low range of the advice. The problem is that the data points are too few and each marathon is different, so there’s not really a scientific way to find out whether the carbs helped from race to race, and I’m loathe to test carbs on training runs because that’s my opportunity to train the fat-burning pathway.
For a rough estimate, the metabolic cart again offers some insight. Recall that at AeT the body begins to utilize at least some glycogen and at AT any additional performance comes from fermentation of only carbs. In the results, a “calories per hour” value is extrapolated from the oxygen consumed, which for me is 616 at AeT and 928 at AT. The body doesn’t switch off fat burning when it is above AeT so it is probably unreasonable to do a simple subtraction and conclude that 312 additional calories must be consumed per hour. For me, these numbers increased by 149 more calories at AT while AeT increased by 102 (compared to 2018), which means that my body got better at burning about 100 calories more fat per hour. Since the body stores so much more fat than it could reasonably burn in a race, it would seem that there is only a 47 kcal/hr gap that could be filled with carbs (half a gel per hour).
With endurance events, the whole idea is glycogen defense. The first line of that defense is always to maximize and optimize the fat-burning pathways by raising AeT. Carbohydrates can be strategically used to fill the hole up to anaerobic threshold, beyond which any additional increase in fueling or effort is futile.
Bad Advice
“Drink before you’re thirsty.” This might make some sense during hot weather or poorly supported races that have hydration stations spaced far apart. With exception of my first marathon, I’ve never ran a race where I couldn’t wait until the next station, and in a cold-weather race like CIM, drinking more just begats the potty. I usually drink just enough to quench thirst and pour the water over my head to cool off. This is the one time in your life where you can litter like a pro (and in fact, you paid for that privilege).
The person in the green shirt was pretty fast. She would stop to get fluids, come from behind to pass me, and I’d see her at the next water stop. I finally ditched her by not having to stop at all.
“Stopping to rest.” I drove the 400 miles from L.A. to Sacramento without stopping. Other than testing the strength of my bladder, I took a page from my younger days of road-tripping — the way to cover more distance in less time is not to drive faster but to stop less. The way to stop less is to have fewer or even no bathroom breaks, by tweaking your digestion and limiting fluid intake. Many runners execute “nuisance stops,” where they take fluids and gels but in reality take a break. It is probably ideal to maintaining a running gait even when hydrating (and you should probably not eat anything that requires chewing). My experience has been that in the last stretches of the race, walking is an evolutionary signal to the brain that it is out of immediate danger, and and the urge to stop will become overwhelming.
What’s Next
Welcome to running purgatory, where I’m fast enough to not suck and not fast enough to qualify for anything. LA Marathon (March 2022) is my next race but I opted to not train for it. I think I am well-equipped by now to finish a marathon without fear that I would suffer a major breakdown, and LA is not the type of race that I would likely PR at anyway.
Another reason why I decided to forgo another training cycle is that there is virtually no more room for improvement at my current body composition. In September 2021, Pasadena Half was a PR at 1:43. However this was only about 100 seconds faster than the PR two years prior at America’s Finest City. Since my AT is already 92% of VO2Max and I have no complications running at the threshold, there are few things left in the “run a better race” basket to tweak.
Another sign that you might have run up against your natural speed limit is if your marathon pace is fairly close to the half marathon pace. If the paces are far off, it’s a sign that the marathon was mismanaged, the training wasn’t adequate, or there was bonking, all problems that can be easily solved. I always knew that my theoretical marathon time should have been significantly faster, but there had been no races during the pandemic to test until now.
Photo Credit Flashframe (CIM Race Photography)
I do not have a stereotypical runner’s body. Eventually, most recreational runners are confronted with the eternal question of having to reduce body fat percentage. I can do more speed work and increase mileage which would help increase VO2Max on absolute terms, but cutting excess body fat is probably easiest. It appears that the potential is significant. My latest DXA scan (days after CIM) showed that I carry about 23% body fat which can probably be cut 10% just by being stricter with the diet. That would mean a 17 lb weight loss which would take me down to about 150 lbs.
While research shows that each pound lost translates to about a 2-second pace improvement, there are reasons to believe that this metric is too obtuse - most people lose muscle along with fat because of caloric restriction which means they also lose overall strength. I plan to continue with my data-driven approach to running, and to apply that to ways to reduce BF%.
All photo credits are mine unless otherwise specified. Cover photo by Lawrence Wong.