Ran 21 Easy Runs in the Chennai Heat.
Here's what the data actually showed.

There was one thing that kept bothering me during this training block. No matter how easy the run was supposed to be, my heart rate always found a way to climb and be noisy in the majority of the runs. Every time it happened, the explanation was like ready made! Heat... Humidity... or Salty air... etc I accepted those answers for long time until I finally looked at the data.

Between April and July, I completed 21 easy runs during this training block. Garmin classified every one of them as an Aerobic Base run. I kept the pace stayed almost for every run, and the effort was meant to be easy throughout. So one question kept coming back to me. If I was running at nearly the same pace with the same training purpose, why was my average heart rate different almost every day?

Everything looks good... until the charts show the real story.

Progress snapshot: April to July

Training Background: the big picture

There is something called a training pyramid. Think of it as a quick way to see whether the foundation of your training is strong enough to support the harder sessions.

Official training block lasted 96 days, from 14 April to 18 July. Already I have done a couple of weeks of training before it started. The only break came when I had a fever and had to take three days off. Other than that, plan is followed consistently.

Strength training took the biggest share of my training time at 40%. That was something I never expected. It simply happened just like that, and each strength session lasted between one and two hours. Looking back, the easy runs often felt much harder than I imagined. There were days when even a 30 minute easy run felt tougher than a one hour strength session. Sometimes, I would already feel tired before even I started running.

Weekly training load, day by day

Each day of the week had a specific purpose in this training block. Tuesday was for speed or threshold work. Wednesday and Sunday were strength days. Thursday was for easy runs, and Saturday was reserved for a long run or a tempo session. Most TTians would already know this routine.

As the training block reached the halfway point, the workouts became more challenging each week. Coach usually shared the next week's plan on Sunday evening or Monday afternoon. Waiting for that message became a feeling that is hard to explain. There was a little nervousness because I wasn't always sure I could finish what was planned or not.

Week 1, from 20 to 26 April, marks the start of this training block. I kept the previous week only as a baseline so I could compare how the training progressed. Week 8 was different. I came down with a fever, and Coach asked me to take a few days off. I missed the Tuesday, Wednesday, and Thursday workouts, and you can clearly see the drop in training load. The weekly load fell to 1310, the lowest of the entire block. The following week, I was back to training as usual. By Week 10, the weekly load reached its highest point at 2884, helped by a 16 km tempo run on Saturday. Week 13 was the taper week. It included a 90 minute easy run on Tuesday, a shorter 5 × 800 m interval session on Wednesday, Thursday and Friday completely off, and a short 30 minute tempo run on Saturday to stay sharp before race day.

The Load Focus chart tells its own story. During the first half of the training block, most of the load came from easy runs. In the second half, full of tempo and threshold sessions.

*Note: Up to 31 May, Tuesdays were dedicated to interval/tempo workouts and Thursdays to easy aerobic runs. From 1 June onward, this pattern was reversed, with Tuesdays becoming easy runs and Thursdays becoming the primary interval/threshold workout. Charts can be interprested accordingly

Source: DRHM 2026 Performance Tracker, Daily/Weekly logging on performance metric manually .

The metric that matters: Its efficiency, not the pace

Like many runners, I always thought pace was the best way to measure fitness. If my pace got faster, I felt I was getting fitter. I spent a lot of time looking at pace, heart rate, and other charts, but hardly paid any attention to the power metrics. Even on the Garmin activity screen, the power chart appears near the bottom, making it easy to overlook. As I spent more time exploring my running data, I realized this was one of the metrics that deserved much more attention.

Efficiency Factor measures how many watts I produce for every heartbeat. In simple terms, it tells me how much running power I generate for the effort my heart is putting in.

Efficiency Factor = Average Power ÷ Average Heart RateShow the math

The higher the Efficiency Factor, the more running power I am producing for each heartbeat. I found this to be a much better way to understand how efficiently my body was working than looking at pace alone.

Efficiency Factor went from 1.536 W/bpm on my first run to 1.642 W/bpm on my most recent one. That's a 6.9% improvement in true aerobic fitness over 13 weeks of training. For a better understanding, a score below 1.4 usually means a beginner, and above 1.8 usually means a well-trained runner. I'm in the middle of the range for a regular recreational runner.

For all my easy runs, I stayed close to the same pace of around 7:20 /km throughout the training block, irrespective to the training load. Even as the weekly training load increased, I found that my body was handling the effort better.

The problem I can't explain away: Aerobic Decoupling

Here's another metric I couldn't ignore. Aerobic Decoupling(Heart Rate Drift) measures how much your heart rate changes while maintaining a steady aerobic effort. In simple terms, it compares how efficiently you ran during the second half of the run when compared to the first half, after excluding the warm up.

As a general guideline, a value below 5% suggests you were able to maintain the same aerobic effort throughout the run. Values between 5% and 8% are generally considered acceptable but may indicate there is still room for improvement. Values above 8% suggest your heart rate drifted more than expected for the same effort. My median value across the 21 easy runs was 7.1%, and only 3 of those runs stayed below the 5% mark.

"Seven of twenty-one runs crossed the 8% line - the strongest within-run fatigue signal in these sessions."

But there's an important factor to consider: 18 of my 21 runs happened at 30°C or hotter. Heat alone pushes your heart rate up during a run, no matter how fit you are. So it's hard to know how much of my decoupling is caused by heat and how much by fatigue. Probably both... The latest run (July 14) also came in at 8.4%, which suggests fatigue is still present.

So I compared each run's weighted average temperature against its average heart rate. The pattern was very clear.

Each extra degree Celsius adds roughly 3.2 beats per minute to my heart rate.Show the math The formula is HR = 3.19 × temperature + 50.7, and temperature alone explains most of the difference in heart rate between my runs. In practical terms: a run at 31.5°C puts about 8 more beats per minute of strain on my heart than a run at 29°C before fatigue, fitness, or pace are even considered.

Two runs stand out. On July 14 (30.3°C, 142 bpm), my heart rate came in about 5 bpm lower than temperature alone would predict. This was the strongest sign of adaptation in the whole dataset, right at the end of the block. On April 30 (30.2°C, 152 bpm), heart rate ran nearly 5 bpm higher than predicted, the largest positive miss, early on before the heat adaptation had set in.

This means comparing heart rate or Efficiency Factor between runs without accounting for temperature can be misleading. A heart rate of 148 bpm at 29°C is a very different effort than 148 bpm at 32°C. Adjusting Efficiency Factor for heat is the more accurate way to track progress here and even after making that adjustment, my trend still points upward.

Week by week: how my body adapted

Looking at weekly averages instead of single runs shows the pattern more clearly, good weeks followed by harder weeks, where the body seems to be recovering and rebuilding before the next gain shows up.

Weeks 2, 3, and 4 (early-to-mid May) were high fatigue weeks. Then weeks 7 and 8 showed the best results of the block. Efficiency Factor at 1.63, decoupling under 7%. Weeks 9 and 10 slipped back toward higher fatigue, weeks 11 and 12 stabilized, and week 13 moved back into fatigue risk. A recovery week is still overdue.

Intervals and threshold: going harder

Easy runs are only part of the story. Alongside them, I ran 17 interval and threshold sessions, and both the volume and intensity of these sessions have been increasing.

A deeper, rep-level diagnostic for these sessions is in progress.Show details It's a composite score that normalizes power, cardiac drift, recovery drop, and pace on a 0–1 scale, so sessions can be compared apples-to-apples. Validated on one session so far; running it across all 17 is still to come. Methodology so far: the notebook is on GitHub.

Strengthening, and whether it helped

Alongside the running, I completed 24 strength sessions, averaging about two sessions each week. Most of them laster for atleast 1.5 hours.

One limitation is that all my strength sessions were bodyweight exercises. Since I wasn't lifting weights, it was hard to measure or track. Even so, the average recovery cost was 59.4 out of 100, which is in the moderate range, and none of the sessions reached a high fatigue level. Overall, the strength sessions worked well alongside my running.

The body itself: composition over 10 weeks

Running and bodyweight training changed more than just my weight. My face became noticeably leaner, but the weighing scale told only part of the story. I lost 2.70 kg during this training block. The real question was whether I had lost fat, muscle, or a bit of both. That's where the body composition data became much more interesting.

Across six body composition measurements, the changes followed a clear pattern. One thing I kept hearing was, "Too much running will make you skinny," or "You'll lose muscle". After all, its only running around 42 km weekly mileage max along with bodyweight strength training. So I wanted to know what I was actually losing. The body composition data answered that question. Body fat steadily came down, while skeletal muscle, muscle rate, protein mass, and body water all increased. Looking at these changes together, it was encouraging to see that my running, bodyweight strength training, nutrition, and recovery were working well together.

Protein mass is now at 15.40 kg (+0.60 kg from baseline), and body water at 56.30% (+1.90%). Both are strong indicators of muscle quality and hydration status. The skeletal muscle increase of +1.00 kg from baseline shows sustained, functional muscle development. Body fat at 23.10% (−2.00% from 25.10%) represents the cleanest recomposition: fat loss paired with lean mass gain.

One thing to keep in mind is that these are only six body composition measurements, not years of data. Home body composition scales are not perfectly accurate to be considered, so small changes from one week to the next should not be overinterpreted. keep it light and low

Most interesting findings from wellness and recovery signals

1. Resting HR improved meaningfully across the period

2. Stress spiked in June, then eased through early July

June 9 recorded the highest stress levels of the entire training block (average 82, maximum 97), despite almost no activity (608 steps, 0.48 km). This occurred during Week 8, when illness forced me to stop training, illustrating how physiological stress remained elevated even as physical activity dropped to its lowest level.

An expanded pull of the stress dataset (94 daily readings, April 16 through July 18, the full window covered by the raw Garmin export) confirms the June spike was a peak, not a new normal: July has averaged 35.4, back below June's 43.9 and closer to the April–May range. Across the full window, mean stress sits at 33.7 (median 32), with 27.7% of days landing in the high-stress band and 25.5% landing low, a wide spread, but one that's trending back down.

3. A very strong recovery block happened in mid-May

The lowest resting heart rates were recorded between May 14 and May 19, reaching a minimum of 44 bpm. HRV data were available for only 11 days, as I did not wear my watch during sleep on most nights. During the period with available data, HRV averaged 62.5 ms and peaked at 69 ms. Combined with the sustained low resting heart rate, this suggests a brief period of excellent recovery and physiological readiness.

4. High-load days often line up with higher stress strain - but not always

Big activity days included June 18, June 20, June 27, and July 4, with top-day volume around 18 to 19.5 km. June 20 is the clearest pairing of the two signals: high volume alongside Medium-band stress (average 65).

Breaking the 101 days into phases

These 101 days were not one smooth trend. They break into clear phases, where training load, sleep, stress, and Body Battery each tell a different part of the story.

One thing worth explaining before the charts: this section leans on peak Body Battery rather than the most-recent reading. Garmin's snapshot has 100s of parameters, out of which for Body Battery, there are four ways to look at recovery for a given day, and each answers a slightly different question. The daily peak Body Battery is better reflects how completely my body recovered overnight, making it a more meaningful measure of recovery across the training block.

Switching to daily peak Body Battery (the highest reading logged each day, usually reached overnight) changes the picture. The relationship is still inverse, but much weaker.Show the fit The fitted line is Peak Body Battery = -0.19 × Stress + 85.0, with a correlation of only r = −0.23 (p = 0.03) across the 94 days with complete data. That makes sense... A day's peak is set mostly by how well the previous night's recovery went, so it is far less sensitive to that day's stress than a same-day "most recent" reading is.

Data quality caveats

The most reliable health trend window is effectively April 16 to July 18, 2026, for resting heart rate, stress, and Body Battery. based on a single consolidated pull from the raw Garmin export covering all 94 days in that span, and one that also confirms a corrected reading for July 4 (originally logged as 80, actually 34). HRV and full sleep-stage data are the exception: they're only present for a short subset (11 days, mostly in May and early June), so long-range conclusions for HRV and sleep staging should still be treated as directional rather than definitive.

The People Behind the Numbers

Over 13 weeks and 79.6 hours of training, one thing became clear. My aerobic fitness improved. Efficiency Factor increased from 1.536 to 1.642 W/bpm, and my resting heart rate came down by nearly 6 beats per minute. These are not huge changes, but they show what consistent training can do over time.

The journey was not perfect. I had a fever in June, stress levels went up, and recovery was not always smooth. The charts show both the good days and the difficult ones. Looking back, I'm glad they do, because that's what a real training block looks like.

None of this happened by myself. TTK Coach Gokul Prasad PB planned every stage of this training block, from the easy weeks to the harder sessions and the taper before race day. Special thanks to Subbu, Jo, Vasi, Vinodh Kumar, Murugesan, Sriram, Dominic, Sridhar, Premi, Lalitha, and Nandha for the encouragement and support throughout the journey. I am grateful to be part of the Towertwisters Chapter. Those early morning runs, conversations, and shared miles made it much easier to keep showing up, even on the days when I didn't feel like running.

This training block had one goal: DRHM 2026. The training is done. Now it's time to trust the work that has already been put in. Whatever happens on race day, this journey has already taught me more than I expected.

Tools & equipment

This research was conducted using a small stack of consumer hardware and Garmin's developer APIs.