6 ms·
Estimation of one's TDEE is pretty simple. Once you have a starting point, adjustment is common sense. If you're not losing weight at the rate you expect, lower
by lrm242 12y ago
Estimation of one's TDEE is pretty simple. Once you have a starting point, adjustment is common sense. If you're not losing weight at the rate you expect, lower your TDEE and thus lower your caloric budget. Now, adding cardio into the mix is where things can go off the rails. One reason cardio can be BAD for a diet is precisely because it is hard to measure how many calories you burn while exercising. Estimate too low and you end up in much too deep of a deficit to maintain the diet. Estimate too high and you're not losing as much weight as you want. These problems are easily solved over a long run if you can accurately do approximately the same amount cardio each time you do it, but this is also quite hard.
Cardio is great for heart health and overall wellness. It's crap for losing weight effectively, IMO. The absolute easiest way to lose weight is to simply estimate your TDEE requirements using a sedentary activity level, eat at a 3,500 calorie deficit each week, preferably at 500 calories per day or cycled if you are lifting weights, and adjust accordingly.
- yummyfajitas 12y agoHere is another stable algorithm for losing weight, which is fully compatible with cardio. Choose a desired rate of weight loss and a measurement period. Each period, measure your inputs and do the following: IF (weight_loss_rate < desired_rate) input -= 200 cals ELSE input += 50 cals END After following this algorithm, your caloric intake will converge to the region [optimal rate - 150, optimal rate + 50]. This is true even accounting for noise in the measurements - that just smears out the -150 and +50 a bit. The nice thing about bodyweight is that it's a stable monotonic dynamical system. Noise and uncertainty don't hurt you much provided you build robust controls.