Thx @Mañ'O'Green and @Lil Dab !
Yeah - the biggest challenge for me as a new grower was 'cutting through all the noise' (as I'm sure it is for all new growers). I love me some scientific method, I agree with the scientific method, but...I'm an engineer. So, I'm usually more concerned about the cost (i.e., effort) / benefit ratio between options. And, I knew there would be variance no matter how much of the grow I tried to control and/or optimize.
I forgot to explain the 2 x std dev numbers above, which is the (nominal) 65% confidence interval based on the variance in the measured results. It should mean that I have a 65% chance of being within the (average +/- 2 x standard deviations) for other grows under the same range of conditions.
For example, using those last numbers for Grows #3 and #5 - 65% of other grows should yield 7.6 +/- 2.7 oz (=7.6 oz +/- 35%) for this strain and feed formula. As an illustration - that means I'd need whatever future 'fancy new expensive supplement everyone's talking about' to improve my yield by 35% or more in order to convince me that it's worth it and not just noise in the data.
The solution is to reduce / confirm the variance through a combination of 1) additional trials and 2) stricter control over more of the grow parameters.
My overall pipe dream would be to farm out #1 to the community and aggregate results in order to avoid #2. May still have too much variance for most things, but could maybe be used to at least get representative per-strain numbers for the breeders on the site, etc.
Perchance to dream - ay, there's the rub...
Yeah - the biggest challenge for me as a new grower was 'cutting through all the noise' (as I'm sure it is for all new growers). I love me some scientific method, I agree with the scientific method, but...I'm an engineer. So, I'm usually more concerned about the cost (i.e., effort) / benefit ratio between options. And, I knew there would be variance no matter how much of the grow I tried to control and/or optimize.
I forgot to explain the 2 x std dev numbers above, which is the (nominal) 65% confidence interval based on the variance in the measured results. It should mean that I have a 65% chance of being within the (average +/- 2 x standard deviations) for other grows under the same range of conditions.
For example, using those last numbers for Grows #3 and #5 - 65% of other grows should yield 7.6 +/- 2.7 oz (=7.6 oz +/- 35%) for this strain and feed formula. As an illustration - that means I'd need whatever future 'fancy new expensive supplement everyone's talking about' to improve my yield by 35% or more in order to convince me that it's worth it and not just noise in the data.
The solution is to reduce / confirm the variance through a combination of 1) additional trials and 2) stricter control over more of the grow parameters.
My overall pipe dream would be to farm out #1 to the community and aggregate results in order to avoid #2. May still have too much variance for most things, but could maybe be used to at least get representative per-strain numbers for the breeders on the site, etc.
Perchance to dream - ay, there's the rub...