burbank Jung said:
I understand those who are talking about statistics but…
There’s no “but”. The statistics still apply. The only question is, do the stats reveal that you are able to learn enough for your purposes from the shortcut data or not.
…I take short cuts for the time and money I'm trying to save. You might say it's a waste of time but I'll take the risk.
There’s no need for unknown risk. The stats can tell you exactly how much risk you are taking of not learning what you want to know. You have to define it, though.
Just as a faint A.M. radio signal is harder to make out when it is small relative to the static noise that comes with it, what you are trying to discern with a ladder test is the signal (the mean value, aka average value), from the noise (the scattering of shots around the mean point of impact or the scattering of velocities around the mean value or both). How precisely you can make that determination depends on how big the noise is, and upon how much averaging your shot count will let you do to filter out the noise.
In terms of shot placement, you can imagine that if you fired an unlimited number of rounds and the gun never wore down in the process, as your shot count approached infinity, the average result would give you an increasingly exact idea of where your true average point of impact is and how much scatter in location (the noise) will mix with it and tend to displace any particular shot. In turn, this tells you how likely you are to land a shot within a particular range of locations.
As to your current data, if you have enough shots in your ladder, you may be able to learn more than simple eyeballing tells you by separately finding the vertical distance of each shot from the point of aim and performing a 4th order polynomial curve fit to it in Excel (or other spreadsheet program like the
Apache Open Office spreadsheet, Calc). This will perform the least-squares fit to the data, which tends to find its trend. The 4th order fit has enough bend directions to find a flat spot. If that doesn’t work, try generating a running average of two and evaluating that result so shot one is averaged with shot two, shot two with shot three, shot three with shot four, etc. If that still doesn’t help, try a three-shot running average, so your data points are the average of shots one, two, and three, and then two, three, and four, and then three, four, and five, etc.
If that still doesn’t bring clarity, you may need to shoot more shots to know what you want to know. How many more can be determined how far apart the shots are on average. I would need to see your data to help you get any more out of it or to determine you can’t know what you want from that little data.
Creighton Audette used single shots of each charge weight in his ladders successfully. You just have to remember the trend of all the shots is being considered, so it isn’t like trying to get data from one shot. On the other hand, he was a champion benchrest competitor and knew how to load to minimize standard deviation and keep conditions for each shot consistent. Not everybody can do that.