Choosing Irons - Brand X vs Brand Y

[fixyurdivot]

1 hour ago, TitleistMike said:

I hear you. I have a hard time getting business leaders to understand variation, std deviation is more important than avg when evaluating alternate solutions. Statistical Test of Hypothesis is the technical term.

 

Oversimplified, if the distance difference is within about 1 std dev, the difference isn't real. It is random variability.  If you did 100 shots with each club and had a gap near 1 std dev....might be real.

 

Genichi Taguchi would have a field day with many of the golf equipment tests we see. 

[Everardo]

On 7/18/2022 at 4:04 PM, TitleistMike said:

Humans want to see differences that are NOT real.  Avg is MEANINGLESS.

They are SAME club effect result...especially with only 10. Give  me 200 or 300 data points and take another look.

Sure but in a fitting a person is not going to take 200-300 shots per club to get data, and taking that many shots, you'll just adapt to the club and it's differences. I wouldn't call taking an average meaningless, we're not robots, so we have to take into consideration best, worst, and median shots. 

[Everardo]

19 hours ago, TitleistMike said:

I hear you. I have a hard time getting business leaders to understand variation, std deviation is more important than avg when evaluating alternate solutions. Statistical Test of Hypothesis is the technical term.

 

Oversimplified, if the distance difference is within about 1 std dev, the difference isn't real. It is random variability.  If you did 100 shots with each club and had a gap near 1 std dev....might be real.

 

std dev is something Foresight and Trackman have added as a way to try and help find differences between a series of shots. Average is still probably the better measurement with that few of shots.

[TitleistMike]

On 7/20/2022 at 10:41 AM, Everardo said:

std dev is something Foresight and Trackman have added as a way to try and help find differences between a series of shots. Average is still probably the better measurement with that few of shots.

It is really deep complex statistics. Quality control people use it to validate a process change is real.

I can say a lot of people make invalid business decisions based on averages and ignoring std dev.

[Everardo]

4 minutes ago, TitleistMike said:

It is really deep complex statistics. Quality control people use it to validate a process change is real.

I can say a lot of people make invalid business decisions based on averages and ignoring std dev.

Just noticed on my latest Trackman report, they renamed it "Consistency"

 

[TitleistMike]

1 hour ago, Everardo said:

Just noticed on my latest Trackman report, they renamed it "Consistency"

 

Yes, std dev has a specific formula. Sample size and othe stuff matters.

The challenge is all this forces users and providers and clients to be subject matter experts in statistics which isn't their comfort space.

When club brand A is 3.2 yards longer than B we want to declare it better. It feels right.

 

I will add Pepsi lost a lawsuit vs Coke based on std dev.