trickybeck wrote:
TirithRR wrote:
trickybeck wrote:
Add high school level math to the long list of things you are terrible at.
Fun to hate on Gbaji and all, but what he's saying is not false. Units make no difference in a person's inability to grasp percent increases (or decreases).
We're talking about the reciprocal ratio, not the unit swap.
No. Smash was arguing that people perceive linear differences in one value as "the same" regardless of the effect on the actual ratio. So the difference between 15MPG and 20MPG versus 35MPG and 40MPG are going to be seen by consumers as the same because they're all a 5MPG difference even though the relative improvement is much better in the first case rather than the second.
The problem is that, assuming we accept his premise (and I'm willing to for the sake of argument), we run into the same problem either way (well, except going in the opposite direction). If we assume that someone will perceive a 5MPG improvement as equally valuable, then they'll also perceive the same linear improvement in fuel consumption over distance as equally valuable. But this is where his argument goes off the rails:
If we take the same values, but flip them around, then we have one case (15MPG to 20MPG) where we go from 6.66G/100M to 5G/100M. That's a 1.66G/100M improvement. But in the second case (35MPG to 40MPG) we go from 2.86G/100M to 2.5G/100M. That's only a .36G/100M improvement. Put another way, assuming Smash's premise is correct, in order for people to have as much perceived fuel efficiency value in the second case, we'd need to be able to drop the fuel usage rate from 2.86G/100M to 1.2G/100M (cause that's 1.6G/100M improvement, right?). That's a problem though, since that would require a much bigger actual fuel improvement increase.
I agree with what Smash started to say, I just disagree with the direction he was going with it. If your goal is to encourage people to keep buying more fuel efficient cars, you want to use a measurement that makes even relatively small improvements appear large enough to be worth it. The linear fallacy angle applies in both cases, but IMO it creates a disincentive for consumers to buy more fuel efficient cars as the fuel efficiency increases if you represent it as G/100M versus MPG. It's relatively easy and cheap to increase relative fuel efficiency when you first start caring about such things, but it becomes increasingly difficult and costly as you improve. We have to accept that we can only get marginal increases over time going forward, but if we assume that every bit helps, then we want to express those small improvements in a way that actually makes people think they're getting something more.
MPG is a better way of doing this precisely because of the false perception Smash was talking about. You want them to continue to place similar value on incremental fuel improvements as fuel efficiency increases as they did when fuel efficiency was relatively low. MPG does this quite well if we assume the perception problem Smash talked about. And if we don't, then it doesn't matter since both represent a ratio. As Smash correctly pointed out, a linear increase in MPG results in successively smaller ratio improvements as fuel efficiency improves, while the opposite is true with G/100M (or the same in metric). Where I think he's wrong is which is actually 'better'.
Edited, Sep 13th 2013 7:05pm by gbaji