I said I was going to check out, but now there’s an entire new post with claims about my predictions and views. So I’m going to restate my predictions and views to avoid misrepresentation; interpret my last comment as merely checking of the previous discussion; interpret this comment as checking out of the new discussion; and try to keep myself to restating my views rather than saying anything new.
Clarifying predictions:
I said that the single-hose AC will be 25-30% less efficient when cooling from 90 to 80 degrees. That would mean the 2-hose AC has a temperature difference 33-43% larger. If there is a larger temp difference (and especially a higher outdoor temp) you will see a larger efficiency gap. You’d get to 50% if you were instead cooling from 97 to 80 degrees, or from 80 to 60 degrees (though that looks impossible in your test). These differences are measured for the 1-hose cooling.
Clarifying my calculation:
I estimated the efficiency loss at (outside temp—inside temp) / (exhaust temp—inside temp).
This was for constant exhaust temp. I agree that higher exhaust would would make a single-hose AC more inefficient, but it doesn’t seem related to Goodhart, since: (i) it is reflected in the BTU rating on the box, (ii) it is reflected in the temperature of air leaving the unit. So I don’t see how a consumer could be making a Goodhart-related mistake based on this type of inefficiency. It’s also not related to this experiment since it won’t be changed by adding a cardboard tube.
The reasoning was: at constant exhaust temp, each AC pays the same amount per unit of heat moved from the cool air to the hot air. But for each unit of air leaving the 1-hose unit, you have infiltration of the same quantity of hot air from the outside. This undoes some fraction of the work the AC did. What fraction? Well, each unit of exhaust moves C*(exhaust temp—inside temp) of heat outside. And each unit of infiltrated air brings C*(outside temp—inside temp) back. So you lose (outside temp—inside temp) / (exhaust temp—inside temp).
Reposting my final comment, which was left out of the summary:
I still the 25-30% estimate in my original post was basically correct. I think the typical SACC adjustment for single-hose air conditioners ends up being 15%, not 25-30%. I agree this adjustment is based on generous assumptions (5.4 degrees of cooling whereas 10 seems like a more reasonable estimate). If you correct for that, you seem to get to more like 25-30%. The Goodhart effect is much smaller than this 25-30%, I still think 10% is plausible.
I admit that in total I’ve spent significantly more than 1.5 hours researching air conditioners :) So I’m planning to check out now. If you want to post something else, you are welcome to have the last word.
SACC for 1-hose AC seems to be 15% lower than similar 2-hose models, not 25-30%:
This site argues for 2-hose ACs being better than 1-hose ACs and cites SACC being 15% lower.
This site does a comparison of some unspecified pair of ACs and gets 10⁄11.6 = 14% reduction.
I agree the DOE estimate is too generous to 1-hose AC, though I think it’s <2x:
The SACC adjustment assumes 5.4 degrees of cooling on average, just as you say. I’d guess the real average use case, weighted by importance, is closer to 10 degrees of cooling. I’m skeptical the number is >10—e.g. 95 degree heat is quite rare in the US, and if it’s really hot you will be using real AC not a cheap portable AC (you can’t really cool most rooms from 95->80 with these Acs, so those can’t really be very common). Overall the DOE methodology seems basically reasonable up to a few degrees of error.
Still looks similar to my initial estimate:
I’d bet that the simple formula I suggested was close to correct. Apparently 85->80 degrees gives you 15% lower efficiency (11% is the output from my formula). 90->80 would be 20% on my formula but may be more like 30% (e.g. if the gap was explained by me overestimating exhaust temp).
So that seems like it’s basically still lining up with the 25-30% I suggested initially, and it’s for basically the same reasons. The main thing I think was wrong was me saying “see stats” when it was kind of coincidental that the top rated AC you linked was very inefficient in addition to having a single hose (or something, I don’t remember what happened).
The Goodhart effect would be significantly smaller than that:
I think people primarily estimate AC effectiveness by how cool it makes them and the room, not how cool the air coming out of the AC is.
The DOE thinks (and I’m inclined to believe) that most of the air that’s pulled in is coming through the window and so heats the room with the AC.
Other rooms in the house will generally be warmer than the room being air conditioned, so infiltration from them would still warm the room (and to the extent it doesn’t, people do still care more about the AC’d room).
I said I was going to check out, but now there’s an entire new post with claims about my predictions and views. So I’m going to restate my predictions and views to avoid misrepresentation; interpret my last comment as merely checking of the previous discussion; interpret this comment as checking out of the new discussion; and try to keep myself to restating my views rather than saying anything new.
Clarifying predictions:
I said that the single-hose AC will be 25-30% less efficient when cooling from 90 to 80 degrees. That would mean the 2-hose AC has a temperature difference 33-43% larger. If there is a larger temp difference (and especially a higher outdoor temp) you will see a larger efficiency gap. You’d get to 50% if you were instead cooling from 97 to 80 degrees, or from 80 to 60 degrees (though that looks impossible in your test). These differences are measured for the 1-hose cooling.
Clarifying my calculation:
I estimated the efficiency loss at (outside temp—inside temp) / (exhaust temp—inside temp).
This was for constant exhaust temp. I agree that higher exhaust would would make a single-hose AC more inefficient, but it doesn’t seem related to Goodhart, since: (i) it is reflected in the BTU rating on the box, (ii) it is reflected in the temperature of air leaving the unit. So I don’t see how a consumer could be making a Goodhart-related mistake based on this type of inefficiency. It’s also not related to this experiment since it won’t be changed by adding a cardboard tube.
The reasoning was: at constant exhaust temp, each AC pays the same amount per unit of heat moved from the cool air to the hot air. But for each unit of air leaving the 1-hose unit, you have infiltration of the same quantity of hot air from the outside. This undoes some fraction of the work the AC did. What fraction? Well, each unit of exhaust moves C*(exhaust temp—inside temp) of heat outside. And each unit of infiltrated air brings C*(outside temp—inside temp) back. So you lose (outside temp—inside temp) / (exhaust temp—inside temp).
Reposting my final comment, which was left out of the summary: