In terms of total kWh per year cheap panels win. In terms of the value per year morning and evening power is becoming increasingly valuable due to the duck curve. Tracking solar panels still aren’t viable, but large instillations will sometimes aim a percentage of panels to the east or west to shift power at the cost of total kWh per year maximize value produced.

So, 20% more cost for 10% more power could be worth it even outside of space constrained applications.

I'm not sure I understand your statement that tracking solar panels aren't viable. The vast majority of utility scale solar PV installations in the United States use single axis tracking systems. These orient the modules on a north south axis, which then tilts from east to west throughout the day. Energy capture over an equivalent optimized fixed-tilt system is significant. The land area required is larger, however.

Most smaller or behind the meter systems such as carports or rooftop mount solar use fixed tilt, and fixed tilt is also used in certain types of terrain or where costs for tracking are prohibitive.

Yes single axis tracking are common, but dual axis racking isn’t.

Horizontal Single-Axis Solar Tracker’s are normally aligned north south, but shifting slightly off axis from pure north vs south can increase output in the morning or evening.

Tilted Single-Axis Solar Tracker’s are closer to the performance of dual Axis trackers, but have a similar question do you pick a tilt angle to maximize annual, summer, or winter output.

The duck curve is heavily misunderstood.

The repositioning of panels to capture more morning and evening light is a real thing. At the theoretical extreme it can make the PV output a square wave.

But, that actually makes the duck curve worse, since the duck curve is about the gradient of that switchover.

But the duck curve has never really been an actual problem so it's kind of academic. And if it was we have solutions to it.

The gradient is one issue, the other is sending wholesale prices to zero for extended periods most days. That messes with the economy of baseline power generation especially nuclear.

Yes, I suppose cheap energy can be seen as a problem if you're an energy seller, not a buyer or user.

Buyers will eventually face a related issue.

To simplify if nuclear/geothermal or whatever was 10c/kWh if sold 24/7 but if wholesale prices are 1c for 8 hours a day then base load generation needs to make up 8x9c = 72c over the other 16 hours. Now prices would hypothetically jump from 10c for those 16 hours to 14.5c for those 16 hours.

It should still be a net savings for the average consumer, but if your business model is based on cheap electricity at 10am to 4am that might flip to 10am to 4pm.

Cheap panels’ output drops as temperature goes up.

I hadn’t thought of it until just now, but that should flatten the duck curve.

Another way to flatten the duck curve is to buy smaller inverters and more panels.

There are probably applications where 10% more power at 2x the cost make sense, like on a satellite or a solar powered aircraft, maybe.

Indeed, though there are other places of the design space to look. For example in satellites, triple junction cells (generating power from additional parts of the spectrum) are worth the much higher cost as they weigh the same as standard cells, thus generating more power without increasing launch cost.

How much more mass, though? Mass is likely to be at a premium anywhere you can't just also slap down a few more cheap cells, e.g. where there's not much space.

Good points by you and your sibling poster. Maybe this tech really would have a hard time finding a niche when considering mass along with cost/performance characteristics, given other premium or economic options. Maybe on the roof of an expensive high rise it would make sense? On boats? Hard to say.