Those of you who have had the (mis)fortune to engage me in canard conversation in the last 2 years will know that I had my engine rebuilt due to assembly deficiencies. After breaking-in the engine, it became clear that due to the new engine having more horsepower than the old engine, cooling was inadequate.
For the last 18 months I have been making all sorts of modifications to the cowlings on my plane to try and bring down the CHTs to what I consider to be a safe operating level. During that time I have hardly ever run the plane at full power in level flight - I could not do so due to CHT issues.
I can report that I made a breakthrough last week. After failing to get the CHTs below 375 at 2500 rpm for a while, I flew home last weekend after getting my prop dynamically balanced. At 2540 rpm, 3500 feet and OAT of 65 degrees, I saw CHTs as follows:

#1: 355
#2: 350
#3: 370
#4: 350

I achieved this significant improvement by simply blanking off 50% of the opening in the top cowling where cooling air flows through the oil cooler (my oil cooler is on the top of the firewall, with exit air exhausting out onto the dorsum).
Oil temperatures in level flight were around 190 degrees.

So...working on the principle that I might as well go the whole way, I then blanked off the entire opening. This time, at similar outside conditions, I saw the following:

#1: 350
#2: 350
#3: 360
#4: 350

The oil temperature was around 195 degrees.

I now believe that the CHT issue was caused by the fact that the opening in the top cowling was allowing too much pressurised air to bypass the cylinders. The air was taking the line of least resistance.
Prior to trying this modification, I had made all sorts of other changes including adding armpit scoops to the lower cowlings, adding exit ramps over the cylinders in the top cowling, changing the airbox profile and angle inside the cowling etc. etc. None of those modifications made any significant difference, still leaving me 30-40 degrees higher than where I think I need to be for safety.
I will now blank off the additional armpit scoops to see if I can get a further reduction in CHTs. (With multiple inlets in a cowling it is possible for air to enter one inlet and exit from another inlet if there is any significant pressure differential between the 2 inlets. This effectively loses cooling air out of the lower cowling, reducing airflow through the cylinders).
I intend to fly to the EAA Hondo fly-in this weekend with the plane in that modified configuration, to assess performance on a long flight.
I may yet need to reverse all of the other cowling modifications that I have made over the last 18 months.
This, folks, is what I think they meant when my elders talked about "learning from your mistakes"...