David,
In my case, it was a series of problems, which started with a poor design by Steve Lawing of SAS. His aerodynamic chart and calculations for the cooling system looked like something you would expect from a bright high school freshman at the local science fair, not from an experienced builder/engineer/savant, etc.
The radiator was in a bad place originally and couldn't get enough air through it to make it work. Lawing had riveted a 3" x 5" bent piece of aluminum to a frame member to deflect air to the rad. The stock 2008 Legacy rad was completely open with 2 puller fans and no shroud.
I reconfigured the entire cooling system on my mid-engine application. I moved and redesigned the radiator, fan, and shroud which now sits just underneath the traditional deck lid on our replicas, with twin puller fans and a shroud that forces all the air through the radiator.
I changed the intercooler from air-air to air-water, moved it to the frunk with its own pump and twin heat exchangers. Not coincidentally, a tuner discovered a clever bit of sabotage. Someone had placed a brass nipple reducer in a water line: 1/4" npt to 1/8" barb fitting, with just enough room in the hose for some water to bypass. I took someone's name in vain when I found out about that.
I try to look back on that time as a learning exercise, like making lemonade out of lemons, eh? I did learn a lot about air and water movement, radiator and fan design, etc. If that's as bad as my troubles get, I'm a lucky guy.