Like I said before, EGT is only a good indication of Stoich near peak (not at peak as we've learned) however even there, AFR varies with the type of fuel- 100LL, Swift, Summer/ Winter E0/ E10 blend Mogas etc. We might be at 14.4 for one type of fuel 15.4 for another- a full point variation.
This is a limitation of this method especially when you throw in changes in ignition timing that may be in play with slight changes in MAP and rpm (FP prop) as you change mixture. You cannot tell with any great precision what AFR might be on different engines on either side of peak. In the end, many people may simply lean to engine roughness and richen slightly to set LOP operation despite and cockpit instrumentation giving additional information. Having multi-probe EGT and WB AFRs just allows us to quantify in some way where mixtures are at. With low GAMI spreads and modern EIs, we may be able to get all the way into the 18s for AFR which is where best efficiency is predicted however the roughness issue will limit leaning no matter what whiz bang gauges we have in the cockpit.
I'd argue that AFR is what matters with regard to what we are trying to achieve and EGT is an indirect method to ballpark AFR since it isn't AFR. This is a reason why today in the modern automotive performance tuning and OEM ECM tuning world, direct AFR measurement is used as the primary tool and EGT as a rather secondary tool. We have the advantage there that individual port flow is usually pretty close due to better design of the entire induction system so cylinder to cylinder AFR variation is generally much smaller than what we'd see on a typical OEM Lycoming.
With equal airflow, equal fuel flow, equal exhaust flow and equal EGT probe placement/ orientation, we should have near equal EGT in each cylinder and near zero GAMI spread. This is the expectation in the auto engine world and the goal probably achievable through modification in the Lycoming world (flow benched matched components and revised aftermarket intake systems).
Why this is so important in the automotive world is emissions and catalyst performance. We can't have one cylinder at 14 and another at 16 AFR to have ideal catalyst performance. Most people would be staggered if they found out what factors are considered and designed around to achieve the best possible performance on the least amount of fuel and emissions in the automotive world. This advances understanding and raises the bar IMO over what is being done at the OEM level over legacy aircraft engines today.
Granted, aircraft piston engines don't have to worry about emissions (luckily) or operating at continuously varying load and rpm levels like auto engines and they are pretty darn fuel efficient running LOP at steady state cruise but they are a ways from optimal as delivered from the OEM. This is the reason for aftermarket intakes, flow benched cylinders, GAMI injectors and injector restrictors- attempts to correct design/ manufacturing deficiencies present in the OEM product.
I'll say it again, low GAMI spread achieved by changing individual injector flow rate is useful but really a partial patch for the actual cause in the first place. Makes people feel good watching that close alignment but all cylinders are not doing equal work which is what we should really be striving for. People have become fixated on low GAMI spread in the aircraft world with only a few companies (mainly on the race and performance side) tackling the actual problem with better designed and produced components.
As I side note of interest, The one customer who has 1700+ hours on the EFI got 400 hours out of his first WB (4.2 sensor). It was still working fine but he installed a new exhaust system and sensor at the same time. He now has another 300 hours on the second sensor. He burns 100LL exclusively and does not use a lead scavenger chemical. He runs LOP all the time in cruise.