Here's something that I can maybe help on...
You have three things going on at any one moment in your kiln. #1 is the heating elements are cycling on and off trying to maintain a constant temp. #2 is the termocouple is reading some temp within the kiln. #3 is your steel in the kiln is searching for equilibrium. The thermocouple is absorbing radiant heat (unless it is shielded), and so is the steel (unless it is shielded). The thermocouple is also being heated by convection, and so is the steel.
At low temperatures, the steel is still pretty shiny. It reflects a lot of the radiant heat, so it's not so big a deal. At austenizing temps, the oxide layer has built up on the steel, so it no longer reflects radiant heat, it absorbs it. Now, you have a problem. If you put the steel in you kiln when both are cold and turn on your kiln, the coils immediately ramp to 2600° and stay there. The steel is shiny, so it reflects the radiation, and it heats by convection at a rate that is roughly eqivalent to the rest of the kiln. Once you hit ~900°, you have a black oxide layer on the steel, and its emissivity in the modified Stefan-Boltzman equation has changed. Now it absorbs much more energy that it reflects, not only because of its black color, but because of the matte surface. Since the thermocouple does not accumulate an oxide layer, its emissivity does not change, and does not absorb any more radiation. Therefore, the thermocouple is oblivious the the fact that your steel is heating much faster than itself and that unless it shuts down the coils the steel may very well overshoot the desired temperature.
Now, consider this scenario; you start the kiln, and all this stuff happens without the steel inside. Once the kiln reaches equilibrium, you can be pretty much assured that what the thermocouple is reading is the actual temp of everything in the kiln. The coils are cycling, unlike before, when they are just pegged at "afterburn". Since modern PID controllers strive to maintain very tight temp control, I would bet it cycles every 2 seconds or so. If you're using an infinite switch, even better. Cycling coils will not have a chance to get to ~2600°. They may cycle around ~100-150° (Don't hold me to these numbers. The point is the coils have to be hotter than the rest of the kiln for heat transfer to take place) hotter than what the thermocouple is reading, then the average temp in the kiln goes up enough and the controller shuts the coils. Coils on an infinite switch may be ~75-150° hotter than the thermocouple reading, and will be a consistent temp since they don't cycle to maintain an average temp. Now you put your steel in. It's getting heated by convection and radiation. It builds an oxide layer and heats faster. But, since the hottest thing in the kiln is now only ~150° hotter than the setpoint, there is much less tendency to overshoot your desired steel temperature.
Before, the hottest thing in the kiln was 2600°, which would be 1100° hotter than your setpoint if you were shooting for 1500°. That temperature difference is what drives the steel to overshoot your setpoint temp.
In either case, your steel is going to be hotter than what your thermocouple is reading unless your thermocouple is laying on top of a coil. How much hotter would require a transient heat transfer analysis that would make your head spin. I personally think that if you are letting the steel sit in the kiln while it heats you could see steel temps that could be in the ballpark of 300° hotter than what you are expecting UNTIL the entire kiln come us to temp. As that happens, the average temperature of the coils comes down and so does the steel temp. In other words, radiative heat transfer from the coils is the controlling factor in steel temp in a cold kiln. In a hot kiln, radaitive heat transfer from the entire kiln will be the controlling factor. Since the temperature of the entire kiln is right at the setpoint and the only thing over the setpoint is the coils(and them only by ~150° or so), the temperature of the steel will be much closer to the setpoint; it could be within 15-50° or so. It may be much closer, but I would have to have an IR thermometer and take quick readings to know for sure.
You can get around all of this in 3 ways... heat treat with foil to shield the steel from the radiative heat transfer, make a firebrick box inside your kiln and put the steel in it to shield it, or wait until the kiln comes up to temp. I build a box AND wait until it comes up to temp, but I'm anal like that.
Now, if your thermocouple is inaccurate or improperly placed, that's a whole 'nother can o' worms.
), we need to be careful not to confuse normalizing with annealing, as many people often do. I will always attempt to address them separately, thus I may stress the importance of high heat in at least one or more normalizing cycles, but then stress the importance of subcritical temperatures in spheroidizing operations. 
