To time the engine, take the vehicle to a dyno. An EGT gauge should be used if possible.
You want to be reducing timing as much as possible, until the EGT begins to climb notably, and the torque drops notably at the same time. Once you find this point, there is a diminishing returns with respect to the rate at which you add back in timing advance, and the EGT will drop/torque will improve. I usually start around 10* of timing for 7-10psi of boost at peak torque region for an engine which was originally turbocharged from the factory (2jz, sr20 for example). If the engine has a falling torque curve due to engine VE you can add a couple of degrees as engine RPM increases; if the torque drop is due to compressor maxing out you do not want to be adding any timing because the air is getting hotter when that happens, so be aware of why torque is falling.
That said, say you start with 10* and notice EGT is very high (1500*F for example in the turbine outlet is too high for 8-10psi of boost. It should be 1000 or 1100*F perhaps). You would want to add a couple of degrees of timing. If the EGT drops significantly in the next run (say from 1500*F to 1250*F) then torque will also improve significantly and you are going in the right direction. If you add a couple more degrees and the EGT stays about the same (it will almost always drop some, don't be fooled: you are looking at diminishing returns, stay out of that region. You want only to see a LARGE change in output/EGT, avoid making timing adjustments more advanced when the return is tiny) then you should be reducing back to where you were before. The engine is 'tuned' when the timing is minimal and results within 5-12% of maximum torque.
Example, pretend scenario
with 10* of timing EGT was 1380*F in the outlet of the turbine and torque is 200
with 13* of timing EGT was 1220*F in the outlet of the turbine and torque is 220
with 16* of timing EGT was 1200*F in the outlet of the turbine and torque is 224
with 19* of timing EGT was 1190*F in the outlet of the turbine and torque is 229
The ideal timing in the above scenario is between 12 and 13* btdc. Notice how more timing always seems to give more torque; if you examine the torque curve carefully you will notice this is often due to tiny spikes or bumps, 'glitches' in the torque graph where it seems to rise suddenly some places. This is a sign that the combustion events are becoming more 'peaky', that is, pressure is rising higher and higher peaks due to increase timing. This is showing up as a slight torque improvement, but it is very very bad for the engine because these pressure spikes are what damages head gaskets and pistons, they will get worse when the engine gets hotter or is run in longer gears for longer loads and something will break or pop. You want the minimal timing number which gives a reasonably low EGT and torque within 5-10% of the maximum possible.
Engines which run more competition states and have more 'toys' to deal with temperature increase, like blankets, methanol, coatings, wraps, shields, higher capacity oil coolers/intercoolers/water systems, can run closer to that bleeding edge to gain that last 5% of torque or so. But this is NOT daily driver territory, it is not good for we who wish to preserve our engines.
And just for reference, I run 9* of timing in sr20det engines at 15-19 psi of boost pressure with 93 octane with great success, through to redline as torque is fairly flat, 20 years of driving and perhaps 30-40 tuned cars with no issues; 1100*F~ in the turbine outlet when EGT is available. This is with a large camshaft, high VE engine (nearly 100% ve across the board). With the 2jz-gte engine and a stock camshaft, they can use slightly more timing around 11-12* of timing through to redline. I have never needed more than 12* in those engines, for 5+ psi of boost situations and 93 octane Florida pump gas with 10% ethanol.
Here is a graph to help illustrate what I just went through above