According to the math, the -4BTDC is actually a double negative as it is read. It could be written as 4ATDC and mean the same thing, but for the sake of continuity, IVO events are always listed as BTDC. In this case, it's just that the .050" duration numbers put it ATDC so they use the negative.
Add up the degrees. There are 180 degrees in the intake stroke. It closes 28 degrees after that 180 degrees, so that makes 208 degrees. Since the intake lobe is 204 degrees, you subtract that 4 degrees from IVO, and (by the power of maths) yup, it opens 4 degrees ATDC.
In truth, those are the numbers at .050" lift. In real life, it will open BTDC but the number of degrees is up for question. Advertised duration is just that... an advertisement. Some cam manufacturers measure advertised at seat-to-seat, others advertise at .006", some measure at .005". Some take the actual measurements and select a "catchy" number that is close. If the seat-to-seat measurement is 249 degrees, they might call it a 246 advertised duration, because two four six is a psychologically more memorable number. The actual seat-to-seat duration is also of less importance given the miniscule amounts of flow that happen in those initial lifts, especially given the variables of lash, hydraulic lifter compression, pushrod deflection, etc. As a reference, .006" (actually .003" times two) is about the thickness of the oil film that hides between the ends of the pushrod and the lifter/rocker, much of which gets squished out during lift, so stressing about advertised numbers being seat-to-seat or .006" is splitting hairs. What is MOST important is the comparison between advertised and .050" duration measurements as it is an indicator of ramp angle/speed. Bigger differences in those numbers indicate a slower ramp speed. Faster ramp speeds make more power since they introduce more area "under the curve" during valve lift, but they do so at the expense of lobe and lifter reliability.
Given your cam card, I would stab it in straight up (leaving the 6 degrees as-ground, or using the 0 degree setting on the timing set). You mentioned that your timing set is adjustable at the crank, right? So the drive cog has different keyways for the crank? You can play with it, although that means constant removal of the timing cover. I would run some dyno simulations with Dyno2000 to see what they do to the torque curve as-ground, and with your 4 degrees of wiggle room. It should be available as a free download. There is a nice box where you can simply select advance/retard so you can compare curves side by side without going nuts re-entering data. As long as the peak of your torque curve is within about 250 rpm of your stall speed, choose the one that suits what you want.
Trust me on one thing, though. Before you start playing with Dyno sim software, read the entire user manual. It is not as simple as plugging in your specs. You can't just take it at face value. For instance, if you were putting in a roller cam and selected "hyd roller" from the drop down menu, it assumes the most aggressive ramp profile as a pre-set assumption. It is better to manually enter lobe profile numbers as well as head flow numbers at each lift point. Too many times people plug in numbers and use the pre-set profiles and suddenly their stock 351 is making 450 hp just by switching to headers. It's difficult to use simulations to accurately predict actual hp/tq numbers, but one thing they do very well is predict trends and use VE to show you where the peaks are in the RPM band. That is to say, it might be off by 20 hp at the peak, but at least you will know, for example, that peak torque will be at 2700 rpms and peak power at 5400 rpms. Playing with the cam advance will let you see how it changes peak torque RPM.
Either way you go.... definitely check valve-piston clearance. My guess is you'll be OK whatever you do, but advancing the cam will start the valves lifting sooner, especially the intake valve. It will be that many degrees further advanced and therefore further open as the piston approaches TDC.
