Overlap is the period where both valves are slightly open. Overlap helps scavange the exhaust gases. Less overlap helps build combustion chamber pressure. Generally, low compression ratio and low RPM likes to have short duration intake and exhaust, and little overlap. High compression ratio and rpm like longer duration and more overlap. There won't be a big differrence in lobe separation, just a few degrees.

I assume intake centerline is the point of highest lift.

Go with what the cammaker recommends for a racing motor. Go with what they reccomend for a street motor, too. Thats probably a little milder than what you think. Street motors spend very little time at 6000 rpm, and too big a cam isn't fun at 2000 rpm and sucks gas.

.
.
.

We used to call them "hairy bump sticks"

That's because the "bumps" on the camshaft were higher (get more gas in, and let the exhaust out faster),

and made the cars performance "hairy"!

I was building engines by '72 - just about to build another one (for myself).

Just as an aside - remember a few idjuts that decided to put in a hairy cam - and wondered why it ran like shit - - -

umm - well that's because that's ALL they did - just put in a high-performance cam - no carb changes, no timing curve changes, no headers, etc., - nadda - -

BUT - they idled real tough!



CC

Unless you're doing this with good simulation software (I use Engine Analyzer Pro), the best thing to do is figure it out by narrative - Basically, look through the listings for your engine, and pick a series of cams that matches your application. then narrow it down within the series by RPM range and cruise RPM. Then "degree it in" using the EXACT procedure on the cam card, because it may differ depending on the lobe profile, even from the same manufacturer. And be sure that your head(s) are set up with sufficent travel to accomidate the cam's lift - this is EZ with near-stock stuff, and a little spooky if things get radical.

I've tried fiddling around with lobe center and cam advance/ retard (same as intake c/l, just a different means of stating it) to fine tune combinations, and, most of the time, the mfr. reccomendations are on the money if you are using the cam as intended.

If you have more questions, PM me. I do full-on modeling and optomization for your combo for $$, and I use a LOT of data I've compiled, as well as the fmaster lobe catalogs if necessary.

One of my combinations in action:

That raised my heart rate!

Solid lifter roller, engineered to live on the street. That engine was designed to have an extremly broad powerband, and to be mechaniclly efficent - not to win bragging rights on the dyno.

used desktop dyno? How accurate is that software? I do have a copy of that.

In general, the simulation software is pretty good to A/B 2 choices. The better the information you cam use, the more accurate you can be. EA Pro lets me get to a serious level of detail - like flow numbers at 5 different valve lifts, and similar data for cams. Rod length, combustion quench, collector taper - great for an obsessive like me.
If you are dealing with pretty mainstream stuff, I think Desktop Dyno will do OK for you. The basic engine files for stock configurations of popular engines are gonna be pretty close, and the same for popular series of cams.

In VERY general terms, your original questions: overlap/lobe center - decreasing lobe center increases overlap. Wide lobe centers are better at low rpm, and may help top end a bit. Conversley, Narrow lobe centers help midrange, especially if the torque peak is in the midrange as well - but the whole range ain't all that big.
Advance/retard - advance tends to move the power peak lower, cam retard moves the peak higher, As timing chains stretch, the cam retards - so, if anything, you want to shade a new build a couple degrees toward advance, so that it will come right as the engine breaks in. A lot of mild cams thes days have 2-4 degrees of advance "ground in", They seem to work best in this range, and this allows them to be installed with a normal timing set and hit the mark most of the time - and if tolerances stack, they won't end up retarded. Retard tends to not work out well for any sort of "normal" combination.