Yeah, the historical rule of thumb is that you need a 3:1 force ratio to defeat a fortified line. Assuming for simple numbers that you're using forces that consist entirely of on-tech PW infantry (so every hit kills), and that every shot hits in the absence of fortification, consider a 3:1 advantage against fort-6 infantry:
Start: Attacker 3000, Defender 1000. 3000 shots=500 hits=500 deaths, 1000 shots=1000 hits=1000 deaths.
After 1 round: Attacker 2000, Defender 500. (omitting for brevity)
After 2 rounds: Attacker 1500, Defender 167.
After 3 rounds: Attacker 1333, Defender defeated.
So the attacker loses 56% of their forces to take the planet, with a 3:1 advantage. The defenders fall, but they take out 167% of their own weight in the process.
Let's try again with 2.5:1 advantage:
Start: Attacker 2500, Defender 1000.
After 1 round: Attacker 1500, Defender 584.
After 2 rounds: Attacker 916, Defender 334.
After 3 rounds: Attacker 582, Defender 182.
After 4 rounds: Attacker 400, Defender 85.
After 5 rounds: Attacker 315, Defender 19.
After 6 rounds: Attacker 296, Defender defeated.
So 2.5:1 forces still win, but they're almost destroyed in the process, and the defender killed 270% of their own weight in the process.
Let's say the defenders are only self-fortified, so x3, but the attackers are only 2:1 advantaged.
Start: Attacker 2000, Defender 1000.
After 1 round: Attacker 1000, Defender 334.
After 2 rounds: Attacker 666, Defender 1.
After 3 rounds: Attacker 665, Defender defeated.
So basically, an attack of equal force composition, without orbital/fighter support, should be 3:1 advantaged against CON-fortified troops, and 2:1 advantaged against self-fortified troops. You'll lose a bit more than you kill, even so, but you'll take the planet. That seems pretty fair to me.
Don't forget the great advantage of the attacker - force concentration. You can pick your targets and pile up forces at them. The defender needs to be strong everywhere, and thereby spreads their forces into penny packets that can be gobbled up by attackers. Tests which don't consider that are inherently incomplete. (They can still have a lot of value tactically, of course, but they don't cover the strategic side).