You all have to be a little careful here.
First, aluminum is OK to ground to unless it is too thin or not solidly connected to the block. The same is true for steel. The thing you have to watch is corrosion, or electrolysis from dissimilar metals. You have to pick the metals that contact other metals. I like to use a stainless buffering material, like a stainless fender washer, to stop dissimilar metals from interacting.
Sometimes we beg for problems.
When we ground to the heads how do we think the current gets to the starter?? Hmmmm. The head is gasketed on all major contact surfaces to other parts.
The current flows though any bolts connected to the head that also happen to have a path to the block. Now it just happens that locktite is a great insulator. So if all those big head bolts have loctite on them, we might wind up depending on something else like an intake bolt to get to the intake for a ground. But wait, that intake sits on gaskets! So now what? The current has to find a path to the block some other way. Maybe through the rockers and pushrods and through the cam? Maybe some small wire for some sensor? We just don't really know, do we?
This is why the factory grounds with that big battery cable to the block or to something they are **totally sure** will never be insulated from the starter housing. Then they run a smaller ground wire to the chassis from the battery negative post to power everything else, or maybe from the block or alternator bracket to chassis ground. There is a reason for this, and it has to do with tearing up parts and starting nasty things like car fires.
We are far better off and far safer to ground to a bell housing bolt or some other large bolt that is non-critical for torque, never has loctite, and that is connected directly to the starter through non-critical things.
I'm not saying the heads won't work 98% of the time, just that there is a reason the manufacturer almost never does that. (If they are SURE the heads will directly contact the starter housing all the time in all cases they will.)
Also, we really only need the same size battery in the rear as we had in the front so long as we use a large enough cable. The problem is not the current capacity of the cable, but the total resistance of the path. An AWG-4 cable is super for a six foot run, but it starts to get marginal for a trunk mount battery. There are also safety reasons to not use cheap #4 cables.
AWG-4 copper has a resistance of about .00025 ohms per foot. With a 200 amp starter load that means .05 volts per foot. You've lost a volt (about 10% of the available cranking voltage) in the starter to battery hot cable alone in most rear mount batteries, and if you ran dual cables (one for the ground) it would be 2 volts with number 4 wires. It can get worse if the starter is hot, compression high, weather cold, or the timing is advanced.
I use 00 welding cable for my remote batteries. Why? It has 1/3 the resistance of number 4 cable. I can run it three times further with the same drop. It has a tough jacket.
So why does a bigger battery help? The internal resistance of a bigger battery is lower so the battery itself does not sag as much when you load it. If you have a good small battery that works fine up front and it fails to work in the rear, it is not because the starter suddenly needs a bigger battery. It is because you added so much other loss that the system just can't tolerate what the battery voltage is under load any longer. You made up for resistance in the cables you never should have added by removing it from the battery.
As for trunk grounds, be careful! Think about what would happen if the positive battery cable was pinched or the insulation fails. The only thing that limits current is the internal resistance of the battery and the path from the battery to the chassis and back to the short at the hot cable.
The result will at best be a wire fire, and at worse a battery explosion. You should actually use a full negative return cable to the engine block and ONLY ground to the chassis with a fusable link cable that will melt if a hot lead goes short to the chassis.
Now most of us won't do that. Instead we will want to use the chassis as a return to save running a second cable.
Doing that, we now have a problem in that full starter current has to pass through what normally would serve as a fuse link. This is another reason why you should use a BIG positive cable with high temperature insulation!! Then you can use a short number 6 AWG link from the battery to the ground and keep that wire clear of flammible materials. The fusing current in that #6 wire would be about 700 amperes. So it would melt and open the battery from ground before the main feed wire caught fire or started melting through metal.
The resistance in a number 6 wire is .0004 ohms per foot, so at 200 amps it would only drop .08 volts per foot. It would not hurt the system if it was less than two or three feet long.
Finally, be sure the chassis connection is to a fairly wide fairly thick section of chassis that is an integral part of the chassis. Be sure you ground the BLOCK to the frame very well. Think about where the current will go if you do not solidly ground the block to the frame and you DO solidly ground the battery to the frame. The path is likely up the driveshaft, through transmission parts, and into the bellhousing. It might not be good if you run 200 amps of starter current through some of the needle bearings in a transmission.
I know this was long, but I hope it stops a car fire or damage! At least consider it. There is a reason things are the way they are!
(former designer of tachs, ignitions, timing lights, dwell meters, gauges, sensors, regulators, and other things electrical)
89 LX coupe with 363, modified PT 7675 CEA turbo with billet wheel (at moment), Super Glide transmission, street legal with full interior. Runs low mid 5's in the 1/8th, mid 8's in the 1/4 through mufflers on Pro Radial tires