what I use to grind rubber is similar to an angle grinder but with a much slower speed and a grinding disk like you might use to grind welds.
If you go too fast it burns. You can cut rubber with a knife, but the pressure of the knife causes deflection. that often causes a lot of inacuracy.
cutting tools like the bit of a metal lathe or a chisel of a wood lathe or a drill bit won't work but you can use tool post grinder and the grindstone should turn opposite the piece being turned. a dremil with a stone in it will grind rubber.
a good table may attach the rail to the slate with bolts, or it may have steel backed rails which provide weight. when the inertia of the ball hits, the more dead and heavy and well bolted into place that rail is, the better bounce you will have and on a good table it's a lot more silent. materials used to make the rails matter. ideally they are heavy wood, and not some junk particleboard.
If you move the rails to tighten the pockets, then you may only be relying upon the clamping force of the bolts holding the rails (because the bolts , when loose, are a loose fit in their holes )
and that could have some "give" The rail may "give" very slightly because there is not a good fit in the bolt hole. If on the other hand , any slack is taken up by moving the rail outward as far as the rail allows and then tightening it, then it physically can't move further.. (we are only talking about a few thousandths of movement and it will return to where it is after impact) ,
if the rail is shifted ( outward) in such a way it takes up that slack in the bolt hole than the ball can't move it further upon impact.
there is some initial engineering to this but its good to be aware that you can have a rail that is not solidly mounted and that may not really appear loose but won't have the same bounce as one that has a good firm mounting.
to be exacting , I think every table manufacturer has a template and that is really the real reference to the position. Manufacturers may or may not be very free about their templates as they may have some proprietary designs.
my suggestion is that if you do intend to reposition the rails, you might make an exact template first, that way youll be able to observe the differences. you can cut plywood or use cardboard or other things or some combination. fit that to the rails and measure the openings of the pockets, if you dont like your new positioning then its reversiible.
I was working on an older small brunswick and that table merely had screws up into the rails. what I found worked was to carefully lay a bit of electrical wire under the rail so when I tightened it down it became a bit like a shim at the outtside edge and that helped. If the rail twists even just slightly ( from flat to the slate) then this will also change the distance fromt he table to the ball. if the felt or staples are bunched up under the rail that can interfere too.
that proportion is very important. if the rail is too low it will launch the ball upon rebound, if its too high it will push the ball down to the felt, creating an undsrable gulley along the rail over time, and not performing so well. I'd suggest taking careful measurements so when you replace the rails you know where it started. I'd check and measure from the table to the contact area and bear in mind that contact area changes depending how much the rail is deflecting. If you take a ( moderately) hard shot at a rail and see the ball bounce and leave the cloth on its return, your rrails are too low or maybe the balls are too big. If you have someone else shoot balls at the rails you can go feel them, try to detect if the rails are actually moving upon impact, grab the rail and see if it can rock or move. If it's loose at all, fix that.
id take some practice shots at the rail and note how or if the ball lofts on it's return path. make some notes of that so if you make changes you can retest in the same way and see any differences. If you are pulling a rail off it's good to note these things and then you can look for any differences upon reinstallation.
that proportion, I think its something like 64 1/2 % of the ball size but that's information you can look up.. I could be off on the digits my point is this is directly proportional to ball size.
its common for the rails to have rubber facings which are made of rubber. some cheap tables might use a synthetic rubber. I found the right type at a place that sells parts to service water pumps, they sell real (butyl) rubber in sheets that is grippy and bouncy , rather than the fake synthetic rubber that you may get if you order the online. common o rings are often neoprene, thats not porous and not real rubber, but it does have bounce characteristics. you can order drums of rubber like that , put it in molds and harden it up with heat.
synthetic rubber is not porous, it can be poured into molds, real rubber is made by a different process, there is a great video online that shows Brunswick rubbers being made and you can see the related machinery.
Rubber can be ordered by its durometer reading ( a shore durometer is the tool to measure it) 30 is softer than 80 .. rubber goes harder with age and there is some preferable durometer number, you can measure that with the tool. it would be possible to track the rubber hardness with a measurement this way and there may be some decision to replace them upon reaching a certain hardness number. This is how the hardness of rubber is usually quantified.
rubber hardens with age so your rails become less "bouncy" its attacked by age, and by sunlight and by ozone exposure.
if you look at the cheeks of an existing pocket you will probably note that they are not parallel but a bit "splayed" open.. open more at the opening, so a ball can rattle and be thrown out, changing this , will change how it plays.
also not that on the rails of a pool table there is a shape that drives the ball down upon impact, ( at least if the cushion is triangular) but where you look at the cheeks near the pockets you may note that the ball can hit that pocket inner surface there, and not be driven down.. this is an area where if the ball hits with a lot of "top" it can easily launch right off the table, so you may want a bit of an angle on that, Don't just assume it should be vertical.. I think you may take pics of the setup you have. and make templates even if they are just cardboard and paper..
I'm not a pool table mechanic, i just learn a bit by doing and read what I can about it. An experienced mechanic who does this for a living would have different info and maybe I'm of on some points, if so feel free to call me out on it. snooker table may have a lot of differences in the roundness near the pockets and the shape of the cushions.
If you take the felt off the rails you may not that the cheeks are glued onto the rubber and may extend into the pocket beyond where the cushion rubber seats against the rail. they can be different thicknesses or doubled up to change the pocket size.
on my little Brunswick I lined this edge with real rubber and that made the pockets a bit smaller which I wanted. this was mainly because the OEM rubber is obsolete and it resembled the plug out of an old tub or basin, white and cracked... those boots are quite thick rubber..
so I had to pretty much make my own pockets and line them with rubber, what I did was bought cheapo plastic made in china pockets and put those overtop of real rubber. the cheapo pockets were thin but served their purpose in dressing up the look .. on most modern tables you can get new buckets. It's not something I'd change just for fun.
I also placed a strip of rubber, ( I used a hunk of a ladies belt) the I put the staples to hold the buckets in through the leather. the reason for this part is so if someone slams a ball into the pocket, it hits that ridge and is driven down rather than back out, or up in the air.
if you are interested in tournament play, then you can argue that it makes you a better shot to have tight pockets on your practice table. You can also argue you are just changing the conditions and how you learn to cheat the pocket is important so having a table as similar as possible is better... I dont know that there is a definite answer to that theory, perhaps that is open to your own interpretation.