If you do a battle where you have an 80% chance of winning and 20% chance of losing the entire game, when you have a perfectly good alternative of having a 97-99% chance of winning, that is a bad choice and bad gameplay. Letting Italy get Southern France on I1 or going after Normandy on G2 are perfectly acceptable options that dramatically lower your chance of a Paris disaster. If you want to play with the dice to be in a slightly better situation with a win, but potentially game-losing outcomes if you are not lucky, you are not a good player.
Getting diced is when you make smart moves that have a high probability of success, there are no other alternatives that would be safer yet yield essentially the same outcome, but lose because of bad rolling. If Germany does a Moscow bombing run with 4 bombers and loses all of them to AA fire, that is getting diced. It was a good tactic with a horrible outcome that will change the game. If you lose the entire air force on G1 when going after the UK fleet, that is getting diced. If you lose in Yunnan on J1, that is getting diced. If you have two Italian tanks plus two Italian mechs fail to kill a single Russian blocker, that is getting diced. If your big battle in the Pacific that should have a +100 TUV outcome results in a -100 TUV outcome, that is getting diced.
I agree with Kreuzfeld that overambitious and unnecessary attacks combined with unlucky rolls is NOT getting diced. There is absolutely no need to have less than 97% chance of winning the Paris battle. That still leaves 3% chance of loss and I do have great sympathy of people who get diced on that combat.