You can do a lot of calculations yourself. You just take 2 creatures and look at all the best move orders. You do not need to take stupid move orders into account (Like instant distract on an immune). But when I calculate, I look at the first moves. Which is the most logical, and then go from there. Here is an example: Indom gen 2 vs Procerathomimus
Turn 1: Procerath can go for 3 moves: ES, ID, and null strike
Indom can go for 3 moves: Definite strike, MF, and cloak.
Procerath would never use ID and Indom would never use MF, so you can disregard those moves. You then have 2 possible turn 1 options for both, so you NOW have have 4 scenarios
Turn 1 (1 of 4): Indom can go cloak and procerath goes null. Procerath will use rampage next turn and win
Turn 1 (2 of 4): Procerath can go ES and Indom uses strike. Indom will rampage and win
Turn 1 (3 of 4): Procerath can go ES and Indom uses cloak. Procerath uses null and wins
Turn 1 (4 of 4): Procerath and Indom both use strike, and they can both move on to turn 2
Turn 2: There are really only 2 moves they can go for. Procerath can go for Rampage and win. But then Indom can use cloak and dodges the rampage, it wins. However, if procerath predicts the cloak with a strike, it wins. But if Indom predicts the strike with a rampage, it wins
You can do a lot of these yourself and can use the Gamepress sim as a general guide, but they aren’t always correct. For example, the sim says Baryonyx/Tryostronix vs Mammolania is 37% in favor of Baryonyx and Tryostronix when it should be 60% win rate for Bary and a 50% for Tryo. Same with Gorgosuchus vs Mammolania. Gorgo wins 95% when it says 92%. And they don’t represent tossup matchups like the one seen above with procerath and Indom gen 2. It just says Procerath wins 100% of the time. The same goes with entelomoth vs geminititan or Tenontorex. And then speedties can skew results as well. You always should do these matchups by hand just to be sure.