So tharos_mt, a fellow redditor has given me datamined critical rate formula, and it was the following:
(Critical Chance) = (Critical Rating)/(Critical Rating Divider) + (Level Difference)x(Level Difference Multiplier) + (Base Critical Chance)
Luckily, this formula is linear! And, since level difference can easily be set to 0, I knew this was going to be rather easy task.
However, I knew that EA/CG could have set different Base Critical Chance for different game modes, as they did with animation speed, and since I had absolutely no clue about what level enemy heroes were, I could not conduct research.
However, recent update unveiled enemy heroes' levels! Now I could finally analyze how critical ratings worked. I conducted over 30 battles, recording around 400 hits to come to following conclusion. But, before going in to the results, I need to assume several things.
1) Above datamined critical chance formula is correct
2) I have not made human error in documenting critical hits
3) The game has not made an error in displaying enemy level or critical hits
4) The game has not made an error in generating random numbers
5) Critical Rating Divider (CRD) is equal for all heroes
6) Critical Rating Divider is equal for both Physical and Special Damage
7) Base Critical Chance (BCC) is equal for all heroes
8) Developers have not changed neither the Divider nor Base Chance while I was recording critical hits
Since I know the formula is a linear one, I conducted linear regression to get coefficient 1/CRD and y-intercept BCC. My R-Squared reached 96.9%, and I decided to stop conducting, because crystals are scarce, and quite a decent portion of my data started to settle as my observations neared 400.
My work(Google Spreadsheet) can be viewed here
My linear regression result showed that CRD is 1736.41 and BCC is 0.045. Now I thought that EA/CG would not have chosen such random numbers, and must have chosen 1750 for CRD and BCC for 0.05. This is purely my speculation, but one I can stand by.
Thus, my conclusion is that :
(Critical Chance for Same Level) = (Critical Rating)/1750 + 0.05
You will see that my guesstimation has minimal impact on the overall predictability of my regression result.
Now I know what Critical Rating means! This work was important for my upcoming Role-Adjusted Tier List, which will be epic in proportion.
One interesting thing about this chart is that, special CC is lower than physical CC for ALL CHARACTERS.
|Name||Phy CR||Spec CR||Phy CC||Spec CC|
|IG-86 Sentinel Droid||516||20||34.49%||6.14%|
|First Order TIE Pilot||487||10||32.83%||5.57%|
|Coruscant Underworld Police||453||0||30.89%||5.00%|
|Jedi Knight Anakin||449||0||30.66%||5.00%|
|Clone Sergeant - Phase 1||423||0||29.17%||5.00%|
|First Order Officer||409||10||28.37%||5.57%|
|Hoth Rebel Scout||378||25||26.60%||6.43%|
|Clone Wars Chewbacca||325||20||23.57%||6.14%|
|Hoth Rebel Soldier||290||80||21.57%||9.57%|
|Obi-Wan Kenobi (Old Ben)||284||10||21.23%||5.57%|
|Grand Moff Tarkin||259||25||19.80%||6.43%|
|Jedi Knight Guardian||255||35||19.57%||7.00%|
|Poggle the Lesser||237||85||18.54%||9.86%|
|First Order Stormtrooper||216||10||17.34%||5.57%|
|Grand Master Yoda||5.00%||5.00%|