Added more accurate random and its model file.#1980
Added more accurate random and its model file.#1980kanemasa1987 wants to merge 6 commits intoPokemonGoF:masterfrom kanemasa1987:random_mod
Conversation
pokemongo_bot/human_behaviour.py
Outdated
| def normalized_reticle_size(factor): | ||
| minimum = 1.0 | ||
| maximum = 1.950 | ||
| return uniform( |
There was a problem hiding this comment.
Why not use your RNG for this too ?
There was a problem hiding this comment.
Gps errors should be approximated by gaussian.
I'm not quite sure for reticle and spin rngs.
Speaking of myself, I try to get simple 'Great' and don't throw spin balls because I'm bad at it.
However, that might be my preference and most of the users might be acting otherwise.
Besides, I'm also not certain what distributions are proper at this point.
It should be other than uniform to prevent users from getting banned in the future, but I decided to keep it as it is for now, before going deeper and messing around with the codes (by adding ball throw cfg etc).
|
Very nice work! |
| if random() < success_prob: | ||
| break | ||
| return trial | ||
| time.sleep(1.8*(trial+random())) |
There was a problem hiding this comment.
This sleep will never execute.
There was a problem hiding this comment.
Oh, I forgot to remove the checking code! 🙇
|
Now I realized the branch is wrong. |
3 participants
Short Description:
Added and modified random functions to simulate a human player better by using other random generator.
However, parameters for log-normal are tricky (ex: for average 50ms and 20ms std mu=-3.00, sigma=0.132,) so I have added pre-calculated model file and its calculation code.
Fixes:
"Updated Fixes":
Time and reflex distributions
reticle size distributions
spin distributions
