Make success rate relevant again

[Dan33]

3 hours ago, V3G3TTA said:

I've test it, 182 I missed 69 high jumps, success rate around 62%

 

I guess you have been really unlucky, i doubt it's something really related to the 90% accuracy not being real, i clicked the move around 30 times, without missing a single time.

Edited May 19, 2021 by Dan33

[V3g3tta]

20 hours ago, Dan33 said:

 

I guess you have been really unlucky, i doubt it's something really related to the 90% accuracy not being real, i clicked the move around 30 times, without missing a single time.

I will give you 1 million if you stream live games, using 30 times high jumps in succession and not missing a single, otherwise you can give me 1 million  

[Dan33]

12 minutes ago, V3G3TTA said:

I will give you 1 million if you stream live games, using 30 times high jumps in succession and not missing a single, otherwise you can give me 1 million  

 

ahahaha i just reported what happened to me, i don't like to gamble. Cause i know 1/10 hjks should miss.. i got lucky i didn't miss, you got unlucky you missed that's it, pure probability 

[Kary0plasma]

On 5/19/2021 at 1:15 PM, V3G3TTA said:

I've test it, 182 I missed 69 high jumps, success rate around 62%

 

Math time:

The likelihood of this success rate is calculated by the probability mass function of the binomial distribution f(k,n,p) = (n over k) * p^k * (1-p)^(n-k).

k is the number of successes in independent Bernoulli trials, so in your case it means, how many HJKs you hit, which is 182-69 = 113. n is the total number of attempts, so 182. p is the chance of success, which is 90% in this case, or 0.9. (n over k) represents the binomial coefficient and is equal to n!/(k!*(n-k)!). We can just dump all that in f(k,n,p):

 

(182 over 113) * 0.9^113 * 0.1^69 = 1.17 * 10^-23 = 0.00000000000000000000117%

 

You are as unlucky as 1 in 85400000000000000000000 (85 sextillion) people.

 

As a comparison, if I would mark one single grain of sand on planet earth and then tell you to pick a grain of sand at random, you would be just as likely to pick the one I marked.

 

I suggest to release your Medicham. It's cursed. Also, you shouldn't offer 1 million to that other guy. Hitting 30/30 HJKs is roughly 4%, slightly worse than encountering an HA poke without BMS.

Edited May 22, 2021 by Kary0plasma

[Shinohara]

On 5/18/2021 at 9:47 PM, V3G3TTA said:

Hello team,

 

So I've been playing the game for some time now, a few weeks ago I started testing the success rate on low and high attack.

 

Over the period of 2 months using Lopunny, I have used high jump 182 times, the success rate was 62%, you can argue that the sample is not big to get conclusions, however it's a beginning.

 

Two days ago, my opponent landed 7 hydro pump in succession with Keldeo and missed the 8th, 87.5% success on the battle. Today my heracross missed pin missile which has 95% while Tornadus landed Hurricane (70%) and every hurricane after that point, you might say it's just pure luck.

 

Again we tend to say luck, it was my day and other stuff, however I believe the success rate needs rework, it's pretty irrelevant right now and it causes a major dissatisfaction.

 

How many times have you seen people raging because it doesn't make sense?

 

Honestly speaking, what you are describing just looks almost impossible. Only explanation I can give is one of the following:

1. You are lying and just increased the amount miss to make it more... "OMG!";
2. You are one of the most unlucky person in the world, cause what you describe has  0.00000000000000000000117% chance of happening;
3. There is something that decrease High Jump Kick accuracy and that has not been found, however after several tests we didn't find it.

I have personally made a Python script to run 182 attempts and see effectively the amount of success and fail with 90% chance (High Jump Kick has 90% accuracy).

After that I have run that script 700 different times (so basically it made 182 attempts, then other 182, then other 182, etc).

 

The lowest I was able to go to was 150 success and 32 fail which is 82.42% of accuracy. Do keep in mind that this is the lowest possible I have encountered.

 

---

After that, to verify if there was any issue in the game we have proceeded and 5 different testers actually tested it.

 

USERNAME	ATTEMPTS	SUCCESS	FAIL	% ACCURACY
TIGRERRA	64	56	8	87.50 %
MALAMAGMAR	99	88	11	88.89 %
METAGUROSU	80	75	5	93.75 %
ACTIVECAMO	100	90	10	90 %
GOUDA	80	68	12	85 %

 

 

As you can see, the lowest is 85% and the highest is 93.75%. Obviously, the higher you go and the more the % will stabilize itself around 90%.

Honestly speaking, numbers there looks pretty correct.

 

---

 

There are obviously things to keep in consideration like Bright Powder, however on a number such as yours (182), it looks really impossible to have the total accuracy go down to 62% unless you have played against only pokemon holding that item.

[Eaty]

Heyho, I used dozens of bot accounts that spammed HJK. Results are following:

 

Hits 54930
Misses 6070
AVS 0.9004918032786885%

 

From my point of view everything seems fine.

 

Edit: Added a video while bots are battling: High Jump Kick %testing

 

[V3g3tta]

On 5/22/2021 at 6:26 PM, Shinohara said:

 

Honestly speaking, what you are describing just looks almost impossible. Only explanation I can give is one of the following:

1. You are lying and just increased the amount miss to make it more... "OMG!";
2. You are one of the most unlucky person in the world, cause what you describe has  0.00000000000000000000117% chance of happening;
3. There is something that decrease High Jump Kick accuracy and that has not been found, however after several tests we didn't find it.

I have personally made a Python script to run 182 attempts and see effectively the amount of success and fail with 90% chance (High Jump Kick has 90% accuracy).

After that I have run that script 700 different times (so basically it made 182 attempts, then other 182, then other 182, etc).

 

The lowest I was able to go to was 150 success and 32 fail which is 82.42% of accuracy. Do keep in mind that this is the lowest possible I have encountered.

 

---

After that, to verify if there was any issue in the game we have proceeded and 5 different testers actually tested it.

 

USERNAME	ATTEMPTS	SUCCESS	FAIL	% ACCURACY
TIGRERRA	64	56	8	87.50 %
MALAMAGMAR	99	88	11	88.89 %
METAGUROSU	80	75	5	93.75 %
ACTIVECAMO	100	90	10	90 %
GOUDA	80	68	12	85 %

 

 

As you can see, the lowest is 85% and the highest is 93.75%. Obviously, the higher you go and the more the % will stabilize itself around 90%.

Honestly speaking, numbers there looks pretty correct.

 

---

 

There are obviously things to keep in consideration like Bright Powder, however on a number such as yours (182), it looks really impossible to have the total accuracy go down to 62% unless you have played against only pokemon holding that item.

Hello mate,

 

Thank you for testing it, my weeks have been long lately.

I'm specifically referring to Lopunny when using high jump, as this is the one I tested in pvp, the number that I gave above is absolute.  if you are saying everything is fine, I don't have the means to prove you are wrong or provide a valid evidence, apart from playing and keeping notes.

If you are surprised from the outcome, as I'm the one experienced, imagine how I feel. I would love to have been one of the people who tested the above as that would prove, what I'm saying.

I want to end this, as there is no point arguing, however I would like to express my gratitude for testing it.

[V3g3tta]

On 5/23/2021 at 10:41 AM, Eaty said:

Heyho, I used dozens of bot accounts that spammed HJK. Results are following:

 

Hits 54930
Misses 6070
AVS 0.9004918032786885%

 

From my point of view everything seems fine.

 

Edit: Added a video while bots are battling: High Jump Kick %testing

 

It's very convincing, I can see the number doesn't fall below 89%, which is fantastic, however as I said previously I'm referring specifically to M. Lopunny while pvping. I don't understand how the coding works, is it possible to change from pokemon to pokemon? could be affected from your pvp points or the fact that you are pvping? 

 

 

[V3g3tta]

On 5/22/2021 at 1:15 PM, Kary0plasma said:

 

Math time:

The likelihood of this success rate is calculated by the probability mass function of the binomial distribution f(k,n,p) = (n over k) * p^k * (1-p)^(n-k).

k is the number of successes in independent Bernoulli trials, so in your case it means, how many HJKs you hit, which is 182-69 = 113. n is the total number of attempts, so 182. p is the chance of success, which is 90% in this case, or 0.9. (n over k) represents the binomial coefficient and is equal to n!/(k!*(n-k)!). We can just dump all that in f(k,n,p):

 

(182 over 113) * 0.9^113 * 0.1^69 = 1.17 * 10^-23 = 0.00000000000000000000117%

 

You are as unlucky as 1 in 85400000000000000000000 (85 sextillion) people.

 

As a comparison, if I would mark one single grain of sand on planet earth and then tell you to pick a grain of sand at random, you would be just as likely to pick the one I marked.

 

I suggest to release your Medicham. It's cursed. Also, you shouldn't offer 1 million to that other guy. Hitting 30/30 HJKs is roughly 4%, slightly worse than encountering an HA poke without BMS.

Looks good and impressive, that's the whole argument the 90% is not statical but variant in this case. Bernoulli trial confirms there is 0% possibility if the event probability is 90% to fail 69 times
 

[Kary0plasma]

On 6/4/2021 at 10:57 PM, V3G3TTA said:

Looks good and impressive, that's the whole argument the 90% is not statical but variant in this case. Bernoulli trial confirms there is 0% possibility if the event probability is 90% to fail 69 times

 

It's unsurprising that the calc came to the same conclusion since it is based on the same formula. Btw you can safely slide the number of decimals after the point to the max the calc supports (20) and you will still get 0.0%. That's how unlikely 69 misses are.

 

In other words, I don't believe OP's story.