Collecting Data on Player / A.I. cascades Part 2

Unknown

This is the second round of 1000 matches from this post:

https://d3go.com/forums/viewtopic.php?f=7&t=30201&p=363751#p363751

Again I marked down every time I got a cascade and every time the AI got a cascade. For this experiment I also counted a cascade as 4 or more matches in one move. A couple things I changed based on feedback from the last posting: I didn't take chances on three dropping into a 5 match, I prioritized Horizontal over vertical. Again all these matches were done with ProfX/Hood/Thor. The matches were done in the Shield Simulator, Gauntlet, and the Prologue.


A.I.
1752

Me
1540

Now I also recorded every game where I used at least one Health pack. 491 games

A.I.
621

Me
922

So in the 514 games where I did NOT use a health pack

A.I.
1131

Me
618

Now the community asked for some other data:

Cascades right before a victory (the ai/me was on last character with 1-2 hits left):

A.I.
136

Me
11 (I only lost 14 matches in the 1000, so this number seems insanely high.)

A.I. Used ability right before victory (the ai was on last character with 1-2 hits left)

A.I.
287

Cascades received after final opponent downed

Me
112

All of this is based on my play style. I am not trying to "prove" anything. I am just presenting you with the data I received. Please remember to take it with a grain of salt. You can look through the previous post and answer most of your questions about how I play in that thread. I changed a couple of things and they are listed above.

Unknown

I didn't take chances on three dropping into a 5 match

Want to firm up your results: If the board was like this at the top, would you take the blue match?




Or (again at top)




If you didn't take the match, did you at least neutralize the match?
Are you counting cascades that come from ability use differently?

Unknown

daibar wrote:

I didn't take chances on three dropping into a 5 match

Want to firm up your results: If the board was like this at the top, would you take the blue match?




Or (again at top)




If you didn't take the match, did you at least neutralize the match?
Are you counting cascades that come from ability use differently?

In both cases I would have taken the Yellow match (mainly because of prioritization of ap for my team)
If they weren't yellow specifically I would have matched the black away depending on what else was on the board above/below it

No, the reasoning is that even though I am a patient person, I don't have the patience to play the same prologue mission 1000 times. It would have been a lot more data points to track if I counted post ability cascades. I would have had to post who the cascades happened for (ie, cyclops, Thor, Hulkbuster etc...) That was more data than I was willing to track at this time. I am considering only counting prologue fights but that would probably take me a month and none of the prologue missions are scaled high enough that I could really simulate a full length "normal" PvP/PvE match.

PeterGibbons316

Thor generates tiles of a certain color. That creates more cascades. Having a full health Thor means he is less likely to die, which means he is more likely to get cascades in his matches.

Using a health pack before a match has no bearing on the amount of cascades you will get, but going in to battle with a full health Thor will generate more cascades than going in to battle with a low health Thor.

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

I'm just going to spoiler this for the time being. In the future try and express your opinions without such condescension.

Unknown

PeterGibbons316 wrote:

Thor generates tiles of a certain color. That creates more cascades. Having a full health Thor means he is less likely to die, which means he is more likely to get cascades in his matches.

Using a health pack before a match has no bearing on the amount of cascades you will get, but going in to battle with a full health Thor will generate more cascades than going in to battle with a low health Thor.

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

At no point in my statement did I say "you should use healthpacks to get more cascades". I was asked to track that data and I did. This post is about the numbers I received in a 1000 game span. I am not advocating the use of healthpacks or one character over another. I am just presenting you with what happened, what I tracked and the things people asked me to do.

Unknown

traedoril wrote:

daibar wrote:

I didn't take chances on three dropping into a 5 match

Want to firm up your results: If the board was like this at the top, would you take the blue match?




Or (again at top)




If you didn't take the match, did you at least neutralize the match?
Are you counting cascades that come from ability use differently?

In both cases I would have taken the Yellow match (mainly because of prioritization of ap for my team)
If they weren't yellow specifically I would have matched the black away depending on what else was on the board above/below it

No, the reasoning is that even though I am a patient person, I don't have the patience to play the same prologue mission 1000 times. It would have been a lot more data points to track if I counted post ability cascades. I would have had to post who the cascades happened for (ie, cyclops, Thor, Hulkbuster etc...) That was more data than I was willing to track at this time. I am considering only counting prologue fights but that would probably take me a month and none of the prologue missions are scaled high enough that I could really simulate a full length "normal" PvP/PvE match.

I'm afraid that there's been a misunderstanding then. These 2 cases are times when you SHOULD take the risk; you have the advantage. Not taking it will lead to less cascades for you. The AI is perfectly willing to take this risk, so may get these cascades.

The difference between:



Or




AND




is that in the first 2, the odds of you getting the match 5 are greater than the odds of the AI getting the match 5. In only the bottom one, there's 2 places where the 5th matching tile could fall (above or below) as opposed to 1 for you.

grunth13

PeterGibbons316 wrote:

Thor generates tiles of a certain color. That creates more cascades. Having a full health Thor means he is less likely to die, which means he is more likely to get cascades in his matches.

Using a health pack before a match has no bearing on the amount of cascades you will get, but going in to battle with a full health Thor will generate more cascades than going in to battle with a low health Thor.

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

You were very condescending the first time around when the original test came through and you are again condescending with the new set of data points. The OP has repeatedly said that he was showing numbers for his test. That is how the scientific method works. He has not made any claims one way or the other. At least he is showing his work. The devs have never showed any of their work when they reply to the cascade issue. They simply come in and say "our tests show that its equal". Never have they came in and said how they did their test, who they used to do the test, in what environment the testing was done,etc. So cut the OP some slack as he is just sharing his info with some physical facts vs. the devs who have never shown their work. In the end, the devs may be right that there is not causality between healthpacks and cascades or player vs ai cascades numbers, but only time will tell. Many scientist believed the earth was flat until it was proven that it wasn't. Many doctors believed that leeches cured sickness until it was proven that it didn't. The OP is just furthering knowledge so that the next person can improve on his testing to MAYBE find the truth. Or the devs can just open up their code and let the player base look at it whenever they want so that we can confirm that everything is random.

shatland

PeterGibbons316 wrote:

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

you forgot to read this part of the OP:

traedoril wrote:

I am not trying to "prove" anything. I am just presenting you with the data I received. Please remember to take it with a grain of salt.

unwarranted, unneeded and unbelievable.

simonsez

PeterGibbons316 wrote:

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

You're sounding awfully antagonistic and defensive to a guy who's only presenting data. Typically reaction from someone when they encounter information that is counter to their world view.

Malcrof

traedoril wrote:

A.I. Used ability right before victory (the ai was on last character with 1-2 hits left)

A.I.
287

Are you including suicide abilities that kill the opponent? for example, a headbutt that kills Juggs, or a sunder that kills Ares, etc..

Unknown

Malcrof wrote:

traedoril wrote:

A.I. Used ability right before victory (the ai was on last character with 1-2 hits left)

A.I.
287

Are you including suicide abilities that kill the opponent? for example, a headbutt that kills Juggs, or a sunder that kills Ares, etc..

yes there were 4 total

Hulkbuster killed himself 2 times, Juggernaught once and ares once

scottee

I just read the original thread and this thread. The OP is 100% simply presenting their data. Then readers of said data are often interpreting it incorrectly. This is still in no way a scientific experiment because there are still too many decision factors that can't be controlled. The only true scientific test would be able to set up the exact same board state and run it repeatedly, sometimes with you making the match, and sometimes with the AI making the match, and to record the results. This is impossible to set up in the actual game, and a simulator would of course not give the results we want.

I think it's highly likely that most people's confirmation is caused by a simple part of the AI programming: The AI is better at looking for chain reactions than humans. The AI can evaluate exactly what the board state will be after a match. This would automatically lead to the AI having more cascades than the player, even if NO new tiles dropped from the top.

Denying cascades doesn't only mean making sure there are no match-4s or match-5s for the AI. It means looking if there's a match available on the known board that sets a chain reaction for 2, 3, or more matches to follow. If you can start to notice these matches, you'll get more cascades than the AI.

A central question that I think could improve a lot of players' games is, "Based on the known tiles, what will the board state look like after I make my move?" And, "What moves could the AI make?"

Advanced would be, "What is the probability that the unknown tiles that will drop will give the AI a favorable move?"

PeterGibbons316

grunth13 wrote:

PeterGibbons316 wrote:

Thor generates tiles of a certain color. That creates more cascades. Having a full health Thor means he is less likely to die, which means he is more likely to get cascades in his matches.

Using a health pack before a match has no bearing on the amount of cascades you will get, but going in to battle with a full health Thor will generate more cascades than going in to battle with a low health Thor.

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

You were very condescending the first time around when the original test came through and you are again condescending with the new set of data points. The OP has repeatedly said that he was showing numbers for his test. That is how the scientific method works. He has not made any claims one way or the other. At least he is showing his work. The devs have never showed any of their work when they reply to the cascade issue. They simply come in and say "our tests show that its equal". Never have they came in and said how they did their test, who they used to do the test, in what environment the testing was done,etc. So cut the OP some slack as he is just sharing his info with some physical facts vs. the devs who have never shown their work. In the end, the devs may be right that there is not causality between healthpacks and cascades or player vs ai cascades numbers, but only time will tell. Many scientist believed the earth was flat until it was proven that it wasn't. Many doctors believed that leeches cured sickness until it was proven that it didn't. The OP is just furthering knowledge so that the next person can improve on his testing to MAYBE find the truth. Or the devs can just open up their code and let the player base look at it whenever they want so that we can confirm that everything is random.

Sorry if I come off as condescending, it's probably because I think people who believe the game is rigged against them are ignorant, and people who still think the game is rigged against them after reading the countless discussions on the matter are just plain stupid. Doesn't sound like the OP fits this description, and good on him for trying to collect some data to answer a question. The only problem is that his data is meaningless without looking at other factors that impact his games (teams used, play style, etc.) All of these things were mentioned in the last thread and hardly any new factors were considered during this "study."

simonsez wrote:

PeterGibbons316 wrote:

Look, believe whatever conspiracies you want, use your health packs if you really think it is giving you an advantage. But don't try to convince the rest of us to waste our health packs on this nonsense.

You're sounding awfully antagonistic and defensive to a guy who's only presenting data. Typically reaction from someone when they encounter information that is counter to their world view.

I didn't quote the OP because my comment was not directed at him. It was meant for those who think the game is "rigged" against them. I am not being defensive as I have nothing to defend. Call it antagonistic if you want, I'm simply trying to prevent the spread of misinformation.

Unknown

Also, everytime someone screams the cascades are out of control:

http://www.smbc-comics.com/index.php?id=3752