D3 better step up & fix AI cheating ASAP

john1620b

I want to chime in here: I've never believed people about the AI using abilities without AP, but in the last PvP, it happened to me. I specifically looked at a Sentry team to make sure they couldn't Supernova me, and they had 7 red AP. So I took the last red match on the board, which cascaded into another red match. The very next turn (when the AI should have had 7 red AP still), I got hit with a Supernova.

It's possible it was a display issue, and they really did have more than 7 AP, but I've played long enough to know that it wasn't goons (because it was PvP) and it wasn't a Team-up (because Sentry was on the team). Unfortunately, I don't have any further proof of what happened, as I was not recording my screen at the time.

Unknown

Has anyone experienced this problem? You've got 2 goon counters at timers 1 and 2.
The timers go down to 0 and 1, setting the first one off.
However, after that happens, the countdown timer at 1 goes off too! (off by 1 error?)

I'm not sure if I'm imagining this, but I'm almost certain I've seen in at least 4 matches throughout my MPQ career.

Dayv

daibar wrote:

Has anyone experienced this problem? You've got 2 goon counters at timers 1 and 2.
The timers go down to 0 and 1, setting the first one off.
However, after that happens, the countdown timer at 1 goes off too! (off by 1 error?)

I'm not sure if I'm imagining this, but I'm almost certain I've seen in at least 4 matches throughout my MPQ career.

You know the countdown timers count down one at a time, right? If one is going off on a higher row, the one marked with a "1" further down still has it's turn to count to "0".

It takes each row, left to right, starting at the top, and processes countdowns one at a time.

gambl0r312

IceIX wrote:

NorthernPolarity wrote:

I think that the majority of those accounts can be explained by people not knowing that goons generate AP. I saw some posts like that of people wondering how Yelena was able to cast recon so quickly, without noticing that she had 2 purple goons generating AP for her. Feel free to post any examples that don't seem like this is the reason though: it'll be like Mythbusters MPQ Edition!

We've investigated quite a few of these reports internally and our results too are about the same. In each case, we discovered that it was a case of minions gathering AP and the user just not taking this into account. Admittedly, the mobile UI doesn't do the best job of constantly updating the player on opponent AP totals at current , but we haven't hit a certifiable situation where AP mysteriously gets generated after quite a lot of concerted testing on the matter.

Have I got a case for you then!

viewtopic.php?f=9&t=11814

I filed this months ago in the Bugs forum, and the circumstances were simple enough that I have no explanations for how IM40 fired that Recharge early. As you can see, it was simple enough for me to recall each turn, moves used on both sides, and AP totals of both players....please point me toward the correct explanation...

Also, no goons to auto-generate AP ( it was LRs )

Unknown

gambl0r312 wrote:

IceIX wrote:

NorthernPolarity wrote:

I think that the majority of those accounts can be explained by people not knowing that goons generate AP. I saw some posts like that of people wondering how Yelena was able to cast recon so quickly, without noticing that she had 2 purple goons generating AP for her. Feel free to post any examples that don't seem like this is the reason though: it'll be like Mythbusters MPQ Edition!

We've investigated quite a few of these reports internally and our results too are about the same. In each case, we discovered that it was a case of minions gathering AP and the user just not taking this into account. Admittedly, the mobile UI doesn't do the best job of constantly updating the player on opponent AP totals at current , but we haven't hit a certifiable situation where AP mysteriously gets generated after quite a lot of concerted testing on the matter.

Have I got a case for you then!

viewtopic.php?f=9&t=11814

I filed this months ago in the Bugs forum, and the circumstances were simple enough that I have no explanations for how IM40 fired that Recharge early. As you can see, it was simple enough for me to recall each turn, moves used on both sides, and AP totals of both players....please point me toward the correct explanation...

Also, no goons to auto-generate AP ( it was LRs )

I read your post of the other thread.

One explanation might be that AI might be equipping boosts? (Not sure about this, but it randomly hit me as it is the only way for them to have AP since move 0). Can the AI equip boosts along with TUs?

Another question would be: If I use boosts for my last fight in a PVP event, do those boosts stay in my defense team? Meaning that if you come to face me, the boosts that I equipped the last time will be used by my defense team. So if I used "+3 to all AP" boost I would start with 3 AP from move 0...

I would like some clarification on this because I have not found a clear answer on this yet.

EDIT: I had typed ISO instead of AP

Unknown

DayvBang wrote:

daibar wrote:

Has anyone experienced this problem? You've got 2 goon counters at timers 1 and 2.
The timers go down to 0 and 1, setting the first one off.
However, after that happens, the countdown timer at 1 goes off too! (off by 1 error?)

I'm not sure if I'm imagining this, but I'm almost certain I've seen in at least 4 matches throughout my MPQ career.

You know the countdown timers count down one at a time, right? If one is going off on a higher row, the one marked with a "1" further down still has it's turn to count to "0".

It takes each row, left to right, starting at the top, and processes countdowns one at a time.

Ah! So if the position of the countdown shifts, it could actually go down by 2 in one round? Thanks! I'll look for this next time I see the set up.

Unknown

I don't think that's what he meant, but that would sure be an interesting bug. I'm sure someone would have noticed that by now.

Dayv

Hydrus wrote:

I don't think that's what he meant, but that would sure be an interesting bug. I'm sure someone would have noticed that by now.

That's not what I meant, but with the right cascade, I imagine such a position shift could occur. It would be interesting to know the results of that.

Unknown

IceIX wrote:

NorthernPolarity wrote:

I think that the majority of those accounts can be explained by people not knowing that goons generate AP. I saw some posts like that of people wondering how Yelena was able to cast recon so quickly, without noticing that she had 2 purple goons generating AP for her. Feel free to post any examples that don't seem like this is the reason though: it'll be like Mythbusters MPQ Edition!

We've investigated quite a few of these reports internally and our results too are about the same. In each case, we discovered that it was a case of minions gathering AP and the user just not taking this into account. Admittedly, the mobile UI doesn't do the best job of constantly updating the player on opponent AP totals at current , but we haven't hit a certifiable situation where AP mysteriously gets generated after quite a lot of concerted testing on the matter.

Curse those AP generating minions! Curse them I say!

Kolence

DayvBang wrote:

Hydrus wrote:

I don't think that's what he meant, but that would sure be an interesting bug. I'm sure someone would have noticed that by now.

That's not what I meant, but with the right cascade, I imagine such a position shift could occur. It would be interesting to know the results of that.

It occurs exactly like that. If the second CD resolving in any way creates a cascade that causes the first CD (that has already reduced it's count that turn) to drop "lower and to the left" or "after" the position where the second CD was, its count is reduced once more that same turn.

DayvBang wrote:

It takes each row, left to right, starting at the top, and processes countdowns one at a time.

It can also happen with your own tiles. And if it happens to IM40 recharge tile, he will immediately stun himself and leave the other recharge tiles, if some are left, unresolved for that turn (plus next turns too since he stunned himself). I had it happen to me with IM40/MMN. Magneto's red CD exploded and dropped a 2nd recharge CD out of 6 that was already passed by the "CD checker" and caused IM40 to stun himself for 1 CD and leave other 5 with mismatched CD - the first at 1, and the rest at 2. Ruined a game almost! And I had 14 and 16 spent before that fiasco.

Avalokitesvara

I am really tired about this. Lately PVP has become unplayable, I must avoid certain characters (OBW, any thor, any ares), because they get insane cascade on their FIRST TURN by matching random 3 tiles.
Last one was ridiculous, I was with lv 104 patch, lv 114 magneto and lended colossus and a lv 65 OBW singlehandledly killed me with cascades in 2 turns. Please. It's insulting, I don't even have the will to keep playing right now as my best character have been killed so easily by a low level character.

Unknown

Avalokitesvara wrote:

I am really tired about this. Lately PVP has become unplayable, I must avoid certain characters (OBW, any thor, any ares), because they get insane cascade on their FIRST TURN by matching random 3 tiles.
Last one was ridiculous, I was with lv 104 patch, lv 114 magneto and lended colossus and a lv 65 OBW singlehandledly killed me with cascades in 2 turns. Please. It's insulting, I don't even have the will to keep playing right now as my best character have been killed so easily by a low level character.

You're not alone buddy. Most of us have suffered humiliating defeats at the hands of the AI, and still do. You just have to push your chips forward when playing this game knowing that you're going to get wiped sometimes. A top player in my alliance was commenting just today that his entire (45+) roster was wiped on a high level PVE node. You just have to laugh it off after a round on the punching bag and consider it a badge of courage moving forward. It's the bitter losses that make the wins in your future all that more sweet.

Great MPQ players are first forged in the fires of MPQ. It is this way as with all things. Thucydides, in his history of the Peloponnesian War (the Melian Dialogue) summed it up aptly: "the strong do what they can and the weak suffer what they must." This is the one few truisms of world affairs, and it is no less true of MPQ.

Never give up.

Pylgrim

What is really, really infuriating about a cascading AI is that in theory, it should get less/worse cascades than a player, but it's generally the opposite. I mean, the player is the one who can carefully study the board. He or she can make what would normally seem weaker matches that make the visible tiles fall in ways that creates a few more matches and from the added board shake and tiles removed, hope for more matches from the invisible tiles that will fall from above (which still, doesn't happen that much). On the other hand, the AI is dumb and goes for hard-programmed moves (such as making a 4-match if possible) ignoring much better moves that the character can see. And yet, the AI is rewarded often and ridiculously by the falling tiles. So it goes something like

"oooh a 4-match, must have! ooo what a coincidence, the new row that fell included a 3-match in my best colour that makes a 5-match with two tiles on the board! ooo what a coincidence, the resulting critical tile plus the new falling tiles made a 4-match that cleared the column! ooo what a coincidence, the new column included a fully formed 7-match in my best colour! ooo what a coincidence, the critical tile that fell to the bottom made a 4-match row on the bottom that cleared that special tile that you had tried to hide in a corner! ooo what a coincidence, all this board shake caused several 3, 4 and 5 matches that basically replace the whole board while dealing thousands of damage even with the reduced cascade damage! ooo what a coincidence, I collected enough AP to use Thunder Strike even though I had 0 yellow at the beginning of the turn! ooo what a coincidence, the resulting green tiles were "randomly" placed in such ways that created one green 4-match and two green 5-matches! Call the Storm! What's that? oh right, I still get to move! oooh a 4-match, must have! ooo what a coincidence................"

GuntherBlobel

Pylgrim wrote:

What is really, really infuriating about a cascading AI is that in theory, it should get less/worse cascades than a player, but it's generally the opposite. "

Ok great, you have a theory, but where is the data?

Forumites have collected huge amounts of data on Token draws that time, and time again, confirm D3's probability tables. Why should this be any different? Tracking cascades will be more difficult, but I bet that if you slow down, stop caring about your win/loss ratio, and dispassionately track your matches you will see this in a new light.

Pylgrim

GuntherBlobel wrote:

Pylgrim wrote:

What is really, really infuriating about a cascading AI is that in theory, it should get less/worse cascades than a player, but it's generally the opposite. "

Ok great, you have a theory, but where is the data?

Forumites have collected huge amounts of data on Token draws that time, and time again, confirm D3's probability tables. Why should this be any different? Tracking cascades will be more difficult, but I bet that if you slow down, stop caring about your win/loss ratio, and dispassionately track your matches you will see this in a new light.

I actually made this exercise for a couple weeks if only for my own satisfaction, since I am aware that it's anecdotal data. The results were something like

Cascades that generate 10 or less AP - Approximately same number for me than for the AI
Cascades that generate 10 to 20 AP - AI got slightly more than me
Cascades that generate 20 to 30 AP - AI got almost the double as I
Cascades that generate over 30 AP - AI got exactly three times as many as I.

Nevertheless, please point out what are the holes in my "theory". Are you implying that in theory the AI should get better cascades? How is this so?

Unknown

Weird. My evidence has shown the opposite: I get way more cascades than the AI, like a multiplier of 3-6X for those of 10AP or more. However my data is heavily biased. I tend to use strike tiles so aim for cascades, even several turns in advance. I also keep an eye on non-matching tiles so if there's a 1/7 chance that I'll make another match I'll take it over the same color match elsewhere. Also I actually match T-5 opportunities (as opposed to the AI), and crits usually lead to cascades.

Several memorable cascades for myself in the past few days: Made a vertical match which was then replaced by 3 twice in a row!
A 3-3 vertical match led to replacement of 6 vertical + 3 on the side, total of 9 team-up matched in one match! This led to further cascades.

The exceptions of course are when playing against or , or the combo. Those guys are cascade monsters.

Unknown

I just got through a game where I boosted green, first turn AI gets was Dormammu's Aid steal green, make a green match 3, 3 more green fall in and matched itself. Next turn World Rupture. I think that was a 'we're not even going to try to pretend this isn't rigged'. Yes 7 green AP is actually very tame but the way the whole sequence played out feels like someone forgot to even pretend it's fair! But good thing I had secretly obtained a copy of the AI 'homing special tile destroyer hack' and destroyed all but one of the CD before they went off, so the joke's on the AI!

You're going to have more big cascades for player in short games because you get the first move and it's usually pretty obvious where the best opportunity for those are. As the game drags on the computer tends to get more because they never miss a row/column match, and that they always gamble for the miracle cascade. That is, if you take a non row/column match 5 most likely the crit tile just matches 2 tiles immediately which stops anything interesting from happening. The memorable cascades is the stuff where you destroyed a row and 6 of the 8 tiles that fell are the same color and immediately matched the whole row... and then again! You can't get lucky if you don't gamble and while taking a non row/column 5 probably has better average turnout, it's unlikely to propagate very far because the crit tile does a good job at stopping a miracle cascade usually by immediately matching itself with any 2 of the same colors.

Spoit

Sure the AI might gamble more often, but I'd certainly hope that humans are more reliably able to trigger cascades in their numbers. Between actually being able to target targeted abilities rather than casting them randomly, and being able to actually see multi-level-matches, I'd be pretty concerned if the AI was getting anywhere near an even number
(by mutli-level matches, I mean like)

Unknown

Phantron wrote:

The memorable cascades is the stuff where you destroyed a row and 6 of the 8 tiles that fell are the same color and immediately matched the whole row... and then again! You can't get lucky if you don't gamble and while taking a non row/column 5 probably has better average turnout, it's unlikely to propagate very far because the crit tile does a good job at stopping a miracle cascade usually by immediately matching itself with any 2 of the same colors.

It's probably a matter of different perceptions, but I'm going to have to disagree with this one. A lot of the times a crit created by a match 5 creates a row/column clear with a match 4. Also, the crit can match in multi directions. I even had one instance where the crit created 4 match 3's and further cascades! Multi-matching a crit isn't that uncommon, and double crit matches often lead to further cascades due to the multiple directions and more overall tiles whose relative positions are affected.

Perhaps it's also OBW use -> mid-air crit matching which might make it less impressive? I don't use her as often.