Loop Prevention System - More Details *Updated (3/12/19)

Scottyp_123

I like to play with Naru Meha and Deploy the Gatewatch and while this doesn't exactly create and infinite loop it can take a very long time to resolve. By the time it does I usually have three creatures with 150/150 or more power/toughness.  I would love to just be able to tap the menu button and have an option to break the loop at will when I have enough to put my opponent down. This way I don't have an arbitrary timer cutting me off and I don't have to sit and wait for the loop detection to kick in. 

How about we use loop detection for the AI and a voluntary loop break option for players? 

IM_CARLOS

It would have been a disaster. But thanks to the pitchforks.... 

I wonder if there is something like beta tests for this game. 18s should be trigger every single match. But definitely no one of dev-team is playing this game, or have very different point of view of this game.

Luckily they changed it last minute. 

Now eyes on the bugs or unintended features. There is at least one gamebreaker to be fixed asap! 

Machine

ertaii said:

I'd like to add my thanks to the list for listening to us and increasing the time limit . I already ran into it a few times at 90 seconds. I can't imagine what It would have been at 18 seconds...

If they had listened, the LPS wouldn't exist anymore.

810Bourne

I just played a match. Greg played Bolus 1.  One turn took several minutes  I was forced to watch as Bolus ended one 90 second loop followed by a second 90 second loop. 


Each time a card was drawn  I took damage  

107 damage on one turn  that's a 39 draw turn  

arNero

This timer seems to have horribly failed in even trying to attempt to stop infinite loops..... BSZ into Thunderherd Migration into fatties into BSZ into Thunderherd into fatties into......

To be fair, I can't think of a more suitable-for-players solution right now, but still.....

Mburn7

arNero said:

This timer seems to have horribly failed in even trying to attempt to stop infinite loops..... BSZ into Thunderherd Migration into fatties into BSZ into Thunderherd into fatties into......

To be fair, I can't think of a more suitable-for-players solution right now, but still.....

The timer will eventually stop BSZ loops.  Just not before there is a ton of stuff on the board.
I'm pretty sure that's what we wanted with the timer, though (not that we wanted a timer at all, but whatever).  It won't stop you from winning, but it'll make it less boring to face.

Tremayne

@Brigby - the LPS system is not working!!!

I just finished a match where the LPS system kicked in at 25 minutes past and 10 minutes later the AI was still in the progress of finishing its turn - with no end in sight I might add!!!!

As far as I could tell it was just two cards that kept the loop running, but the deck included a third card which would prolong the loop (and was used to get the loop running).

I’ll be happy to send you the names of the cards, but I will not post them here, because that would most likely destroy several events.

This example shows the fallacy of your current solution, please reconsider your approach ASAP!!! The player base has already suggested several alternatives, if you insist on this approach. As far as I can gather the solution with the most control is based upon counting actions and then clear the action queue and move to the next phase.

Laeuftbeidir

@Tremayne I (and likely others as well) found out about how to work around the loop **** by accident. It's not the cards, it's surprisingly way more simple than that!

Tremayne

@Laeuftbeidir - great, but as far as I can determine the main reason for LPS is to avoid Greg running amok. That goal has not been achieved. The fact that you can circumvent it during your own match is just another nail in the coffin for this lousy implementation of LPS.

OmegaLolrus

Tremayne said:

@Laeuftbeidir - great, but as far as I can determine the main reason for LPS is to avoid Greg running amok. That goal has not been achieved. The fact that you can circumvent it during your own match is just another nail in the coffin for this lousy implementation of LPS.

So not only is a cruddy solution, it's also a broken one. Awesome.

810Bourne

Just put it back to 18 seconds. S3 with B/G/R just had way too many loops.

This game has turned to straight garbage 

arNero

Okay, I know what I'm saying will be shot down immediately, but I'll just say this once then abandon this thread since I've already said that I don't like the LPS system.

My personal idea of a solution to prevent the loops is to make BASIC CHANGES to the gameplay system, which in this case, limit the number of cards that can be cast on any given turn (such as, only 1 card can be cast on turn 1, 2 cards on turn 2 and so on until we give a hard limit of, say, 6 cards castable on turn 6 onwards, no matter how many million mana anyone gets). This helps stop loops coming from the more problematic card/combos such as Prism Array, Seasons Past, Omniscience and BSZ, although I admit it probably does nothing to Path of Multitudes drag.

I do admit this solutions have a LOT of hurdles aside from breaking apart the (already fragile) game codes as well as not a catch-all solution; namely, this hard limit will make it next to impossible to complete Speed objective (win in X or less turns), although in that case I might as well propose retire that objective permanently.

Gunmix25

Tremayne said:

@Brigby - the LPS system is not working!!!

I just finished a match where the LPS system kicked in at 25 minutes past and 10 minutes later the AI was still in the progress of finishing its turn - with no end in sight I might add!!!!

As far as I could tell it was just two cards that kept the loop running, but the deck included a third card which would prolong the loop (and was used to get the loop running).

I’ll be happy to send you the names of the cards, but I will not post them here, because that would most likely destroy several events.

This example shows the fallacy of your current solution, please reconsider your approach ASAP!!! The player base has already suggested several alternatives, if you insist on this approach. As far as I can gather the solution with the most control is based upon counting actions and then clear the action queue and move to the next phase.

I figured out how a timer can be extended. If the timer triggers and you with swaps left... if you swap a 5 match or more then the timer adds 10 seconds to your play time. In that time stuff can be cast and so on until it runs out and you need to resume finishing out swaps or the turn ends. If you continually make a5 gem match swap. The timer just goes dormant for each additional 5 gem swap. So if you swapped a lucky 5 gem swap 3 times in a row after the timer triggered you have an additional 90 seconds of play time to cast things. I know because I have done it a few times now. The first time it occured I was confused and thought... "oh a bug."

It's not a bug. After re-reading the notes I realized that is how it was designed, to be honest I thought it was clever. 

The many complaints I see in here I don't think are valid complaints. Call me out, flame me, whatever...  but this timer was to stop infinity loops, not to keep you from losing to someone's big combo. I think those combos are the lifeblood of this game. Infinity loops of course should be shut down. Having a Neha respawning and cast the one card that brings her back out over and over and over and over with no way to win the match or end the turn... that is what the timer is for.

Tremayne

@Gunmix25 - let me see if I understand your post correctly, you are saying that the timer is coded to prolong the time indefinitely under the right circumstances? 

Assuming that is correct, then please tell me - from a practical point of view - what is the difference between a loop that never ends and a loop that manage to prolong the timer indefinitely?

OmegaLolrus

arNero said:

Okay, I know what I'm saying will be shot down immediately, but I'll just say this once then abandon this thread since I've already said that I don't like the LPS system.

My personal idea of a solution to prevent the loops is to make BASIC CHANGES to the gameplay system, which in this case, limit the number of cards that can be cast on any given turn (such as, only 1 card can be cast on turn 1, 2 cards on turn 2 and so on until we give a hard limit of, say, 6 cards castable on turn 6 onwards, no matter how many million mana anyone gets). This helps stop loops coming from the more problematic card/combos such as Prism Array, Seasons Past, Omniscience and BSZ, although I admit it probably does nothing to Path of Multitudes drag.

I do admit this solutions have a LOT of hurdles aside from breaking apart the (already fragile) game codes as well as not a catch-all solution; namely, this hard limit will make it next to impossible to complete Speed objective (win in X or less turns), although in that case I might as well propose retire that objective permanently.

I like the idea better than the timer, certainly.

I would limit it to a card by card basis... though it would be essentially the same thing. If BSZ and Omniscience can only be cast once a turn, then... May have the same end result.

rafalele

arNero said:

Okay, I know what I'm saying will be shot down immediately, but I'll just say this once then abandon this thread since I've already said that I don't like the LPS system.

My personal idea of a solution to prevent the loops is to make BASIC CHANGES to the gameplay system, which in this case, limit the number of cards that can be cast on any given turn (such as, only 1 card can be cast on turn 1, 2 cards on turn 2 and so on until we give a hard limit of, say, 6 cards castable on turn 6 onwards, no matter how many million mana anyone gets). This helps stop loops coming from the more problematic card/combos such as Prism Array, Seasons Past, Omniscience and BSZ, although I admit it probably does nothing to Path of Multitudes drag.

I do admit this solutions have a LOT of hurdles aside from breaking apart the (already fragile) game codes as well as not a catch-all solution; namely, this hard limit will make it next to impossible to complete Speed objective (win in X or less turns), although in that case I might as well propose retire that objective permanently.

Excuse me, but this way is not Magic the Gathering at all.

Of course I do not like the timer.

Gunmix25

Tremayne said:

@Gunmix25 - let me see if I understand your post correctly, you are saying that the timer is coded to prolong the time indefinitely under the right circumstances? 

Assuming that is correct, then please tell me - from a practical point of view - what is the difference between a loop that never ends and a loop that manage to prolong the timer indefinitely?

Easy. 

If I understand your post correctly this is my answer. But I want to lay out a few things before I get to it. 

For sake of this
conversation:

Looping combos by definition are not infinite.
I.e. trostani, path to discovery, tokens, Zendikar's resurgeance. This loops for a long time,  but is not infinite. There is a win condition and the looping combo does end... it takes awhile, yes. But it does end.  

Infinite loop combos by definition are loops that continue in a repeated pattern in which the same two, sometimes three, cards are stuck being cast over and over and over with no win condition in place, no way to break the combo because it does not break the pattern to end the loop. You could leave this for 12 hours and it'll still be going.  That is an infinite loop.  

Swap: manually connecting gems into a chain. 5 gem swap counts as an additional swap. As do any cascades that result in a 5 gem chain from swapping. 

Conversion: any 5 gem chains do not count as an additional swap.  

Now my answer:

A loop that never ends is an infinity loop. The timer is built for that.  An infinity loop has no access to additional swaps because the loop denies the player (Greg or player) access to the battlefield to make a swap for a 5 gem connection to add time. Gem conversions do no count as additional swaps. The timer will eventually kick in to break this cycle this ending the turn.  But.... Should the player have extra swaps available due to stitch or having made a 5 gem swap prior the infinity loop kicking in, the turn doesn't exactly 'end'. If a 5 gem swap is available during this phase, the player or Greg could manually swap for that and gain 10 seconds to cast, draw ... or whatever.  If more than 1 extra swap is available,  for each 5 gem swap they make,  they could potentially add 10 seconds for each one to their turn before it officially ends. Please keep in mind that the likelihood of such a thing occurring is extremely slim.  


Prolonging a loop indefinitely manually is impossible. The level of luck would have to be insanely high. Conversions as you know,  do not add swaps. If a player or Greg recieved extra swaps after a long combo gets shut down (not an infinite combo) because the timer kicked in because of stitch. They could potentially add 10 seconds If a 5 gem swap was available. if there isn't one,  then Greg or player simply swaps out all his swaps and ends the turn.  If by lucky chance if a 5 gem swap was available every time a swap was made,  for discussion let's say say a crazy 3 of them were made,  then 30 seconds was added... in which you can do something. But when that ends and if you've no extra swaps... your turn ends. Your second half of the question requires an amazing amount of luck to prolong a long turn. This extra time does not prolong the combo that triggered the timer. That ended when the timer ran out and Greg or the player are forced to make another action if able... in most cases they cannot and the turn immediately ends with your creatures attacking and resolving said phase. So technically it isn't prolonging a loop and more prolonging a turn.

Hope this explains it. 


You posed a very good question. Thank you

OmegaLolrus

rafalele said:

arNero said:

Okay, I know what I'm saying will be shot down immediately, but I'll just say this once then abandon this thread since I've already said that I don't like the LPS system.

My personal idea of a solution to prevent the loops is to make BASIC CHANGES to the gameplay system, which in this case, limit the number of cards that can be cast on any given turn (such as, only 1 card can be cast on turn 1, 2 cards on turn 2 and so on until we give a hard limit of, say, 6 cards castable on turn 6 onwards, no matter how many million mana anyone gets). This helps stop loops coming from the more problematic card/combos such as Prism Array, Seasons Past, Omniscience and BSZ, although I admit it probably does nothing to Path of Multitudes drag.

I do admit this solutions have a LOT of hurdles aside from breaking apart the (already fragile) game codes as well as not a catch-all solution; namely, this hard limit will make it next to impossible to complete Speed objective (win in X or less turns), although in that case I might as well propose retire that objective permanently.

Excuse me, but this way is not Magic the Gathering at all.

Of course I do not like the timer.

How about this. We stick 4 of each card in your deck to build your library. If you go through those 40 cards, you lose. No more than 4 of a card per deck and drawing from an empty library is pretty Magic.

I'm being facetious, of course.

Yeah, Magic doesn't limit the number of cards you can cast in a turn, but it also has a hard limit of the number of cards in your deck. In a standard event, with your average deck, if you cast more than 60 cards in a turn, you lose.

As much as it draws and is inspired by Magic, this game isn't Magic. It has a lot of the same DNA, but they're going to have be different on some fundamental levels... otherwise, you're just looking at whatever the last Duels of the Planeswalkers game was.

ElfNeedsFood

I don't like that they missed the fact that the removed the nerf they put in place specifically for Naru, so we're back to trivial Naru loops in competitive standard matches...  Sure, the Quasi-Naru loop only goes on for (well a lot more than 30 seconds), but it's still just the kind of one-off fix that we need to see in lieu of a catch-all timer that re-enables this triviality...

Tremayne

@Gunmix25 - thank you for a well written reply. I’m happy you liked the question. 😄

I agree that the player should still die by a “big combo” (as you called it). However, if the LPS system fails to stop the action sequence so the big combo can deal lethal damage, then what is the point of LPS?

The Naru Meha / Sirens ruse loop is the basic example of the LPS system justification. This loop occurs on the outside and can be stopped by removing adding new cards to the action stack. “Outside” in this context means that card A cast card B and card B then cast new card A and so on, so the action stack only contains one action at a time.

Your answer is great at explaining the intention of the LPS system, when you only look at this kind of looping.

You failed to take into account that the loop can happen inside the action stack (and possibly elsewhere in the game), so 
Card A puts one action C on the action stack and the resolution of action C creates one (or more) action(s) D on the action stack.
Similarly, the resolution of one action D creates one (or more) action(s) C on the action stack.

Since I do not know the queue structure used to implement MTGPQ, I can only guess what now goes wrong, but from my actual experience my guess is that as follows card A creates 10 C actions, each of which creates 5 D actions, which each creates 10 C actions, which again creates 5 D actions which ...

So initiating the “loop” results in 5*10*5*10*5*10*5*10*5*10*5*10*5*10*5*10.... actions that must be resolved before proceeding to the phase where damage is actually dealt.

That takes forever and leaves the user with the pseudo-infinite action stack, that looks quite similar to an infinite loop, like NM/SR. In my view that is a flaw in the LPS system, which should be rectified. You may disagree.

I hope this clarifies things.

Caveat - Since I will not reveal the cards in question I can’t be more specific in my example above, but please trust the example even if it is a bit vague.