# Potions Explosions are hurting more than helping?

I’ve had the same feeling that it was targeted 4/5 gem matches, but also haven’t recorded any data. What I have noticed it that it doesn’t always remove these matches. I’ve seen in many cases where it explodes a gem away from one of these matches and it leaves the match intact.

That doesn’t invalidate this theory but that’s just my observation.

10 characters

The board is 8x8, there are 64 gems. A 6x6 area (36 gems) will explode the full 9 gems. The non-corner 24 gems along the border will explode 6 gems. The 4 corner gems will explode 4.

So the 9-gem case happens 56% of the time and explodes 14% of the board. On average, 7.56 gems or 11.8% is exploded at the start of a round. That already makes me itchy for the 4/5 match.

Let’s imagine the board having columns A-H and rows 1-8. If we imagine a 4-match spanning from, say, D3 to D6, we know one gem at D4 or D5 has to be in C or D in order to have the match setup. That means, roughly, 25 gems around its perimeter including the involved 4 gems could disrupt the match. So that’s close enough to 30 of our 64 gems, putting us at about a 46% chance of a vertical 4-match being disrupted.

Vertical 5-match is worse, it ends up being exactly 50%, as comfortable as I am using the word “exactly” here.

Horizontal matches are more of a nightmare in terms of probability. Imagine a 4-match set up with gems at {A1, B2, C1, D1}. If the explosion happens in any column from A through E, that match is disrupted. That’s 40/64 = 62.5% of all possible explosions. Ouch. For a horizontal 5 match in row 1, it gets worse. That’s 6 columns or 48 gems = 75% chance of disruption.

The probabilities per row are (floored, not rounding, sue me):

• 1: 0.625 | 0.75
• 2: 0.625 | 0.75
• 3: 0.546 | 0.656
• 4: 0.468 | 0.562
• 5: 0.390 | 0.468
• 6: 0.312 | 0.375
• 7: 0.234 | 0.281
• 8: 0.156 | 0.187

(These are slightly off because I forgot to account for a 3-gem hump around the offset if and only if it’s above the match. Deal. It’s worth about 2% for the lower 6 rows and I’m already throwing lots of error and estimation around.)

This is harder to read, here’s my takeaways:

• If you think 46% is “close enough” to 50%, 62% of all horizontal 5-matches are going to be disrupted at least 50% of the time.
• 50% of all horizontal 4-matches are going to be disrupted at least 50% of the time.
• A row 8 4 or 5 match is “close enough” to 15% if we’re conservative. That’s only 12.5% of all horizontal 4/5 matches.

So, breaking it all down:

• Vertical 4 and 5 matches are roughly 50% likely to be disrupted.
• Top row horizontal matches are more likely to be disrupted than not.
• Bottom-row horizontal matches are safest and will only be disrupted 15% of the time.
• On average, horizontal matches are something between 40% and 50% likely to be disrupted.

So technically explosions don’t “hurt more often than help”. But theoretically “they are almost equally likely to hurt as help, with a very strong bias towards not hurting horizontal matches towards the bottom rows, which are less than half of all possibilities”.

But.

I haven’t even started to try and model “What are the odds of a 4/5 match being FORMED by the explosion”. I don’t know what that case is, it’s probably too hard to model. But if you’re comfy with considering your odds at about 45% - 50% for “hurt”, and you think there’s even a 5% chance of “help” in the aftermath, then the overall odds of “explosion doesn’t hurt” and “explosion helps” together give you a positive outcome. Someone else can try to work that one out.

5 Likes

@Slypenslyde did you take into account adjacent Gems being selected for explosion? It kind of seems like it shouldn’t happen, but often the exploded area is a lot less due to overlap (I’m pretty sure, but tbh I don’t watch these in slow motion enough). It’s honestly a little disappointing for me when it happens.

On the whole, I’d agree though that 4 random gems exploding on the starting board have a decent chance of disrupting any existing 4+ matches (from an unreliable, intuitive statistical perspective ) – and I appreciate the breakdown you provided (actually helped me to conceptualise/analyse it a little!).

I also know the feeling of those matches being targeted it can definitely be infuriating seeing a beautiful L-match being destroyed to give you nothing, even with a Storm present. I tend to adopt 4.21’s approach of ignoring whatever starting board was present, but the inconsistency of Storms makes me a little bit sceptical sometimes that the Board turns out as it ‘should’ have.

Starting the match with a Storm is definitely the way to go, and Mithran’s take resonates with me – no Storm generally means a tiny bit of gained Mana and a slightly worse board; starting with a chosen Storm generally lands you in positive territory, occasionally getting a crap board, occasionally getting ridiculous amounts of Mana (whole team filled) thanks to cascades, esp. if using TPK.

For me, I guess the bottom line is that I want Storms to be more consistent and I’m general pro-optionality/choice in terms of AWR’s request to toggle potion effects. Iirc, the typical Dev response is to want to ‘fix’ the issue rather than allow conditionally positive (/negative) effects to be turned on/off (e.g. weapon upgrades and certain traits). On the whole, I don’t think it’s bad enough to warrant removal, but I do think it could be improved, and should be investigated (with details being reported back to assuage everyone’s concerns!!!).

Aren’t they the same odds as the starting board? As in, the odds of a starting board without potions having a 4/5 match vs. the odds of a starting board AFTER explosions having a 4/5 match imo would be exactly the same. In fact if you have a storm the odds would actually significantly increase after the explosion.

Whether the explosion ruins a 4/5 match is irrelevant.

2 Likes

Call it bad luck. I have seen more 4/5 combos removed by potion explosions than left alone, but I have seen them survive as well.

Actually it has been confirmed those odds are not the same.

The starting board is completely random. The gems that fall are subject to logic that cares about if it creates “too many” 4/5 matches.

Also, consider how 4/5 matches work. Imagine we have gems at { A4, C4, D4 }. On the starting board, we are ready for a 4/5 match if a same-colored gem is generated at either B3 or B5. But if an explosion is causing gems to fall towards that setup, the only possibility is if the gem falls to B3, because B5 is already set. And only explosions on row 1 or 2 and columns A, B, or C can create that. That’s 6/64 gems or only about 9% of the total explosions to begin with. The startup logic might pick B3 or B5. Since it gets 2 chances I inherently believe its rate is higher. (Note I’m excluding the case where the gem is already in B5 because that’s not “a near match” but “a match”.)

For an explosion to create a 4/5 match for you:

• The explosion has to be in the right place (I estimate probably about 12%-18% at best once verticals are accounted for).
• The gem has to be the right color (Roughly 1/8 = 12.5%, doomskulls make that a little wonky.)

That doesn’t suggest it’s very often. On the other hand, I don’t have a good idea for how to answer, “Well, what’s the rate of non-explosion boards that start with a 4/5 match?”. If I did, I’d also have an answer for, “What’s the rate of boards that have a near match that could be set up by an explosion?”

1 Like

Waaaaait I might have to re-run the numbers, or abandon them entirely. How many gems are exploded? I was only considering 1.

I see a suggestion of 4 and that’s a little too complex for me to model other than perhaps treating it like a distribution that will explode between 4 and 36 gems. (The cases being “all 4 explode the same corner” and “all 4 do not overlap”.)

Let’s say I go with the midpoint there and argue most of the time between 12 and 28 gems are being exploded. That’s higher than the 1-gem average of ~7 by a fairly wide margin, and still higher than the 9-gem case. So the odds of your match being disrupted will only get higher.

I used a vertical 4-match example where about 29 gems represented “a possible disruption” for about a 45% chance. I think (someone can fact-check this) the odds of 4 gems exploding and hitting 1 of 29 gems on the board is either:

• 1 - ((1 - 0.45)^4) = ~90%
• 1 - 0.55^4 = ~90%

But they work out to the same damning number so either I’m completely wrong or yes, explosions hurt more than help. That’s also the “no overlap” case, an upper bound. I’m not going to try to reckon it out but my gut tells me at best that lowers it to 75%?

I seem to recall that logic only kicking in after five consecutive extra turns taken, after which any further consecutive extra turn has a cumulative 20% chance of intentionally arranging gems in a mismatched pattern on board manipulations.

There’s another factor to consider. Even though the explosion might not set up an extra turn match for you to take, it could resolve any number of 3+ matches as part of the explosion cascades. Take the classic Skeleton Key team as an example, it just explodes one gem at the start, you may still get a surprising amount of mana, up to entirely filling Egg Thief.

I don’t think there’s any good mathematical approach to tackle this, there’s just too many factors and edge cases to take into account. A better approach might be to start 1000 explore battles instead, both with a standard team as well as a Sins of Maraj team that involves four Omen traits. Collect as many matches as you can without passing the turn on, note the total mana gained and move on to the next match. Sum up and compare how well each team fared, possibly broken down into blocks of 100. Picking teams that cover all mana colors might provide the best measurement results.

Thank you Slypenslyde for all the math. RNG is a big part of this game. It makes sense that a high middle match 4 will be disrupted at a high frequency while a low corner match 4 shouldn’t be disrupted. This is not what I’m seeing in game play. I’m seeing low corner match 4s being targeted by the single explosion. If we are strictly looking at where an explosion lands it should only directly hit a match 4 about 1/3 of the time. There are 20 of the 64 gems on the board that can cause a direct hit. This is not what I’m seeing. The match 4s are targeted.

1 Like

Is this a troll post? Lots of commotion over something that has cut event times in half.

6 Likes

Is that a troll comment?
You’re giving a tremendous amount of credit to 4 Explosions that at a max can generate 12 Mana between 4 troops if you have the correct Mana colors.
PS… You spelled discussion wrong.

Not a troll post
Just a post of “words that will get me in trouble” people

Each explosion can gain 4 mana x4 is 16 possible mana, plus any cascades and you still get to go first. With a storm your odds are actually higher at having a 4 match when your turn starts.

1 Like

I actually sort of agree the ~12 mana can be super useful. If you have a weapon like MC it’s not uncommon for it to be fully charged from a half-mana start on turn 1, that sets up a lot of other good things in your favor.

All in all I feel like post-potions I’m chewing through events a lot faster whether or not I get that first free turn. So many of my teams do so much better with that explosion.

1 Like

Just pick up a switch if you want a taste of the “old”. I play on both and switch is a crawl in comparison.

1 Like

I have a Switch… But my 2.5 years of experience is on my PC/mobile account. So though I appreciate your insight. Think before you speak (specially since… The Switch just got potions as well.)
4 Explosions is 24 potential gems. That’s the max possible. Despite usually my explosions only affecting 12 potential gems. I digress, explosions can only generate up up 50% Mana based on the gems they effect. Each explosion has a grid base of 6. And yeah it’s possible Cascades can help. It’s just as possible that Cascades can help my 4 or 5 extra turn that the Explosions took away.

No where in the OP or any comment am I suggesting the permanent removal. Let those who enjoy the game have all the explosions they could ever want. Just also let those who know how to play have the ability to turn them off when they themselves believe it’ll be strategically beneficial. Freedom of choice is always a good thing. Just remember boys and girls… That includes the freedom to be silent when you don’t do your research before weighing in on matters.

Most likely it starts with 6 Mana, then you have 2 Mana probably from enchant… So yeah if you have 8 brown gems exploded. Your instantly full.

Dude, it’s the internet. I posted yesterday about a time when I couldn’t use my fancy phone app to skip the checkout line at a grocery store because they didn’t put the produce code on the sign. In 10 minutes I had 8 downvotes, 4 people saying that never happens, and one person had actually driven to the store to take a picture of the sign to prove that I must have totally missed the code.

You could post that the sky is blue and I can already think of people pointing out:

• Actually the sky doesn’t have a color, you’re seeing light refracted via particles in the atmosphere.
• Actually at sunrise and sunset it’s different colors.
• Actually storms can also alter the color.
• Actually if you’re travelling at relativistic speeds the color will shift.

It doesn’t matter that none of that matters to anyone. What matters is you are WRONG on the INTERNET and someone’s going to get a single like for posting it.

I am playing switch as we speak. No potions. Also I dont know what explosions you are experiencing that only get 6 gems. Are yours only exploding 2/3 of the way? But as usual I am just stupid and ignorant. Oh yea and all my information is just incorrect and I dont know how to play.