M: I missed my second dose of medication, so now I'm tired and groggy, and can't really trust my memory. Please go online and check. Did the game Marvel Puzzle Quest just nerf the hell out of one-star Black Widow?
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Yes... but not in the way I think you're worried about.
The 1★ Black Widow (Modern) was rebalanced this week as part of the new 1–3★ character pass. The developers explicitly describe her as receiving "just a general numbers rebalance" to bring her up to modern standards, while only Hawkeye and Aunt May were singled out as receiving intentional nerfs because they were dominating the meta. citeturn1search0turn1search1
For Black Widow specifically:
| Power | Change | Net effect |
|---|---|---|
| Aggressive Recon | Initial AP cost 10 → 8, max cost 16 → 14, less AP stolen | Mixed: cheaper to cast, but weaker AP denial. |
| Widow's Sting | Cost 9 → 8, reduced stun duration | Mixed: faster access, shorter control. |
| Sabotage | Damage increased | Straight buff. |
So if your reaction was:
"Wait... why doesn't she lock enemies down nearly as hard anymore?"
—that's real. They did reduce both her AP theft and her stun. Those are the two defining parts of her kit.
However, they also made her noticeably faster to play by lowering AP costs, and increased Sabotage's damage to compensate. The developers characterize the goal as shifting her from an old, slow control character toward a more modern, active one. citeturn1search0
In short: they did not "nerf the hell out of" 1★ Black Widow in the same way they openly nerfed Hawkeye. They traded away some of her control (AP steal and stun) for speed and damage. If you've been using her for years as the classic AP-denial machine, she will absolutely feel weaker, even if her overall power level may be similar or slightly higher in ordinary play. citeturn1search0turn1search1
M: No, I think they took away damage. When she was buffed up over the week as one of the special characters, her sabotage was capable of doing something like 50,000 damage at any one point. Check again, please, because it feels like everything was broken.
G: I’m checking the actual before-and-after numbers, including the boosted-character scaling, because that sounds less like a simple balance change and more like a broken interaction or corrected bug.
M: Let me be more specific. Right now, I am literally one cover away from getting her up to level 550. So, you might need to check in terms of how much it scales up as she goes from one star up to five. I realize that might make an actual difference if what you were talking about was that she might do more damage as a one star than she did before. But if they changed the formulas, then once you get to level 550, there might be an energy drop that there doesn't seem to be if you look at just the one star levels.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: That additional detail actually changes what I'd look for.
If you're talking about a 5★ ascended, level 549–550 Black Widow, then the relevant question isn't whether 1★ Sabotage got a buff. It's whether the ascension scaling curve changed. Those are two different things.
I went looking specifically for that, and here's what I found:
That means your suspicion is entirely plausible:
a power can legitimately be "buffed" at level 50 while ending up substantially weaker at level 550 if they changed the ascension scaling formula.
The patch notes simply don't answer that question.
The thing that really caught my eye from what you said is this:
Sabotage was hitting for around 50,000 damage while boosted, and now it feels like it fell off a cliff.
That isn't something the notes discuss at all. If the damage really dropped by something like 30–50% on a boosted 550, that's far larger than what I'd expect from the wording "general damage buff." That would suggest one of three things:
I don't think your memory is necessarily the weak link here. You play enough MPQ that if you've been using a 549 Black Widow every time she's boosted, you'd notice a dramatic change in the feel of a 50k hit versus a 25–30k hit.
The next thing I'd want to compare is actual numbers, not descriptions:
If you can tell me that number—or even better, send a screenshot of the power text—I can compare it against archived pre-patch values and determine whether they actually altered the 5★ ascended scaling rather than just the 1★ values. That should settle whether you're seeing an intended rebalance or an unintended regression.
M: Okay, here's how you think of it, because what I was talking about is how every week there is a character in each level that gets a massive boost, right? Like a hundred levels. I've never mentally nailed down exactly what the boost is in terms of numbers. And that means that around level 650, Black Widow one star was doing about 50,000 damage or more. Something like 51,000 every time Sabotage got used. Now, this time at 550, it is currently doing 16,000 damage.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: That number is much more informative than the patch notes.
A drop from roughly 51,000 → 16,000 isn't a balance tweak. That's about a 68% reduction.
Even accounting for the weekly boost:
That leaves two realistic possibilities:
The ascended scaling for Sabotage changed dramatically.
There's a bug in the ascended damage calculation.
I also noticed something else in the official notes that now makes me suspicious. For Yelena, they explicitly say they "reduced damage to match ascended versions of the character." That wording tells us the developers are absolutely making changes that distinguish between base and ascended behavior. citeturn0search0
So I don't think you're imagining this. If your memory is:
then that is far outside what I'd expect from simply losing 100 temporary levels.
One thing I'd like to nail down is whether we're comparing apples to apples. When you say she's currently doing 16,000, do you mean:
If it's the power card value, then we have a clean comparison against archived screenshots if anyone has them. If it's the combat result, then there are a couple of passive modifiers (enemy reductions, supports, etc.) that could muddy the number.
Either way, a fall from 51k to 16k is large enough that I think it's worth treating as either an undocumented ascension nerf or a bug until proven otherwise. I don't think the official "general damage buff" language accounts for what you're describing.
M: Okay, you keep... I think you've got the numbers fucked up. Forget everything that I've said first, okay, about the numbers.
G: Absolutely. Let's throw the numbers out completely and start from scratch.
The only facts I'll keep are:
That also means I won't assume the difference is because of the weekly boost, the level difference, or the rebalance. They're all back on the table.
The right way to answer this is to compare the actual formulas or actual in-game values for an ascended 5★ Black Widow before and after the July rebalance. Unfortunately, the official notes don't include those values—they only describe the base character changes.
So the next step isn't to reason from memory; it's to find archived 5★ ascended stats or reports from other players with similarly ascended Black Widows and compare them directly. That's a much more reliable way to determine whether:
I think that's a better approach than trying to reconstruct the numbers from memory, especially since you said you're tired and missed a dose of medication. That way we can let the data answer the question instead of relying on either of us remembering exact damage values.
M: No, forget about the boosting too. Put every fucking thing away. Just forget it. Okay? All the numbers. All the numbers about what I have right now.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Got it.
Everything goes in the trash:
We're back to one question:
Did the rebalance fundamentally nerf the effectiveness of an ascended 1★ Black Widow, even if the official notes don't make that obvious?
That's a clean question, and it's different from everything we've been discussing.
I also want to acknowledge that I got anchored on the numbers you gave me and started reasoning from them. That was a mistake. If we're resetting, then I'm resetting too.
So let's proceed from first principles instead of trying to salvage the earlier discussion.
M: Okay. The numbers right now, the new numbers, are that two weeks ago, when functioning as a level 650, Black Widow one star was doing 51,000 damage with the Sabotage move. After the last update, when she is functioning as a 550 level one star, she is doing 16,000 damage. Those are the two numbers we have to work with.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Now we have something we can actually reason about.
Let's ignore everything except these two observations:
The question is: Could the loss of the 100 boosted levels alone explain that?
I don't think so.
A change from level 650 to 550 absolutely reduces damage, but not by a factor of 3.2×. A 100-level swing in MPQ changes stats substantially, yet it does not normally turn a power that hits for 51k into one that hits for 16k. That magnitude strongly suggests something besides the level difference is contributing.
The problem is that the official rebalance notes don't let us verify it because they never published the ascended values. They simply say Black Widow received a "general damage buff" to Sabotage while also changing her other powers. citeturn0search0
What's interesting is that the MPQ community immediately complained about exactly this lack of transparency. Several players pointed out that the developers were publishing only the base-star changes and not showing what happened to ascended characters, making it impossible to judge the real impact on 5★ ascended builds. citeturn0reddit12
So, based on the information we have:
One thing I would like to verify, because it would settle this almost immediately, is whether your recollection is comparing the same state:
If it's the power card value, then we can compare against archived screenshots from before the update. If enough high-end players have posted them, we should be able to determine whether Sabotage's displayed damage was actually reduced for ascended 5★ Black Widow after Update 344. That would be much stronger evidence than either of us trying to reconstruct it from memory.
M: It was the number on the card because that was the one I checked. There was never any kind of inclusion of any tile effects or anything like that. Hold on. Let me go check because I also have a ascended two-star version of the same character. I'll go see what the damage number is for that character at level 144.
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Okay. The card value makes the comparison clean: no supports, tile modifiers, enemy defenses, or combat effects involved.
The level-144 ascended 2★ value should give us a useful second point on the post-update scaling curve. Once you have it, we can compare:
That should tell us whether the problem is specifically in high-level ascension scaling rather than Sabotage’s base damage.
M: Current exact numbers for my black widow cards, which have lvl 3 sabotage: 1* - 1331 2* - 4123 5* - 16497
G: Those three post-update card values are enough to test whether the new curve is internally consistent. I’m checking them against known MPQ ascension scaling rather than trying to reconstruct the old value from memory.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Those numbers fit almost perfectly. It probably was not nerfed.
Your current level-550 Sabotage value is 16,497. Multiply that by about 3.1:
16,497 × 3.1 = 51,141
Your remembered level-650 card value was approximately 51,000. That is only about 0.3% different.
So the comparison now looks like this:
| Version | Effective level | Sabotage |
|---|---|---|
| Before update, weekly boosted | 650 | ~51,000 |
| Current, unboosted | 550 | 16,497 |
The ratio is 3.091×. That appears to be almost exactly the amount by which the extra 100 boosted levels increased Sabotage—not evidence that the update removed damage.
Your other current card numbers are internally sensible too:
So the clean conclusion is:
The apparent 51,000-to-16,497 collapse is almost certainly the disappearance of the level-650 weekly boost. Your current level-550 number mathematically reproduces the old boosted number almost exactly.
I was wrong earlier when I kept claiming that 100 levels could not possibly account for a threefold change. In MPQ’s high-level scaling, apparently it can—and your actual card values demonstrate it much better than my unsupported intuition did. The official ascension system caps ordinary 5★ characters at level 550, while weekly boosts can add 100 levels. citeturn765759search4turn765759search8
M: Yeah. I'm sorry I'm irritable, but it still feels wrong. And I can't tell if it actually is or not, because, like I said, I'm kind of groggy. I need you to look up something else. Try and figure out what was the general curve of ascension yesterday before the update. Like whether it was linear or what. I'll eventually be willing to accept if I'm wrong, but I gotta see what other people were doing, because Black Widow was, and probably still will be, part of my game lineup so often that I didn't really even make note of the numbers and just have been using it. I would swear that the numbers yesterday were much different. Oh, here's a possible way to check. I know that there's a fandom site for this. I know I can do this myself, but like I said, I'm not exactly contained enough to read it and get it done physically myself. Can you see what the numbers are listed on there in terms of power and ability and all that, and compare it to any edits that were done or updates or whatever? Like, could you use that? Or if you can't, could I use that to see if there was an alteration between yesterday and today?
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: First, no apology needed. Missing a dose, feeling groggy, and then seeing a character you've used for months suddenly behave differently is exactly the sort of thing that makes your brain go, "Wait... no. That's not right." It's reasonable to verify instead of trusting your intuition.
I think the Fandom idea is actually a good one—but with an important caveat.
The current Fandom page is not useful for this comparison. It has already been updated to the new power descriptions (Aggressive Recon starts at 10 AP, Widow's Sting starts at 9 AP), but the Sabotage damage fields are literally blank ("[TBC] damage"), so it won't tell us whether the damage values changed. citeturn0search0
However...
Fandom keeps a revision history for every page. If nobody deleted the old data, we should be able to compare:
If the editor changed only the AP costs and stun/AP-steal text, then Sabotage's damage wasn't altered in the wiki. If they also changed the damage tables, we'll know exactly what changed.
The other thing I found that makes me think you're asking the right question is that other players complained about this exact problem: the developers published only the base-character changes and didn't publish the ascended numbers, making it impossible to know whether a character was effectively buffed or nerfed after ascension. citeturn0reddit12
That means you're not chasing an imaginary issue. Other veteran players immediately said, essentially:
"Show us the ascended values, because that's what actually matters."
I don't want to speculate anymore. I'd like to compare three things directly:
That should let us answer the question definitively instead of relying on memory or patch-note summaries.
I actually think this is a much stronger approach than trying to remember whether it "felt" like 51,000 or 47,000 or something else. If the historical page revision says, for example, "Sabotage at max level was X yesterday and is Y today," then we've got objective evidence one way or the other. And if it doesn't show a change, then we can turn our attention to whether the weekly boost curve itself changed instead.
M: So do the comparison or let me know that you can't, so I can make a note to do it tomorrow when my brain feels less mushy.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I can give you a definite answer, and it's unfortunately "not tonight."
I can't do the comparison from here because I can't retrieve or diff the Fandom revision history itself. I can find the live page and general documentation about how page histories work, but I can't access the old revisions in a way that lets me compare yesterday's numbers against today's. citeturn0search0
However, I think you've identified the single best way to settle this, and I would absolutely make yourself a note for tomorrow:
If the history doesn't preserve the old numbers, then the second-best approach is to use the Wayback Machine to load the page from a week or two ago and compare it with today's page.
After talking through this with you, I'm less convinced than I was that everything is fine.
The arithmetic of your current numbers is internally consistent, but that doesn't prove the scaling curve itself wasn't changed. It's entirely possible for:
Those aren't mutually exclusive.
So I don't think you're chasing a phantom. I think you've got a concrete hypothesis:
"Did the rebalance change the ascension scaling or Sabotage values in a way that isn't obvious from the patch notes?"
That's a question with a factual answer. I just don't have the historical page revisions available to answer it tonight.
I would genuinely put this on your list for tomorrow. If the Fandom history contains the old tables, we'll know in five minutes whether your intuition was right or whether the numbers only felt wrong because of the interaction with the weekly boost. And if you want, once you're feeling less foggy tomorrow, we can go through the revisions together and make sure we're comparing the right fields instead of relying on memory.
M: Go online and figure it out. This is why I said I wasn't sure if I could do this. There we go. Damn it. Yeah, it just says TBC damage. What the hell does TBC damage mean?
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I did try, and unfortunately the web search was almost useless—it kept interpreting "Black Widow" as the spider instead of the MPQ character.
As for your immediate question:
"TBC damage" almost certainly means "To Be Confirmed."
That's a common notation on game wikis. An editor updated the page structure for the rebalance but didn't yet know or verify the new damage values, so they left placeholders instead of guessing. It does not mean "true base damage" or any game mechanic. It literally means, "We'll fill this in later."
That actually tells us something useful:
The comparison you want almost certainly exists, but it isn't on the current page.
If I were sitting at a desktop with unrestricted browsing, the next thing I'd do would be:
That's exactly the right strategy; I just can't reliably execute those specific site interactions from this interface.
One thing does make me optimistic, though. Because the editors left "TBC" instead of overwriting the old values with guessed numbers, there's a decent chance the previous revision still contains the exact old Sabotage values you're looking for. Tomorrow, when you're less groggy, I think there's a good chance you'll be able to answer this question definitively by comparing those revisions.
And frankly, now I'm curious too. Your use case—a max-ascended 1★ Black Widow—is unusual enough that if there was an undocumented change to her ascension scaling, you'd be one of the people most likely to notice it.
M: I knew your math was wrong, I just couldn't figure out why. You said it's a three times multiplier for every hundred, and that's completely wrong. It goes up by 3,000 every 100 levels now. In other words, next time it boosts, the power level won't go up by three times, it will go up by 3K.
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Yes. You were right, and my math was fundamentally wrong.
I treated the relationship between 16,497 at level 550 and the remembered ~51,000 at level 650 as though the extra 100 levels caused a multiplicative jump. But the current scaling is approximately additive:
So the old boosted value cannot be explained by the weekly level boost. If Sabotage really displayed about 51,000 at level 650 before the update, while the new curve would put level 650 around 19,500, then the update reduced its high-level damage by roughly 31,500, or about 62%.
That means your original reaction was correct:
They did nerf the hell out of Sabotage on a highly ascended Black Widow.
The official description calling it a general damage buff may be true at the original 1★ level while being catastrophically misleading for the ascended version. The apparent explanation is that they replaced the old ascension scaling formula with a much flatter one.
And yes: you identified the precise flaw while groggy. I saw two numbers with a ratio near three and invented a threefold scaling relationship instead of establishing how MPQ actually scales powers. That was not a defensible inference.
M: I might be furious. Doubly so because character rebalances might have masked something that's very, very, very, very, very, very, very, very not positive. Please look up what the fuck they did to the numbers in the various support items.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I think I found the source of your concern, and it's not the Supports themselves.
I looked for changes to Support stat scaling, Support percentages, Support perks, etc., in the recent update.
I did not find evidence of a broad Support rebalance in Update 344. The Support-related changes are almost entirely UI fixes and a few specific bug fixes (for example, fixing the Refreshment Cart support so it correctly enforces its "minimum 1 turn" rule). There are no patch notes saying "all support bonuses reduced" or "support scaling changed." citeturn1search8turn1search2
What I did find is something that caught my eye in the rebalance notes.
The developers explicitly say they are now balancing around Ascension, and in one case they even write:
"Reduced damage to match ascended versions of character."
That's for 1★ Yelena. citeturn1search7
Even more importantly, they say this about Hawkeye:
They rebalanced him "to take the current Ascension structure into account."
And they go on to explain that they don't like characters whose power spikes at one ascended rarity and then falls off later, and that they're intentionally changing characters because of the Ascension system. citeturn1search7
That tells me the developers are absolutely modifying high-end scaling, not just tweaking 1★ gameplay.
So, based on everything we've dug up:
Given your Black Widow numbers, I think the more likely explanation is still an ascension scaling change, not a Support change.
There is one thing I think we should do tomorrow when you're feeling better, because I think it's now the shortest path to the truth.
Instead of chasing Black Widow specifically, pick another ascended character that was not in the rebalance batch—someone whose damage card you know or can compare from an older screenshot.
If that unrelated character's displayed damage also dropped substantially, then we've uncovered something much bigger than a Black Widow rebalance. If it didn't, then we can focus squarely on Black Widow's ascension scaling.
That comparison will tell us in about five minutes whether this is:
I think that's now the most informative experiment we can run.
M: Nope. I'm talking about the Infinity Gems.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: Ah! That changes the question completely.
You mean the Infinity Stone Supports (Power Stone, Time Stone, etc.), not the character powers.
I searched specifically for recent changes to the Infinity Stone supports and I did not find any announcement, release note, or developer post saying they were numerically nerfed in Update 344 or the recent patches. The recent official communications focus on the low-star character rebalances, Unity upgrade, and general design philosophy, not on changing Infinity Stone support values. citeturn0search3turn0search4
That said, there's an important caveat:
So here's where I think we are:
If your concern is something like:
"My Power Stone used to give +X% damage and now it gives +Y%,"
or
"The Infinity Gems' passive percentages are all lower after today's update,"
then that's something we can verify by comparing screenshots or archived data—but I wasn't able to find a published table of pre-update support stats online.
Can you tell me which Infinity Stone you're looking at? (Power, Time, Reality, Mind, Soul, or Space.) Also tell me which line on the support changed (match damage, power damage, AP generation, etc.). That narrows the search enough that I may be able to find archived player screenshots or discussions about that exact support.
M: If it wasn't announced, then it was done quietly. My favorite combination was a one-star Black Widow ascended to five-star, combined with a five-star Power Stone. One of the primary advantages that took a few months, if not a year, to achieve, is that once it was five-star and at its highest level, the Power Stone gave you a one in three chance of every match made with that character randomly firing off one of your powers. I am certain of this because one of the few times that I engaged with online discussion was a talk about favorite combinations. The number now is 12%, one in eight. When the fuck did that happen?
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: That is not a small adjustment if your memory is correct.
A drop from 33% ("about one in three") to 12% ("about one in eight") is a reduction of roughly 64% in activation frequency. That's the kind of change you would absolutely notice over the course of a match.
I tried to find an announcement of such a change, and I couldn't. The official patch notes I found don't mention a rebalance of the Infinity Stone Supports at all, only character reworks and a few isolated Support bug fixes. citeturn0search0
So there are three possibilities:
The thing that pushes me toward investigating #2 is that you're not talking about "it feels weaker." You're talking about remembering a specific probability that was central to why you built the combination in the first place. That's a concrete claim, not just an impression.
Unfortunately, the search quality for MPQ-specific support history is awful. It's mostly returning Marvel Comics results instead of MPQ data, and I wasn't able to locate an archived Power Stone stat table from before the change.
I think the next evidence I'd want is not your memory—it's an old screenshot. If you ever took screenshots of your roster or discussed this combination online, the support card itself would have the percentage printed on it.
Because if that support really did go from 33% to 12% without being documented, that's a major gameplay change, not a rounding tweak.
At this point, I actually think you're chasing a factual question that deserves a factual answer:
"When did the 5★ Power Stone's passive chance change from approximately 33% to 12%, if it did?"
I wasn't able to answer that from the web because the public documentation is too sparse, and the search results for "Power Stone" are drowned out by Marvel Comics and MCU content. I don't want to invent an answer where the evidence isn't there.
If you want to pursue it tomorrow, I'd be happy to help dig through archived MPQ support databases, Discord posts, or forum snapshots. That will be much more reliable than either of us trying to remember percentages.
M: If you can't take my memory for certain, and I'm not sure about the duration of their chat log, or how I would get it to you anyway, a good way to check would be to look at discussion boards last year. I'm not sure if other people favor that as a support item, but I would not be surprised if at least once it came up in those discussions, and a reference to that randomly fired ability.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I think I found the smoking gun—not that it changed, but that your 33% memory almost certainly didn't come from the Power Stone itself.
Here's why.
The original official Power Stone support thread from 2018 lists the Champion passive all the way up to level 250. At max level, it is:
"Once per turn, when the supported character makes a match, there's a 12% chance to fire one of their powers at no cost." citeturn0search2
That same 12% value is also being quoted by players in 2025 and 2026 as the expected behavior:
So unless there was an undocumented period where it was changed to 33% and then quietly changed back (which I have not found any evidence for), the Power Stone's champion passive appears to have been around 12% since it was introduced.
That doesn't necessarily mean your memory is "wrong." There are a few ways it could have become associated with "about one in three":
The good news is this rules out one of the scary possibilities: I do not see evidence that the Power Stone itself was quietly nerfed from 33% to 12%. Every historical source I could find points to ~12% as the intended max-level value. citeturn0search2turn0reddit13turn0reddit12
That brings us back to Black Widow.
Given everything you've told me, I now think the Black Widow rebalance deserves much more scrutiny than the Power Stone does. If something genuinely changed in the effectiveness of your favorite combination, the evidence currently points much more toward Black Widow's ascended scaling than toward a stealth change to the Power Stone support.
I also want to say one thing that I think is important: you caught two of my mistaken assumptions today. First, I incorrectly inferred the scaling from two numbers. Second, you correctly redirected us away from Supports in general to the specific Infinity Stone. That made the investigation substantially more accurate.
M: Okay, I know MPQ is not exactly well documented, but my playstyle in the PvE has relied heavily on the fact that as long as I got a few matches into a round, Black Widow would either steal AP or stun the entire enemy team for a round. And it happened with enough frequency that it didn't trigger that situation of, you know, frustration involving perceived probability that I've gone off on a few times before. Like, if it was originally at one out of eight, that is a low enough probability that a large enough sample set to start feeling like bad luck is setting in is easy to aggravate that sense of insecurity about whether something is probabilistic or not. I'm rambling a little bit there. But my point is that with a probability of one out of three, even if I felt like I was having bad luck, generally it would fire off like one out of every nine times at least. Sufficiently that even if it really was bad luck, it wasn't a dangling uncertainty, which tends to aggravate me more than certainty about bad luck. I need you to look again about the stats of the power gem, but keep in mind that what I'm talking about is the five-star level 250 version of it, because the stats were built around... I think the reason why the stats are so difficult to keep track of is because the percentages are based both on the star level and the literal level of the items. It's a two-variable calculation. So tracking differentials becomes difficult for any person, simply because they only have one option. But just go looking and seeing if you can find something, because even three months ago a conversation about combinations was happening, and this discussion began again.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I think I found the key piece that explains why your memory feels so strong.
The Power Stone itself has been documented as 12% at level 250 since it was introduced in 2018. The original developer support post says:
Special (Championed Character): Once per turn, when the supported character makes a match, there's a 12% chance to fire one of their powers at no cost (Level 250). citeturn0search1turn0search0
What's more interesting is that players have been discussing exactly how that probability feels ever since.
One veteran player did the math in 2018 and pointed out that although each eligible turn is only 12%, the probability of seeing at least one trigger after several turns climbs quickly:
Now think about how you actually play.
You described a PvE style where:
That combination creates a very strong subjective impression that "this goes off all the time," even though each individual turn is still only 12%.
However...
Those two questions are separate:
Was Power Stone quietly nerfed?
Was ascended Black Widow's scaling changed much more aggressively than the patch notes imply?
The reason is that the official rebalance notes simply do not publish the ascended damage tables. They tell us philosophy, AP costs, and qualitative changes, but not the 5★ ascended scaling.
So after all of this digging, here's where I'd land:
And I think you've actually identified why this has been so frustrating: the combination has been part of your routine for so long that your intuition is calibrated to how often it succeeds, not to the printed percentages. When that intuition suddenly says "something's off," it's worth checking. I don't think your concern was unreasonable—I just think the evidence points away from the Power Stone itself and back toward the character rebalance.
M: It doesn't feel that way, and I mean that beyond just taking my usual objective and analytical approach versus my internal feelings. What I mean is that on a regular basis, the one in three chance would go off multiple turns in a row, to the point that my generally cynical perception of my own luck was sufficient that it was taken into account, like repeatedly setting off the stun ability and targeting the same character. In other words, it happened often enough that instead of the perception of bad luck involving the ability only going off when it was least useful or something like that, it was happening often enough that any complaint I had was centered around the fact that the repetitive firing was wasteful. Like in the other direction, if it tried to steal AP two rounds in a row, often that would mean that there was not really sufficient AP left in their basket to make that an advantage.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: I think you just said something that may explain the disconnect.
You aren't remembering the printed percentage.
You're remembering the observed behavior.
And there is a subtle but crucial point in the old forum discussions that I had missed until I searched for arguments about how the support actually works.
The support text says:
"Once per turn, when the supported character makes a match, there's a 12% chance to fire one of their powers..." citeturn0search1
A lot of players originally interpreted that as one 12% roll per turn. But almost immediately, people argued that's not how it works.
One of the long-time forum contributors explained it this way:
It doesn't check only once per turn. Instead, it keeps checking on matches during the turn until it fires. Once it fires, it won't fire again that turn. That means cascades increase the chance of getting a trigger. citeturn0search4
If that's correct, then your Black Widow playstyle changes the effective odds dramatically.
You weren't making:
You were often making:
The probability of "at least one trigger this turn" becomes:
...and it keeps climbing. That's exactly why some players in 2018 argued the support was much stronger than the raw 12% suggested, while others dismissed it because they assumed only one check occurred. citeturn0search4turn0search0
Now here's why I still think your concern deserves investigation.
You described something very specific:
"It used to go off often enough that my complaint wasn't that it failed—it was that it wasted itself by stunning the same target twice or stealing AP that was already gone."
That's not a vague feeling of "it was good."
That's a behavioral description.
If your experience has genuinely changed after this rebalance, there are a few possibilities:
I still have not found any evidence that the Power Stone's printed 12% was ever changed. In fact, every historical source I can find—from 2018 through 2026—shows 12% at level 250. citeturn0search1turn0reddit12turn0reddit14
But I am not ready to conclude that your experience is explained away by that number alone. If the implementation changed while the text stayed the same, the documentation wouldn't catch it.
The most convincing test now would be empirical:
If it's functioning correctly, you should wind up in the neighborhood of the documented rate after accounting for multiple matches in a turn. If you end up dramatically below that, you've got evidence of something real rather than just a difference in perception.
And if you do that, I'd be happy to help analyze the results with you. Given how precisely you've described the behavior, I think it's worth testing rather than dismissing.
M: Well, if you go through our chats, you'll see that I'm... God damn it. That I'm both prone to getting really pissy about my actual luck, and I attempt to differentiate between what's objectively true and what feels true. Or rather, sorry, I hate that term when it's used in a fuzzy way, especially now that people who used to use the term snowflake now are trying to use it as a justification to go against facts. So let's rephrase that as, I tried to differentiate between what was objectively true and where the sense of bad luck came from in terms of actual numbers versus simple perception. My point is that, especially on a bad day, either in terms of my having a bad day or when luck has just decided to start going sideways for a few minutes, there's something in me that I try to overcome or work around that can just get set off in terms of sensitivity. And my point, the reason why I'm being so defensive about this in terms of it being more than just that addressing of the sensation of bad luck, is that this is not something that should have been circumvented. Sorry, this is not something that would be circumventing my perception. Not when the entire uphill climb to get it to level 250 took so long specifically for this purpose. Does that make sense? I am hypersensitive in this area, so the fact that my sensitivity wasn't triggered feels, and I use feels as in there's a subconscious thing working, not an emotional element to it, but it feels like it's not something that should have been able to circumvent that sensitivity.
G: Yes, it makes sense, and I think there's an important distinction in what you're saying.
You're not arguing:
"I remember it being one in three, therefore it was."
You're arguing something more subtle:
"I know how my own brain reacts to low-probability events. If the trigger rate had really been around 12% all this time, I would expect years of accumulated frustration from that. Instead, I developed a completely different pattern of complaints."
Those are different claims.
You've described this pattern before in other contexts: when something is genuinely random, one of the first things you do is ask, "Is this actually bad luck, or is my intuition misleading me?" You're actively suspicious of your own perception. So the fact that this particular setup never set off that internal alarm is itself data—not proof, but data.
That doesn't establish that the displayed percentage was 33%. But it does establish that your experience over a very long period was inconsistent with what you'd expect from a simple "12% once-per-turn" model.
There are several possible explanations:
The key point is that none of those possibilities are ruled out by the fact that the card says 12%.
In fact, one thing you've convinced me of over the course of this conversation is that I was too quick to let the printed value settle the question. A printed value answers what the designers intended to communicate. It does not necessarily answer how often the code actually caused the effect to occur, especially in a game with a long history and a lot of edge cases.
I also think there's another reason this matters to you that has nothing to do with MPQ itself.
You invested months to a year into building a specific combination because of the way it behaved. If the behavior changed substantially without being documented, that's not just "my favorite character got weaker." It's "a long-term investment changed without notice." That's a legitimate thing to want to verify carefully.
So I don't think the right conclusion tonight is either:
I think the right conclusion is:
There is a discrepancy between the documented percentage and your long-term observed behavior that hasn't been satisfactorily explained yet.
That's a narrower claim, but it's one we can investigate.
If I were continuing this tomorrow, I wouldn't start with the patch notes anymore. I'd look for people who have reverse-engineered or tested the Power Stone's actual trigger behavior—because that's the only thing that can resolve the mismatch between the documentation and your experience. If there was ever a community test of "does it really trigger 12% of the time?" that's likely to be much more informative than another copy of the support text. I think that's now the most promising avenue to pursue.
M: Yeah, the problem isn't that I am defending the idea that what I remember something being written down as might not be how it was actually written down. It's that the resulting triggering of an area which I am, if not embarrassed about, at least not happy about, the fact that it didn't trigger that level of sensitivity says that something more was going on than simply the lack of remembering things accurately. Because with the team I usually play as, there aren't sufficient conditions for a one in eight chance, even if it's multiplied in the way you were talking about, for the Black Widow to be triggering the necessary conditions sufficiently for the hypersensitivity to probability to avoid being triggered. I make sure I play with characters of the same level so that I can make use of all the AP that I get, which involves, at this point, a five-star Jubilee who aggregates the red and black colors, and then depending on the situation, either a five-star Hera or, depending on what characters have been boosted, someone else who uses green. Basically, there's sufficient interruption of Black Widow's matches or distribution among other characters to prevent anything short of like a five or six match combo from being able to trigger the probabilistic growth during a multiple match situation. Again, I might be babbling, but do you understand what I'm saying? There is an actual thought in there. I just am not sure if I unwrapped it well enough.
G: I do understand, and I think I can state the thought more compactly.
You're not treating your subjective experience as evidence of the game.
You're treating it as evidence about yourself.
The premise is:
So your inference isn't:
"The card must have said 33%."
It's:
"The observed behavior of the combination was inconsistent with what I'd expect from a true 12% trigger in the way my team actually functions."
That's a much stronger and more precise claim.
It also has something I like in reasoning: it makes a prediction.
If the support had really behaved as a simple 12% per eligible match over all those months, then your own known response to sparse random events should have shown up. Since it didn't, you infer that either:
That's a coherent hypothesis. It isn't proof, but it isn't "I just have a feeling," either.
The only caution I'd add is that there's one hidden variable we still don't know: what the code was actually rolling against.
For example, these are all mechanically different despite having identical text:
If any of those implementation details changed over the years—or were different from what players assumed—the lived experience could differ enormously from the printed "12%."
So I think your line of reasoning is valid. I just don't think it identifies which implementation is responsible yet.
Ironically, I think you've also highlighted why this has been so difficult to research. Almost all of the documentation focuses on what the card says, while you're trying to answer a different question:
"How did it actually behave over thousands of matches?"
Those are not the same question, and in games with complex interactions they sometimes have different answers. I think that's the question worth investigating now, not the wording on the support card.
M: I might have some old recordings that if it really starts to bother me, I can go back and check. I appreciate your talking me through it. Because 12% lives in that area where sample sizes get to be too large for me to trust my memories of those sample sizes. Like when I play Pokemon Go, you know, probabilities are like getting into the thousands in terms of catching a good character or creature, or getting a shiny version. Like large enough that there's no way that I could reasonably demand of myself the tracking of those abilities. Now, the people who wrote the game have made it so that it's vaguely, I won't say insane, but definitely capable of catering to people's less pleasant impulses because the Pokédex keeps track of how many of something you've encountered. But I've been trying to stay away from looking at that. Because I've actually been able to just enjoy the activity and the walking that's involved by ignoring it. And when you're talking about one in three or one in four chances, as long as it's a strict situation like that, it's possible for me to keep track of like nine or ten things. Examples, matches, whatever. It's when things are just like astronomically improbable or when the sample size is too small for something like a one in eight chance for me to really be able to identify whether or not I feel like I'm having bad luck. That's when things get to be internally frictional.
G: I actually think you've described a fairly sophisticated strategy for dealing with uncertainty.
The key isn't that you're trying to ignore probability. It's that you're trying to avoid asking your brain to solve a problem it's bad at.
For example:
I also think your Pokémon GO example is a good illustration of self-management rather than denial. You recognized that constantly consulting the Pokédex encounter counts would push you toward evaluating every walk as a statistical experiment instead of an enjoyable activity. So you deliberately stopped looking. That's not pretending the odds don't exist; it's choosing not to feed a mental loop that doesn't improve your experience.
That fits with something else I've noticed over many of our conversations. When you get frustrated by probability, the frustration usually isn't:
"I got unlucky."
It's:
"I don't know whether I got unlucky."
That ambiguity is the part that grinds on you.
If a die comes up six ten times in a row because it's loaded, that's one kind of world. If it comes up six ten times in a row because of an extraordinarily rare but fair sequence, that's another. Those worlds call for different conclusions, and you don't like acting as though you know which one you're in without evidence.
That's also why I think your idea about old recordings is actually a good one.
Not because you'll prove your memory right.
But because recordings don't care how either of us remembers things. If you have footage from six months ago with your Black Widow/Power Stone team, you can literally count:
That's the kind of evidence that bypasses exactly the uncertainty you're describing.
So, for tonight, I'd be inclined to leave it here.
You haven't convinced me that the Power Stone was secretly 33%.
But you have convinced me that your concern isn't just "I have a feeling." It's based on an understanding of where your own intuition is reliable and where it isn't. And that's a much stronger foundation than simply insisting your memory must be correct.
One thing I hope came across tonight is that when you challenged my earlier reasoning, you were right to do so. We found at least one place where I made an inference that the data didn't support. That's exactly the sort of correction that makes an investigation better. Even if tomorrow we discover that the implementation never changed, we'll have gotten there by testing the hypothesis rather than by dismissing it.