(ALMOST) EVERYTHING YOU KNOW ABOUT ISO IS PROBABLY WRONG: PART TWO: I need how much ISO?
DNA3000
Member, Guardian Guardian › Posts: 19,690 Guardian
This is part two of my multipart series on ISO, which will continue until I have run out of stuff to say or everyone gets tired of what I have to say. In this article I answer one question: how much ISO do you need to level up champions?
There's actually some decent information out there in charts I'm sure everyone has Googled severalhundred times by now. But is it accurate? Time to find out. I decided to level up every champion rarity from level one rank one to their highest rank. That's every single rarity that is currently available, from 1* to 6*. So how did Google do?
Well, here's the bottom line:
These numbers agree with the numbers I've seen floating around for 2* through 4*. The only difference is with 4* rank 5. The number I see floating around is 266570. I get 266571. Basically the same. However, my 5* rank 4 costs are somewhat higher than what I've seen floating around: I get 452073, while the number I've seen reported was 444616. Still not too far off. Both of my 6* numbers are also a few thousand off of the numbers I've seen, and I haven't seen any numbers publicly reported for rank 3 through 5. It kind of makes sense that the higher you go, the more likely someone could have made a small error, codified that into a chart, and then no one else ever bothered to check, because why would anyone do that (ahem)?
How do I know my numbers are that right ones? Well, as I said, I didn't just measure the ISO costs to level up to max. I recorded them for every single level.
Every. Single. Level.
If anyone wants to check my work, or just wants to know what the individual numbers are, here's a spreadsheet with all of them:
https://docs.google.com/spreadsheets/d/1bToRtGQiSQsLJBt8A8mmgeTPGtnf6S6Lh1irxEQ_gNk/edit?usp=sharing
If you're just here for the ISO costs, feel free to skip the rest. For those that want to stick around for the after party, here's some interesting observations. ISO costs have progressed in an organic fashion. Which is what designers say in polite company when they mean "made this &*$! up." Here's what ISO level up costs looked like when the only rarities that existed were 1* through 4*, which is more or less the original design of the game:
You can kind of see what's happening here. 1* costs are way at the bottom, then 2* gets a bit higher, then 3* even higher, then 4* higher still. And if you look close, you can see there was a deliberate attempt to make rank 1 and rank 2 level up costs relatively low, and then higher ranks jump up substantially. The general principle was "start low, get higher, accelerate upward." So lower rarities cost low, higher ones cost more, and the increase itself increases. Individual ranks within a rarity also follow this accelerating upward curve. So what happened when 5* champs entered the game?
Interestingly, the rank "acceleration" mostly disappears. Each rank gets higher, in a (very roughly) even amount upward. 5* level up costs start high, and each rank higher they just keep jumping upward, kinda sorta roughly evenly. In other words, the 5* designer looked at the older stuff and went "nope" and just made it simpler. And a lot more expensive. Did that happen with 6* champs? Well...
This is just the same chart from above, but now that we look at it in context, holy cow. The 6* designer looked at the 5* designer and said "hold my beer." Actually, it looks like three separate designers worked on 6* ISO level up costs. The first one worked on rank 1 and 2 and said "screw all these weird cost curves, I'm just going to make one curve for rank 1 and then add a little bit to it for the rank 2 curve. The second one said whoa, this is getting a bit out of hand, I need to lower the costs a bit (rank deceleration for rank 3+ may have had something to do with this) and then the third guy came in and tried to continue the rank 3 and 4 curves but seems to have forgotten to carry a two somewhere.
All of them seemed to think we had too much ISO burning a hole in our pockets, because 6* rank up costs do not follow the original 1* to 4* inflation curve, and they don't even follow the much more aggressive 5* progression. They just jump immediately to Pluto. One way to see this is to look at the costs on a logarithmic scale. A log scale will make accelerating costs seem more linear. Where you see straight lines that's accelerating upward. Where you see evenly spaced data (in the vertical scale) you're seeing exponentially increasing costs.
This is the history of ISO cost acceleration. We can see that costs rise, but on the log scale each curve tends to jump upward initially and then level off. That's not costs leveling off, that's cost inflation leveling off to some small inflationary rise. And you can see each cost curve starting and ending higher and higher in a remarkably even fashion. That reflects the fact that each rank and each rarity is getting more expensive, and the costs themselves are increasing, and the *increase* is itself increasing, which is what you'd expect to see in an inflationary environment.
And then 6* rarity ( the black curves) jump up dramatically. And on a log scale, jumping upward dramatically is not just inflation, and it is not just inflationary acceleration. Its just a sudden dramatic discontinuous leap upward that no progression in costs could account for.
I always knew 6* champs were much more expensive to level up, and I've always simply adjusted to that cost increase without thinking about it too much. But actually analyzing the numbers carefully, I can see that 6* champs are in fact too expensive. Not in the sense that players don't like the cost, and not in the sense that the costs can't be absorbed in theory. Rather, this degree of cost increase is something that cannot possibly be reflected in the game economy. To put it another way, the developers might have been trying to respond to some temporary perceived requirement to drain ISO from the game economy, but there's no way that 7* champs could possibly continue this trend from 5* to 6*. 6* champs are not just more expensive, they are not even exponentially more expensive. They are more than exponentially more expensive. Now, that doesn't mean they are impossible to level up. Obviously. "More than exponentially more expensive" does not mean "astronomically more expensive." Rather, it means the rate of cost increase is expanding at a faster rate than any economic forces could possibly account for in a finite game. Which means at some point when we look back on the 5* vs 6* vs 7* costs, we will almost certainly see 6* costs as an anomaly. Which, in design terms, will mean it was set wrong in the first place.
Now, we can't really change the costs of 6* rarity, the rarity has been around too long and there's been too much level up activity. It would trigger a big discussion involving the C-word. But it suggests that the ISO economy is not really in the place it "should" be. This is a very sticky subject, because ISO bottlenecks mostly revolve around subjective things: how fast should players expect to be able to level up champions. That's an impossible thing to really argue. But maybe there's an objective way to look at ISO costs, within the context of what the 6* ISO cost curve is doing relative to prior precedent. If 5* ISO costs were reasonable for their time and 4* ISO costs were reasonable for their time (if 4* costs were too low they would have tried to compensate for that *with* the 5* costs) then can we say what the 6* cost discontinuity is doing relative to those costs?
When I figure that out, you'll all see Part Three.
For those that didn't see it, here's a link to Part One (Gold costs to use ISO)
There's actually some decent information out there in charts I'm sure everyone has Googled several
Well, here's the bottom line:
These numbers agree with the numbers I've seen floating around for 2* through 4*. The only difference is with 4* rank 5. The number I see floating around is 266570. I get 266571. Basically the same. However, my 5* rank 4 costs are somewhat higher than what I've seen floating around: I get 452073, while the number I've seen reported was 444616. Still not too far off. Both of my 6* numbers are also a few thousand off of the numbers I've seen, and I haven't seen any numbers publicly reported for rank 3 through 5. It kind of makes sense that the higher you go, the more likely someone could have made a small error, codified that into a chart, and then no one else ever bothered to check, because why would anyone do that (ahem)?
How do I know my numbers are that right ones? Well, as I said, I didn't just measure the ISO costs to level up to max. I recorded them for every single level.
Every. Single. Level.
If anyone wants to check my work, or just wants to know what the individual numbers are, here's a spreadsheet with all of them:
https://docs.google.com/spreadsheets/d/1bToRtGQiSQsLJBt8A8mmgeTPGtnf6S6Lh1irxEQ_gNk/edit?usp=sharing
If you're just here for the ISO costs, feel free to skip the rest. For those that want to stick around for the after party, here's some interesting observations. ISO costs have progressed in an organic fashion. Which is what designers say in polite company when they mean "made this &*$! up." Here's what ISO level up costs looked like when the only rarities that existed were 1* through 4*, which is more or less the original design of the game:
You can kind of see what's happening here. 1* costs are way at the bottom, then 2* gets a bit higher, then 3* even higher, then 4* higher still. And if you look close, you can see there was a deliberate attempt to make rank 1 and rank 2 level up costs relatively low, and then higher ranks jump up substantially. The general principle was "start low, get higher, accelerate upward." So lower rarities cost low, higher ones cost more, and the increase itself increases. Individual ranks within a rarity also follow this accelerating upward curve. So what happened when 5* champs entered the game?
Interestingly, the rank "acceleration" mostly disappears. Each rank gets higher, in a (very roughly) even amount upward. 5* level up costs start high, and each rank higher they just keep jumping upward, kinda sorta roughly evenly. In other words, the 5* designer looked at the older stuff and went "nope" and just made it simpler. And a lot more expensive. Did that happen with 6* champs? Well...
This is just the same chart from above, but now that we look at it in context, holy cow. The 6* designer looked at the 5* designer and said "hold my beer." Actually, it looks like three separate designers worked on 6* ISO level up costs. The first one worked on rank 1 and 2 and said "screw all these weird cost curves, I'm just going to make one curve for rank 1 and then add a little bit to it for the rank 2 curve. The second one said whoa, this is getting a bit out of hand, I need to lower the costs a bit (rank deceleration for rank 3+ may have had something to do with this) and then the third guy came in and tried to continue the rank 3 and 4 curves but seems to have forgotten to carry a two somewhere.
All of them seemed to think we had too much ISO burning a hole in our pockets, because 6* rank up costs do not follow the original 1* to 4* inflation curve, and they don't even follow the much more aggressive 5* progression. They just jump immediately to Pluto. One way to see this is to look at the costs on a logarithmic scale. A log scale will make accelerating costs seem more linear. Where you see straight lines that's accelerating upward. Where you see evenly spaced data (in the vertical scale) you're seeing exponentially increasing costs.
This is the history of ISO cost acceleration. We can see that costs rise, but on the log scale each curve tends to jump upward initially and then level off. That's not costs leveling off, that's cost inflation leveling off to some small inflationary rise. And you can see each cost curve starting and ending higher and higher in a remarkably even fashion. That reflects the fact that each rank and each rarity is getting more expensive, and the costs themselves are increasing, and the *increase* is itself increasing, which is what you'd expect to see in an inflationary environment.
And then 6* rarity ( the black curves) jump up dramatically. And on a log scale, jumping upward dramatically is not just inflation, and it is not just inflationary acceleration. Its just a sudden dramatic discontinuous leap upward that no progression in costs could account for.
I always knew 6* champs were much more expensive to level up, and I've always simply adjusted to that cost increase without thinking about it too much. But actually analyzing the numbers carefully, I can see that 6* champs are in fact too expensive. Not in the sense that players don't like the cost, and not in the sense that the costs can't be absorbed in theory. Rather, this degree of cost increase is something that cannot possibly be reflected in the game economy. To put it another way, the developers might have been trying to respond to some temporary perceived requirement to drain ISO from the game economy, but there's no way that 7* champs could possibly continue this trend from 5* to 6*. 6* champs are not just more expensive, they are not even exponentially more expensive. They are more than exponentially more expensive. Now, that doesn't mean they are impossible to level up. Obviously. "More than exponentially more expensive" does not mean "astronomically more expensive." Rather, it means the rate of cost increase is expanding at a faster rate than any economic forces could possibly account for in a finite game. Which means at some point when we look back on the 5* vs 6* vs 7* costs, we will almost certainly see 6* costs as an anomaly. Which, in design terms, will mean it was set wrong in the first place.
Now, we can't really change the costs of 6* rarity, the rarity has been around too long and there's been too much level up activity. It would trigger a big discussion involving the C-word. But it suggests that the ISO economy is not really in the place it "should" be. This is a very sticky subject, because ISO bottlenecks mostly revolve around subjective things: how fast should players expect to be able to level up champions. That's an impossible thing to really argue. But maybe there's an objective way to look at ISO costs, within the context of what the 6* ISO cost curve is doing relative to prior precedent. If 5* ISO costs were reasonable for their time and 4* ISO costs were reasonable for their time (if 4* costs were too low they would have tried to compensate for that *with* the 5* costs) then can we say what the 6* cost discontinuity is doing relative to those costs?
When I figure that out, you'll all see Part Three.
For those that didn't see it, here's a link to Part One (Gold costs to use ISO)
Post edited by Kabam Ahab on
28
Comments
thank you for your service
Actually, one of the more interesting challenges was figuring out how to level up low rarity champs without jumping straight from level one to level twenty two. I had to keep a set of small bricks around and rotate them in and out. I would tap a brick and see the level counter rise, then unselect it and select a bigger brick, then combine two smaller bricks, then unselect them and pick a bigger brick, sort of like trying to make exact change from what's in your pockets.
Getting things like the 1* rank costs was an interesting challenge that involved phoning a (Kabam) friend, as I already had every 1* champ fully ranked. But every single level up cost is something I collected myself directly. None of it is anecdotal data collected from someone else.
Iso curve is messed up Exponentially
Good job!
No, I'm not going to level up every single champion individually next. Well, hopefully not. But if anyone has a list of the ones that are supposedly or suspected to do so, I will see what I can find out.
After discussing this with my source at Kabam, it seems that the higher ISO costs for Sentinel and Ultron are an artifact of a time in the distant past when MCOC designers were experimenting with a lot of ideas (think Karnak) and one of those ideas was that champions might have different "types" that would be designed in different ways and cost different amounts to rank up. Ultron and Sentinel's ISO costs (which are higher, but not really a lot higher) is a design fossil of those experimental ideas that never got beyond those two champions (I think).
When 5* rarity was introduced this idea was long since dead and so those champions reverted to costing the same amount as everyone else, so this difference doesn't exists past 4* rarity. o it *should* be the case that all 5* and 6* champs take the same amount of ISO to level up. If anyone is aware of another exception, either at 4* or lower or especially at 5* and above, post here and I will take a closer look.
Where is the TLDR!
TLDR: ?
I still got a Level-up Cost Guide from 2015 with that notation at the bottom, when 4* was max rarity.
Don’t know if it was Trucos who made the chart, since it didn’t have Trucos logo on it, but it was from way back then.
Design fossils can persist long past the point when the idea behind them is technically dead. That’s why they are called fossils.
I doubt there was a day when someone stood up and said “we’re not doing this anymore.” Rather, someone thought it was a good idea and tried it out, but it didn’t catch on. A year or two later a designer could have started work on Sentinel, pulled up Ultron as a starting point, and just copied what they thought would be the same and overwrote what they thought should be different. Or they might have seen Ultron and said “oh that’s a cool idea” and built it into Sentinel, before they learned MCOC had moved past the idea.
I don’t know the details, but I do know such things happen in game development studios. Sentinel might be a fossil, or he might be an echo of the past. Either way, they are not doing that any more.
Interestingly, the ISO costs to level 7* champions from 1/1 to 1/25 and from 2/1 to 2/35 are almost identical to the ISO costs to level up 6* champs from 4/1 to 4/55 (1215196) and 5/1 to 5/65 (1404845). This is probably not coincidental. We could try to use the 6* ISO costs to project what the 7* ISO costs will be for future ranks, except:
Yeah, it doesn't follow a consistent curve. But actually, the last three ranks almost follow a nice quadratic curve, so we can attempt to project using just those three points, and ignore the other two (damn interns). If we do, we get this:
The costs above the line are observed, the costs below the line are projected. We'll see how close I am in 2025.