"Hey, stop it," Merrin says mildly at her thoroughly unconscious patient. "That's rude." She absently toggles controls on three different machines at once and then goes back to securing all the loose tubing down very firmly with skin tape. 

Her overall affect is simultaneously calm and - not tense, exactly, but in a baseline state of intense hypervigilance. She is very, very focused on the patient, taking in relevant data and ignoring nearly everything else, her moment-to-moment emotional reactions tied intimately to subtle changes in her patient's condition. Things going worse-than-median-predictions make her wince and sometimes swear out loud at the equipment. 

Sometimes, though, things go well, better-than-expected even, and then Merrin beams proudly at her patient and says out loud that he's doing really well, and in this case she was checking peripheral circulation, and gently squeezes his limp hand for a moment before returning her hands to the control console. 

Merrin is fascinating to watch, and to have a little shoulder-model of in his head. 

(Khemeth, to be clear, finds basically all humans fascinating. But Merrin is neurodivergent in such an unusual way, and it's always rewarding to see someone repeatedly exceeding expectations.) 

....It's also incredibly endearing - and, honestly, kind of emotional - to watch her interacting with his little brother in a way that makes it so incredibly vastly obvious how invested she is, how strongly she already cares about a person whose specific existence she wasn't even aware of an hour ago, who she's been interacting with directly for thirty minutes and has in fact arguably not 'met' per se. It's interesting. Khemeth likes people, and also gets invested much faster than average, but for him it usually at least involves a conversation. 

[Suggestion: raise bid to keep her on the case] he sends to the group chat, and half-tunes back in to the Venture Capitalists' videoconference.

There is some sort of intense debate happening over whether they can, by throwing more money at the problem, get some of the specialty supplies over to the Exception Handling medtech team on the fast transport plane, do an aerial rendezvous and transfer to the patient transport plane in time to start the early stages of the neuroprotective protocol while still en route to Default. 

Patient now secured, vital signs re-stabilized after the most recent procedure and transition, and with Merrin having given the go-ahead, the medicopter takes off. They were on the ship for less than thirty minutes. It's been around fifty minutes, total, since the patient was pulled out of the water. 

The medical transport airplane took off uneventfully 20 minutes ago, just about on schedule, and is already 200 miles closer. If everything goes smoothly, the medicopter will reach the plane for rendezvous in around 70 minutes, transition should be complete within 20 minutes, and at that point they'll be about 1300 miles from Default - call it 130 minutes airtime.

Just over three and a half hours, and they'll be in a real medical facility with full backup. 

Merrin is cleared to begin cooling the patient as soon she feels in control of the situation! 

...The markets are not incredibly confident that this will work. The conditions are far from ideal; the patient will be less stable at a colder temperature, they don't know by how much or how well Merrin will be able to compensate by juggling her six-machines puzzle, and being in a moving helicopter in a storm is not great for that. This is roughly why, while the patient's predicted odds did improve when Merrin agreed to run the new stabilization protocol, they are still not nearly as good as the hypothetical where rewarming-with-neuroprotective-protocol starts now. Updates were roughly:

Minor damage: 82% -> 65%

Moderate damage: 52% ->32%

Severe damage: 34% -> 18%

The odds improved a little in response to the absence of anything going horribly wrong or worse-than-expected while Merrin was performing the final procedure, and are now:

Minor damage 62%

Moderate damage 28%

Severe damage 14%

They're about to narrow the uncertainty a little, though, when they learn whether or not they can realistically hit 18 C while still in the medicopter. Because if the patient tolerates cooling poorly (which is, you know, pretty reasonable of him, human bodies are really not meant to exist while that cold) the odds of a fatal complication start to be uncomfortably high even, within their rather ruthless cost-benefit framework. 

A major issue - of course, one they are planning for in advance - is the actual midair transfer from the medicopter to the medevac plane. It's inevitably going to involve some jostling and some g-forces as the two aircraft match velocities and line themselves up while still in a storm region. And then about 100 seconds during which the patient will be inside a protective transport pod while being reeled up to the underbelly hatch of the plane - and Merrin will be nearby, with her basic control console, but without visual contact on the patient and with far more limited screen real estate to keep an eye on sensor data. The plan is for her to do what she does best: work in purely reactive stimulus-and-response, and if she doesn't have a muscle-memory-instinct reaction in her repertoire, and would have to pause and think about it, instead call it out as an Exception to the remote support team and then ignore it and focus on the rest of her interlocking six-part control system, and someone else on her team will implement whatever the Treatment Planning team can come up with in ten seconds. (Which is a serious efficiency loss, since they won't be fully loaded-up with all the context that Merrin is tracking mostly on an implicit and instinctive level, but it should suffice to make sure the patient at least keeps having a blood pressure.) 

If the patient tolerates cooling well, and is stable (or as "stable" as anyone nearly 20 degrees below normal body temperature can possibly be) at the point of transfer, the markets are actually giving it 98% odds that any problems caused will be resolvable, and the patient will reach Default alive, albeit maybe with a lot of extra stress and hassle for Merrin and her team. At that point, given the sheer quantity of advanced life-support equipment and backup they have at their fingertips, the patient's chances of straight-up biological death during rewarming are negligible.

If the patient tolerates cooling very badly - if he's fragile enough at 18 C that Merrin is already working flat-out just to keep his basic bodily functions going - then Merrin cannot be expected to have much slack for the inevitable perturbations of giving that delicate system a hard shove. It's way too likely that she won't be able to stay on top of everything. They'll of course be throwing all the remote support they can at her, but there is far too high a chance - 30%, maybe - of one of Merrin's variables - electrolytes, or pH, or some weird cell-damage detritus that the liver should be clearing - spiraling wildly out of control. Which isn't a guaranteed failure, but in maybe 50% of those cases (15% overall) the patient is either basically dead on arrival, or else accumulates enough additional cell damage that the neuroprotective protocol will be coming far too late. 

They do not want to be in that world, and so if it looks like they're headed there, when Merrin starts cooling the patient, the plan is to stick with keeping the patient at 22-24 C, and if that means losing ground on neurological outcomes, letting the inflammatory cell-death cascade get further before they can interrupt it, then - well, if they kill Kalorm before even reaching Default, they aren't going to have a chance to put the neuroprotective profile into action at all, are they. 

...And, of course, it's most likely that they'll end up somewhere in between those two extremes.

If the markets were expecting one outcome more than another, they would already be updating on that. If there is anything about the patient's underlying condition that makes success or failure more likely, and they can measure that, it's already been taken into account in the current predictions over outcomes.

There are, of course, things they can't measure - but from the point of view of the medical prediction markets and the Treatment Planners, operating under uncertainty, the upcoming turning points are going to feel pretty much analogous to a certain dath ilani children's game involving using quantum randomness generators to simulate efficiently functioning markets. 

First up! Merrin is going to start cooling the patient by about 1 C per five minutes (he's currently at 22.9 C, about 5 degrees from the new goal) - and briefly pause around halfway to the goal temperature, call it around 20.5 C, to wait on the results of some blood tests that aren't included on the continuous sensors, or where the continuous-sensor readings at such a low temperature are imperfectly reliable. 

The market does not think that the patient is going to crash before this point; it's the second half that's going to be especially fraught. Call it a 10% chance, 1:9 odds, that the patient destabilizes enough that Merrin has to abort and re-stabilize him before they reach the 20.5 mark. If the patient copes okay with the cooling, it's a modest but far from definitive good sign, one that mostly impacts the odds on the worst outcomes. 

Call it a 1% drop in the probability of minor damage, a 2% drop for moderate damage, and a 3% drop for severe damage. This is pretty likely! 90% likely, in fact! 

Of course, this necessarily means that in the 10% case, the odds of the bad outcomes will increase much more steeply: by 9% (62% -> 71%) for minor damage, 18% (28 -> 46%) for moderate damage, and 27% (14% -> 41%) for severe damage. 

Is Merrin ready to find out which world they're in? 

Aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaahstressstressstress! 

(Well, the previous improvement in the patient's odds was entirely based on Merrin being willing to try this moderately risky new stabilization protocol! The patient's odds are strictly better if she starts cooling - it's just that they're going to find out sooner if the patient's underlying condition is worse-than-expected, but it's not that they're causing that; they would have found out one way or another eventually anyway.) 

See, but Merrin has secret information about Kalorm that the markets don't, which is that Kalorm is very good and so he has to be okay Well, if Merrin really did think she had information the markets didn't, she really ought to point this out. But she usually shouldn't be more right than the markets - sure, she's the one on-site, but there are more experts and traders, they're smarter, and she's already conveyed the fact that Kalorm keeps doing better than she was expecting.

Sheeeeeeee does not think that even her illegible intuitions actually disagree with "10% chance this won't go well". It's just that her motivation system feels like this should not be allowed. This is not actually how probabilities work. 

Controlled cooling (or warming) is a little harder than it would be if she had the patient on a full heart-lung bypass circuit, but 0.2 C per minute is totally doable even in a moving helicopter while juggling six different life-support machines. 

And??????????

It seems to go basically fine! 

The transition time is obviously more stressful. His blood pressure is confused about what Merrin is doing to it, his CO2 - now actually within the stupidly narrow parameters they handed her - is trying to drop below them, as his metabolism slows a few ticks further and less is produced. 

But he responds to Merrin's interventions and tweaks, and once she's gotten him leveled off at 20.5C, fifteen minutes in - 55 minutes to go before the midair transfer - it's not much harder to maintain the whole system than it was at 23 C. 

Additional bloodwork comes in within vaguely acceptable bounds. 

The markets update! 

Likelihood patient will survive through the end of the rewarming protocol: 93%

Likelihood of at least minor brain damage: 61%

(Corresponding likelihood of no long-term cognitive deficit: 39%) 

Likelihood of at least moderate brain damage: 26%

Likelihood of worst-outcome severe brain damage: 11%

Odds of an unexpected complication en route and change of plan: 35%

Also, they're not done yet. Is Merrin comfortable proceeding, and cooling the patient to 18 C? They'll go a little faster, this time, since nothing disastrous happened and the disastrous outcomes were, if they happened at all, very likely to start happening by now. Is Merrin comfortable cooling by 0.5 C per minute, so they'll be at the goal in 5 min? It took another 5 min to obtain and analyze all the interim lab results, so the rendezvous is now in 50 minutes, and they want Merrin to have enough time to get really very comfortable with the patient's stabilization needs at 18 C, since she'll have to keep doing it under extremely non-ideal conditions during what may be the single highest-stakes 100 seconds of the entire evacuation. 

Merrin is mildly relieved, but not surprised, that the first bit went well. Yeah, she can manage 0.5 C per minute cooling. Probably. Presumably if something goes super awry she'll slow down and switch to a more conservative cooling protocol, so she can actually stay on top of wildly fluctuating vital signs and other bodily functions. 

Yep, they'll tell her if she needs to slow down, though Merrin should of course also use her judgement if she needs to, instead, switch to a very conservative 0.1 C per minute cooling protocol (which will take 25 minutes, leaving Merrin only 20 minutes to get settled in a new rhythm with a patient who will, if they need to resort to that at all, in expectation be a lot more fragile.) 

They think this is relatively unlikely to happen, but more likely than before; they're putting a 25% probability on it. 1 in 4 odds.

If the patient tolerates the further cooling, that's actually a substantial good sign, one that has the largest impact on the moderate-brain-damage odds (the predicted odds on severe damage are already low, and most of the remaining probability mass on it is in worlds where that degree of damage was already done, not worlds where it hasn't yet happened but might or might not happen in the next 3 hours depending on whether they do everything exactly right.) '

The markets are not always right about how far they'll move, but currently are looking at 3:1 odds of a 3% decrease in the probability of mild damage – to 58%, which is actually getting kind of close to even odds of no permanent cognitive deficit at all. 

(It's not under debate that there is currently damage; the sensor readings are enough to confirm that. But now that they've dealt with the worst of the dehydration and Merrin is successfully bringing his sodium steadily down, the patient's brain does, in fact, look basically structurally normal to blinded experts examining his CT scans. Under optimistic models - partly based on their tiny dataset on near-drowning in young children, which wouldn't normally be comparable to a 25-year-old with less neuroplasticity but may, for a bizarrely resilient 25-year-old, be closer to comparable than looking at post-cardiac-arrest outcomes in 50-year-olds with heart disease. The initial damage dealt by those 11 minutes without oxygen, and some-smaller-number-of-minutes without circulation and with CO2 and anoxic-cell-damage byproducts accumulating, may still be within the brain's capacity to heal from and repair, and with high-quality supportive care and eventual rehab - which they can take for granted he'll have access to - the patient could eventually return to his cognitive baseline.) 

The likelihood of moderate damage will be the most affected; the market is expecting to drop by 6%, landing at 20%. (The hypothetical market branch on "immediate access to neuroprotective protocol" actually updated in an optimistic direction after some of those case studies, plus a larger though less-directly-applicable dataset of stroke patients, were reviewed in more detail, and landed on 17%; 20%, 1 in 5, is about as close as they're likely to get conditioning on having to wait 3 hours.)

Severe damage is already looking relatively unlikely, but the market anticipates another 2% drop, to 9%. 

...Of course, in 1 of 4 worlds, they're going to get bad news. If the patient is unstable at 18 C, then either they have to return to a less-ideal higher temperature, or if not they're still risking additional damage accumulating just because Merrin won't be able to as consistently stay within the really-quite-unreasonably-tight parameters on the protocol. (The markets have not yet settled on which of those is better; behind the scenes, there is a very contentious debate going on about it.) 

So, with 1:3 odds, they're instead going to be looking at a 9% increase in likelihood of minor deficits (to 70%), an 18% increase in moderate damage (to 44%), and a 6% increase in severe damage (to 17%).

(Probably! But they are, still, reasoning under massive uncertainty. The patient's brain could be in better or worse shape than they're aware of, and whether his other organs tolerate cooling to 18 C is not directly informative about that. It's possible the patient's actual chances are different than the ones the market has the information to show.) 

Merrin takes a deep breath, and for a precious ten seconds, takes her eyes off the screens, takes one of her hands off the control console, and squeezes Kalorm's hand. "Listen. You're doing really well and I'm so impressed and I need you to keep it up. I know it's hard. It won't be for that long. Just - hang on for me, okay?" 

(Does Merrin expect this to have any effect whatsoever on how the patient's body reacts? Absolutely not. His EEG is a flat isoelectric line; there's no information-processing happening there. But she knows from sims that it does, actually, have a small positive effect on her performance if, when she's about to attempt something high-stakes and very very hard, she takes a moment to– it's not obvious even to her what she's doing, actually, or why this helps, but she's a little calmer and a little more focused.) 

It's more obvious to Khemeth, actually, who is currently running his best attempt at a Merrin-copy in his head. Merrin benefits from - no, stronger than that, she needs to feel that she and her patient are fighting for the same side. 

It's easy, when you're a medtech handling a complicated protocol with a critically ill patient, wrestling dozens of metabolic processes that really are not designed to work under these conditions and keep trying to run off in incredibly inconvenient directions, and you've just signed yourself up for twice as much stress and hassle to buy your patient a few more percentage points of a good outcome, and also you expect to be doing this for up to sixteen hours - it's easy, under those conditions, for the human mind to start feeling like the battle they're fighting has to involve an enemy. It's easy to end up frustrated with the patient for making you work so hard, especially if things end up going not the way you hoped. 

Merrin needs to feel like she and her patient are a team, and the enemy they're fighting is - the vagaries of biology, the forces of entropy, and of course the horror of True Death. Merrin isn't increasing the odds that Kalorm will tolerate being cooled another 2.5 C, but she is, genuinely, improving her own chances of coping gracefully if he doesn't. She won't blame him, because he isn't her enemy, Death is. 

...Also, Khemeth is running a Kalorm in his head, still, because one of his five text chat windows is with the Exception Handling Assistant Financial Coordinator on the case. In the absence of explicit written medical advance directives, and given the fact that Kalorm is a bizarre psychological alien by dath ilan standards and she's still incredibly confused about his preferences, she feels more comfortable if she can sort of just consult the Kalorm Expert every time something comes up. 

(Under different circumstances, there might be a prediction market on this. Khemeth is in fact considering setting up a small secret market, just so he can stop doing this extremely exhausting and emotionally destabilizing mental modeling, but he's only going to do that if he decides that his family members are coping and are not going to set anything on metaphorical-or-maybe-literal fire, and if Kalorm's friends on the boat – the conscious ones, because the unconscious one is now going to emergency cryosuspension, no he's not having feelings about that – recover enough over the next couple of hours to be up for participating in it during the actual rewarming.

But! Guess what! Khemeth has a very good idea of exactly how radioactively furious his brother would be about his internal preferences on whether to live or die being subject to debate by strangers and (even worse) trading experts. And he's not sure it would even help - he's genuinely not sure that people who are neither family members nor close friends of Kalorm's even have useful inputs here. So. No prediction market.) 

Anyway. It turns out that running a mental model of Kalorm, while watching Kalorm's rather skilled, given her age, and very very very dedicated Exception Handling medtech give some quiet encouragement to Kalorm's unconscious body before she embarks on a truly heroic effort to try to save his life–

(and this in the context of having spent the last hour terrified of Kalorm gone forever or, maybe even worse, Kalorm helpless in a hospital bed, probably in pain, certainly confused and scared, not really himself, not really there, but not really gone either, and being the one who has to make the final call on whether to let that remnant of his brother go the rest of the way)

- it turns out that these are conditions under which Khemeth, who has not had an unendorsed emotional reaction in front of anyone save his boyfriend since he was twelve, has to switch off his video and audio for the Venture Capitalists conference call because he is now crying. 

This does not actually interfere with pinging Ranthir in the family group chat to check if she finds the 3:1 odds of Kalorm tolerating 18 C convincing, or if she's also looking at them with a dubious expression and remembering a particular childhood adventure on an in-hindsight-ill-advised family skiing vacation when eight-year-old Kalorm decided to run away, made it like four miles on foot, fell off a 15 meter cliff, and had to be rescued by Exception Handling after Nerdel convinced Finnar that no it was not safe for them to mount a family rescue expedition with no backup. And then was somehow uninjured and completely fine after like twelve hours in -10 C winter conditions.

Khemeth has no idea if this is medically comparable enough to be genuine new information that isn't screened off by the EMTs already knowing literally everything about Kalorm's physiology that can be measured by a sensor, but he is still sort of frowning at that 3:1. 

Ranthir does not very clearly remember that incident. She was five at the time. But, yeah, she's absolutely betting against the markets on those odds. The markets are definitely wrong. Kalorm's actual odds have got to be at least 9:1.  

....okay, some of that is in fact a rationalization from the fact that Kalorm is not allowed to die, and the odds given by thousands of actual medical experts do mean something and should shift her prior. (You get very good at immediately catching that, when you have Ranthir's career.) 

She still thinks the markets are wrong in that direction, though, and will absolutely bid on 4:1. 

Finnar is in his basement workshop - not the secret basement workshop a level below it, just the normal basement - and currently has the full-wall projector screen displaying the View Three vast tangled multicolored visualization of the 5-dimensional structure representing every metabolic pathway and upstream gene-expression pathway known to dath ilan. The other wall display has the full-detail diagram of the metabolic pathways touched by the neuroprotective protocol. The full, secret version of the protocol, that is. 

(Finnar is not really supposed to know about that. He wasn't involved in the Harkanam case, and Khemeth swore a grade-three secrecy oath about it, and neither of them would ever expect Khemeth to violate that even under these circumstances. But Finnar is the sort of person who guesses about the secret project in the Basement of the World at age thirteen, and is fortunately quietly approached by a Keeper before he does anything ill-advised like "tell everyone", and Finnar very grumpily agreed to a secrecy oath, and has also never broken it. His son Kurthin does know, but not because Finnar told him; Kurthin is just enough like his father to have also deduced it, though not until he was fifteen. Finnar tries not to be too disappointed. Not everyone can be expected to keep up with him. In fact, in his experience, almost nobody can keep up with him and the people who can are all working on secret Governance projects.

Anyway. Finnar does not work in the Basement, because that calls for a higher level of legible trustworthiness than anyone has ever assessed him as having. He's not bitter about it, no really. However, once you know about the Basement at all, have not been told per se but have some hypotheses about what the Basement can do with custom protein synthesis, and then spend three minutes reading up on the history of the non-secret protocol that suddenly emerged onto the scene a couple of years ago, it's pretty obvious. He had to do a lot of quietyelling to get access to it, and swear a grade of secrecy oath which he is kind of bitter about, and then right as he was ramping up for a very satisfying argument, it turned out that somehow Khemeth of all people was already read-in on the situation and handling it. But, of course, the important thing here is that it's being handled. Kalorm has to be okay. Kalorm will be okay, even if Finnar has to personally become an expert in neurochemistry and figure out a protocol that fixes brain damage after the fact.) 

He's not an expert in medicine or biochemistry. Not even a little bit. Most of his domain knowledge is in advanced physics and computer hardware. But he is very, very smart, +3 SD thinkoomph from the dath ilan median and more like +3.5 SD on visual-spatial specifically, and he is absolutely not going to let a lack of any domain experience stop him. He will be confidently wrong repeatedly until he manages to be confidently right! This strategy has never failed him! 

Though he'll take five seconds to message Khemeth, who is of course not working on anything nearly as important to saving Kalorm's life, but does have a lot of connections and a lot of money.

He tells Khemeth to pay for a retrieval mission to get the wreckage of Kalorm's boat brought back to Default. 

Khemeth would like to register that he is, in fact, really busy. Why in the superheated depths of the Deep Ocean Trench is retrieving Kalorm's boat a priority here.