The Two General's Problem

tlb wrote:The war with Haven had changed from fleet against fleet to attacks against points of value (which might necessitate such plans); but Haven's mobile fleet had been captured and Manticore's fleet was at Haven's home planet at the time of Oyster Bay. The ships that Honor had could destroy everything that Haven had, so there was no obvious need to worry about bubble sidewalls or blocking ships.

ThinksMarkedly wrote:We know that, but the MAlign did not. When Oyster Bay was launched, the Battle of Manticore had not happened yet and information about it could not have made it into their planning anyway. They had to contend with the possibility that Haven was about. In fact, the whole plan was to give Pritchart and Theisman the ability to finish the war by winning, which required them to have a mobile fleet.

Which they did. Haven's own Python Lump of 600 SD(P)s was scheduled to come out of Bolthole in this time frame. What they didn't have was experienced crews aboard them yet. They'd have just swapped the older, non-pod designs for the SD(P)s, so the new fleet was inexperienced. Their existing personnel experienced on SD(P)s was held as POWs on Manticore.

Back to the point at hand: while the war had effectively frozen after the Battle of Manticore, the two powers were still at a state of war. The Alliance knew Haven had a secret production location called Bolthole that they couldn't find and therefore had to assume was still a threat (they could estimate about how big it was). And the Queen and Empress especially was still very distrustful of the Havenites, still calling them "peeps." So while I think the RMN did not expect Haven would mount another attack, especially not while Honor was on Haven itself, they could not stand down from standard war readiness on HMSS Vulcan.

And the MAN planners could not count on it either. Therefore, OB was planned to hit HMSS Vulcan at a time when its bubblewalls would be down (if it is not all the time) and with enough surprise that they couldn't bring the standard defences up.

That is incorrect, they might not have known that Honor was at Haven discussing an end to the war, but the information delay between Manticore and Mesa is only 5 days because of the streak drive and the wormholes. Oyster Bay was only authorized after the Battle of Manticore. From chapter 33 of Storm from the Shadows:

"I know that after so many centuries, recasting and reorganizing our plans on such short notice is enough to make anyone nervous. But let's face it, we've always known that when the time actually came, we were going to have to change planning and operational tempos. In many ways, I would prefer to continue with our original timetable. Unfortunately, the opposition hasn't chosen to cooperate with us in that respect. In my judgment, and especially in light of the outcome of the Battle of Manticore, the threat Manticore represents to our entire strategy has just increased exponentially. We cannot allow them to consolidate a clear-cut military victory over Haven, especially if they simultaneously manage to deploy four or five hundred podnoughts equipped with whatever it was they used to smash the attack on their home system. The odds of their having the strength to move directly against us under those circumstances if they realize what's really happening would be unacceptable, despite anything the Sollies might do.

-- skip --

"In other cases, some of you may find that accelerating Oyster Bay will prove fatal to your objectives. I recognize that, and I'm afraid it's simply going to have to be the price we pay. In other words, there will be no repercussions for anyone who does find his or her areas of responsibility severely damaged as a consequence.

"And whether that happens or not, my decision has been made. Although we still have to complete the detailed plans for a scaled-back Oyster Bay, our studies indicate that it will be completely feasible for us to do so. And on that basis, I have instructed Benjamin to plan for an execution date of a modified Oyster Bay strike no later than six T-months from today."

tlb wrote:That is incorrect, they might not have known that Honor was at Haven discussing an end to the war, but the information delay between Manticore and Mesa is only 5 days because of the streak drive and the wormholes.

Thank you for the correction.

Though there's also a control loop between Mesa and Darius to be taken into account. We don't know how long that takes, because we don't know what the fastest path is: it could be Mesa-Visigoth, thence via hyper to Warner, then Warner-Mannerheim, then to the Felix Junction..

In any event, your quote also brings up the fact that they couldn't wait any longer, if they wanted to catch the Python Lump in the yards (which they ultimately failed at). They did not have a choice of when to launch Oyster Bay, which means it's less than a 50-50 chance that Sphinx could have been protected by the RZ.

Only RFC can tell us for sure, but my feeling is that the HMSS Vulcan bubblewall can go into full "shields up" mode within 2 minutes. And if that is so, I see no reason why the Weyland or Hephaestus defences would be any different.

And while we know that there was no Hermes buoy system between the A and B components, the MAN could not count on that. In fact, maybe there is a low-bandwidth, military-only emergency FTL channel. This in turn means the attacks on both components had to have been carried out within 15 minutes of each other

On the Two Generals' problem: this is not the case. Either prong of the attack can succeed on its own. If one of them is significantly early, it might compromise the success of the other, but that's not what the Two General Problem is about. There's also nothing to gain in aborting, since there's not a lot of personnel in the line of fire anyway, only hardware that would be expended anyway.

ThinksMarkedly wrote:

tlb wrote:The war with Haven had changed from fleet against fleet to attacks against points of value (which might necessitate such plans); but Haven's mobile fleet had been captured and Manticore's fleet was at Haven's home planet at the time of Oyster Bay. The ships that Honor had could destroy everything that Haven had, so there was no obvious need to worry about bubble sidewalls or blocking ships.

We know that, but the MAlign did not. When Oyster Bay was launched, the Battle of Manticore had not happened yet and information about it could not have made it into their planning anyway. They had to contend with the possibility that Haven was about. In fact, the whole plan was to give Pritchart and Theisman the ability to finish the war by winning, which required them to have a mobile fleet.

(SNIP!

ThinksMarkedly wrote:

Oyster Bay might had been planned before news of the Battle of Manticore reached Mesa, but Albrecht authorized the strike because of the hundreds of Superdreadnaughts that were destroyed there, see chapter 33 of _Storm from the Shadows_.

ThinksMarkedly wrote:

tlb wrote:That is incorrect, they might not have known that Honor was at Haven discussing an end to the war, but the information delay between Manticore and Mesa is only 5 days because of the streak drive and the wormholes.

Thank you for the correction.

Though there's also a control loop between Mesa and Darius to be taken into account. We don't know how long that takes, because we don't know what the fastest path is: it could be Mesa-Visigoth, thence via hyper to Warner, then Warner-Mannerheim, then to the Felix Junction..

In any event, your quote also brings up the fact that they couldn't wait any longer, if they wanted to catch the Python Lump in the yards (which they ultimately failed at). They did not have a choice of when to launch Oyster Bay, which means it's less than a 50-50 chance that Sphinx could have been protected by the RZ.

Only RFC can tell us for sure, but my feeling is that the HMSS Vulcan bubblewall can go into full "shields up" mode within 2 minutes. And if that is so, I see no reason why the Weyland or Hephaestus defences would be any different.

And while we know that there was no Hermes buoy system between the A and B components, the MAN could not count on that. In fact, maybe there is a low-bandwidth, military-only emergency FTL channel. This in turn means the attacks on both components had to have been carried out within 15 minutes of each other

On the Two Generals' problem: this is not the case. Either prong of the attack can succeed on its own. If one of them is significantly early, it might compromise the success of the other, but that's not what the Two General Problem is about. There's also nothing to gain in aborting, since there's not a lot of personnel in the line of fire anyway, only hardware that would be expended anyway.

Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

penny wrote:Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

Well the Sharks launched the graser torps, and dropped the Cataphract pods, at their "deployment point, one light-week from Manticore, with a velocity of twenty percent of light-speed relative to Manticore-A." [SoV].

And the Manticore-A side of the attack spent just over a month maneuvering from where it emerged to that deployment point. It's exceedingly unlikely that they'd run into anybody out there they had to dodge -- but I'm sure the attack plan left extra time between their entry point and the planned deployment point. And they didn't have to hit that particular deployment point at that particular speed to put the attack on target at the pre-planned time. If they had been forced to be late they could intercept that attack path a bit closer to the system and insert the torps and pods where they 'should' have already coasted to.

But, at it was, the torps and pods had to coast ballistically into the system for over a month (~35 days) before they reached the point where the torps 'woke up' and started accelerating into attack position and the pods got ready to launch their cataphracts to hit right afterwards.

I guess, in theory, if someone had gotten in the way the graser torps could have maneuvered clear and then attempted to reestablish the same attack time -- but they'd have to do so on just their internal AI. But the lines of pods following the special particle shield lead pod have no ability to maneuver. If they even saw someone passing in front of them they'd just have to hope they weren't seen and didn't collide.
But essentially for the moment the weapons were dropped and (and, I believe, the Sharks departed) the time of the attack couldn't be altered -- it was all in the hands of physics and the ballistic trajectory the pods were following.

ThinksMarkedly wrote:On the Two Generals' problem: this is not the case. Either prong of the attack can succeed on its own. If one of them is significantly early, it might compromise the success of the other, but that's not what the Two General Problem is about. There's also nothing to gain in aborting, since there's not a lot of personnel in the line of fire anyway, only hardware that would be expended anyway.

penny wrote:Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

ThinksMarkedly is saying that Oyster Bay does NOT represent a case of the Two Generals Problem, because the attack on Grayson, Manticore-A and Manticore-B did not require communication between the three elements beyond the initial agreement on a planned execution time. He does point out that any mistiming or detection of one might result in a suboptimal result for another (but only in the binary system of Manticore, the information delay between Manticore and Grayson is too great). However that is a separate problem from the Two Generals.

Within each attack there was some communication, because the missile pods did receive targeting information from a unmanned communication platform that was uploaded by the Ghost scouts.

PS: Neither Oyster Bay nor the attack on the Peeps holding Trevor's Star can be a case of the Two Generals Problem, because in both the time of attack has been centrally determined before anyone starts out. There is no negotiation required (nor permitted) between the individual elements.

tlb wrote:

ThinksMarkedly wrote:On the Two Generals' problem: this is not the case. Either prong of the attack can succeed on its own. If one of them is significantly early, it might compromise the success of the other, but that's not what the Two General Problem is about. There's also nothing to gain in aborting, since there's not a lot of personnel in the line of fire anyway, only hardware that would be expended anyway.

penny wrote:Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

ThinksMarkedly is saying that Oyster Bay does NOT represent a case of the Two Generals Problem, because the attack on Grayson, Manticore-A and Manticore-B did not require communication between the three elements beyond the initial agreement on a planned execution time. He does point out that any mistiming or detection of one might result in a suboptimal result for another (but only in the binary system of Manticore, the information delay between Manticore and Grayson is too great). However that is a separate problem from the Two Generals.

Within each attack there was some communication, because the missile pods did receive targeting information from a unmanned communication platform that was uploaded by the Ghost scouts.

PS: Neither Oyster Bay nor the attack on the Peeps holding Trevor's Star can be a case of the Two Generals Problem, because in both the time of attack has been centrally determined before anyone starts out. There is no negotiation required (nor permitted) between the individual elements.

That is not what the book says, or rather implies.

Roderick Sung, Task Group 2.2’s Commander of the six Ghost Class scout ships made this comment. There was a close call because The Protector's Own was conducting war games way off of the ecliptic where hardly any ships operate. The Scouts were rendezvousing there.

MoH wrote: In fact, both Sung’s orders and every pre-op briefing had stressed that concern, yet he doubted his superiors would look kindly on the Man who blew Oyster Bay, whatever the circumstances.

So apparently, the Grayson prong of Oyster Bay could have blown Oyster Bay had they been discovered.

The six Ghost ships were scouting Yeltsin then meeting back at a rendezvous point way off of the plane of the ecliptic where the rendezvous was taking place. Communication could have been imperative to abort for another time. I assume the Manticore part of the attack was far more important.

IOW, an agreement or confirmation to abort could be just as important. Since the Yeltsin component was still in the scouting phase, the entire operation could have been blown.

penny wrote:That is not what the book says, or rather implies.

Roderick Sung, Task Group 2.2’s Commander of the six Ghost Class scout ships made this comment. There was a close call because The Protector's Own was conducting war games way off of the ecliptic where hardly any ships operate. The Scouts were rendezvousing there.

MoH wrote: In fact, both Sung’s orders and every pre-op briefing had stressed that concern, yet he doubted his superiors would look kindly on the Man who blew Oyster Bay, whatever the circumstances.

So apparently, the Grayson prong of Oyster Bay could have blown Oyster Bay had they been discovered.

The six Ghost ships were scouting Yeltsin then meeting back at a rendezvous point way off of the plane of the ecliptic where the rendezvous was taking place. Communication could have been imperative to abort for another time. I assume the Manticore part of the attack was far more important.

IOW, an agreement or confirmation to abort could be just as important. Since the Yeltsin component was still in the scouting phase, the entire operation could have been blown.

The Ghosts could potentially blow the entire attack because they had been dropped off way before the attack wave to scout the system.

I believe their sequence was expected to be:
1. Arrive early (dropped off by freighter)
2. Split up to scout the system.
3. Rendezvous to share scouting information and determine final attack details.
3. Program each Ghost's communication platform.
4. Sneak back in and deploy the communication platforms.
5. Slip away (But I don't know if they stayed in the system to perform distant bomb damage assessments, or if they'd hypered out at some point before the cataphracts launched)

We don't know when their rendezvous was relative to when the pods and torps got dropped - though I'd guess it was at some point after they'd started their ~35 day coast in period. And my impression is that it was just the Ghosts rendezvousing with each other -- not that they were rendezvousing with the Sharks (who'd never come within a light-week of the system)

However, if the rendezvous was more than about a week before the attack there'd be time for Grayson to warn Manticore that strange new highly stealthy ships had been sneaking around; and thus Manticore might have heightened its defenses in ways that blunted the attacks.

If it was less than a week before the attack hit then they'd only be able to spoil the Grayson arm of Oyster Bay. But in addition to Grayson raising its defenses any Ghosts it found and destroyed wouldn't be able to participate in the sharing of targeting data, nor would they be able to deploy their communication platforms. So not only would the Grayson stations likely be better protected, the targeting info on some targets would have been lost AND there wouldn't be enough communication platforms. So I'd expect some of the missiles and torpedoes to not get any final targeting update and others getting older or less complete information -- both of which also would reduce the effectivness of the attack.

But none of the plans seems to have required the Ghosts to ever communicate with the Sharks. It seems the Sharks had to simply dump their weapons on a trajectory that would deliver to the pre-planned final attack run launch point and the pre-appointed time and trust that the Ghosts would have done their jobs and the communication platforms would be there and and to provide the updated target info.

penny wrote:That is not what the book says, or rather implies.

Roderick Sung, Task Group 2.2’s Commander of the six Ghost Class scout ships made this comment. There was a close call because The Protector's Own was conducting war games way off of the ecliptic where hardly any ships operate. The Scouts were rendezvousing there.

MoH wrote: In fact, both Sung’s orders and every pre-op briefing had stressed that concern, yet he doubted his superiors would look kindly on the Man who blew Oyster Bay, whatever the circumstances.

So apparently, the Grayson prong of Oyster Bay could have blown Oyster Bay had they been discovered.

The six Ghost ships were scouting Yeltsin then meeting back at a rendezvous point way off of the plane of the ecliptic where the rendezvous was taking place. Communication could have been imperative to abort for another time. I assume the Manticore part of the attack was far more important.

IOW, an agreement or confirmation to abort could be just as important. Since the Yeltsin component was still in the scouting phase, the entire operation could have been blown.

It depends on what is meant by "blown"; I expect that the attack would have still taken place, but might be less effective. If the scouts have already loaded targeting information, then the attack on Grayson would probably been just as effective; since the moon base and shipyards would not be protected by sidewalls.

If warning could have been delivered to Manticore, then it depends on their chosen response as to how the attack might be blunted. Just finding a stealthy scout, does not indicate what sort of attack was about to take place. Particularly since the scout would not have remained in condition to be examined. Again if the scouts' information had already been uploaded, then targeting by the missiles would still have been effective. The only question is whether the sidewalls at the various stations would be raised and I would not expect that. There are parts of Hephaestus where they only could be raised if an active threat was observed, since it was the main shipping node for Manticore.

The attack could not be aborted or rescheduled, the Sharks were already inbound and could not be contacted. If a Ghost had been detected, then parts of the attack might be blunted; but it had taken six months to plan this attack and a second attack would be too late to stop Apollo. Note that the Ghosts are deployed around Manticore in chapter 48 of Storm from the Shadows and the Sharks transit into normal space in chapter 51, three weeks short of weapons deployment.

Jonathan_S wrote:

penny wrote:Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

Well the Sharks launched the graser torps, and dropped the Cataphract pods, at their "deployment point, one light-week from Manticore, with a velocity of twenty percent of light-speed relative to Manticore-A." [SoV].

And the Manticore-A side of the attack spent just over a month maneuvering from where it emerged to that deployment point. It's exceedingly unlikely that they'd run into anybody out there they had to dodge -- but I'm sure the attack plan left extra time between their entry point and the planned deployment point. And they didn't have to hit that particular deployment point at that particular speed to put the attack on target at the pre-planned time. If they had been forced to be late they could intercept that attack path a bit closer to the system and insert the torps and pods where they 'should' have already coasted to.

But, at it was, the torps and pods had to coast ballistically into the system for over a month (~35 days) before they reached the point where the torps 'woke up' and started accelerating into attack position and the pods got ready to launch their cataphracts to hit right afterwards.

I guess, in theory, if someone had gotten in the way the graser torps could have maneuvered clear and then attempted to reestablish the same attack time -- but they'd have to do so on just their internal AI. But the lines of pods following the special particle shield lead pod have no ability to maneuver. If they even saw someone passing in front of them they'd just have to hope they weren't seen and didn't collide.
But essentially for the moment the weapons were dropped and (and, I believe, the Sharks departed) the time of the attack couldn't be altered -- it was all in the hands of physics and the ballistic trajectory the pods were following.

Jonathan_S wrote:

penny wrote:Unless one of the stations was more important than the others. How exactly did they communicate? Radio silence was observed throughout the entire operation as I read. And each component would need to be slotted into place before launching. Unexpected guests could have delayed any component from their planned coordinates. In which case simply agreeing upon a time could have been disastrous; certainly suboptimal.

Well the Sharks launched the graser torps, and dropped the Cataphract pods, at their "deployment point, one light-week from Manticore, with a velocity of twenty percent of light-speed relative to Manticore-A." [SoV].

And the Manticore-A side of the attack spent just over a month maneuvering from where it emerged to that deployment point. It's exceedingly unlikely that they'd run into anybody out there they had to dodge -- but I'm sure the attack plan left extra time between their entry point and the planned deployment point. And they didn't have to hit that particular deployment point at that particular speed to put the attack on target at the pre-planned time. If they had been forced to be late they could intercept that attack path a bit closer to the system and insert the torps and pods where they 'should' have already coasted to.

But, at it was, the torps and pods had to coast ballistically into the system for over a month (~35 days) before they reached the point where the torps 'woke up' and started accelerating into attack position and the pods got ready to launch their cataphracts to hit right afterwards.

I guess, in theory, if someone had gotten in the way the graser torps could have maneuvered clear and then attempted to reestablish the same attack time -- but they'd have to do so on just their internal AI. But the lines of pods following the special particle shield lead pod have no ability to maneuver. If they even saw someone passing in front of them they'd just have to hope they weren't seen and didn't collide.
But essentially for the moment the weapons were dropped and (and, I believe, the Sharks departed) the time of the attack couldn't be altered -- it was all in the hands of physics and the ballistic trajectory the pods were following.

The time of the attack probably couldn't be altered, but the attack itself probably could have been aborted by a self destruct button. Why show your strategic hand if it telegraphs your intentions and can possibly blow the most important part of the operation. ALL of the wedges and sidewalls would have been active at Manticore in the event that the Yeltsin component telegraphed an infrastructure attack. We know the close call had to be an event outside that week window, since Sung was worried about blowing Oyster Bay. And what I think was meant by blowing Oyster Bay is in failing to destroy the new ships plus failing to destroy Manticore's building capacity. Both of which would fail if the strategic objectives of the attack had been revealed at Grayson.

I wonder if a self destruct button to destroy the pods and the g-torps was part of the plan. And whether a streak boat was waiting in hyper to beat the Case Zulu to Manticore.