The other best thing is to help refine and improve the models and information presented here, but it feels like the default failure mode to focus too much on that and not enough on continuing the work.
That’s definitely a failure mode that exists, but I’m not worried about it here, relative to incorrect models. A short list of things I think we need to nail down before concluding this was a success:
I’m kind of shocked to see “they mayor said so” cited as conclusive evidence that a government official acted based on nothing but public outcry, in less than one workday. At a minimum we should expect shipping executives to have raised this issue with the city ahead of time, because of strong incentives to do so.
It’s been 10 days it’s unclear if the backlog has improved? It’s very hard to get data on this- Marinetraffic.com either aggregates by week or doesn’t break out cargo ships, which are a small percentage of Long Beach’s work work, but
This article from 10⁄20 lists “over a hundred” waiting ships
The tweet on 10⁄22 lists 70 waiting ships (but that might be Long Beach only)
This article lists 153 waiting ships as of 10⁄29 (7 days after the tweet)
This article lists ~70 waiting ships as of 11⁄2. That might be big progress, or an inconsistency in definitions. And of course, throughput could increase while keeping backlog the same as ships redirect from other ports. And even then, this wasn’t the only change made in the last month.
I’m all in favor of doing more things and iterating rather than waiting for perfection, but it really seems like we have not squeezed nearly the learning we could out of this one.
Marine Exchange Facebook page has graphs of “container ships at anchor or loitering” for Los Angeles + Long Beach combined. It has been relatively flat in the 70s since mid-October, with an early November dip followed by a bounceback. The peak of 80 was on Oct 24, the current (Nov 8) is 77.
Port of Los Angeles has its own data (current pdf) with POLA Vessels at Anchor; it’s the “historical container vessel activity” which you can get from this page by clicking on the picture that says “Working Container Vessels”. It shows a peak of 40ish from Oct 19-29, which then dropped to 30ish, but is back up to 40 now (Nov 8).
Port of Long Beach has this page listing the container vessels at anchor there. It currently (Nov 9) lists 43 vessels. I can’t find historical data except through internet archive; the most recent archived page shows 31 vessels as of Oct 8.
If I gave the impression that I don’t support additional learning and investigation then I apologize. That was most certainly not my intention. It was more that my default model of what happens is ‘some analysis then basically nothing’ and that seems like the most likely failure mode in worlds where making this work is feasible.
To address the question of the backlog, if I wasn’t clear, I too was unable to find good evidence of the number of ships waiting and how that changed from day to day. I did see claims it had improved somewhat but it’s all confused.
I tried to make it clear here that I am confident it was, in that previous comment’s terms, ‘a first step’ and did at least some good, but that I’m not confident on magnitude. Which, as again I tried to say, I still think is good enough.
I am confused why you think this is good enough, and would like to understand why. I look at the situation and see several extremely important cruxes that are not yet satisfied. I’ve specified some of those cruxes above, albeit not super rigorously and I can’t promise they’re exhaustive.
I think I get the failure mode you’re worried about here- that comments like mine will raise the friction of doing anything so high that nothing is even attempted, and it’s more important to try more things and learn from them than to nail down any particular one. There are versions of that I would very strongly agree with. But I don’t think trying things with the level of rigor and intentionality shown in this post and the last one will generate good data. At best it’s measuring the ability of charisma and emotion to compel action, which is a symmetric weapon, but I don’t think it even does that if you don’t check the causality chain (as stated above, I don’t consider the mayor telling a newspaper he acted in response to constituent demand to be strong evidence, when it’s not even clear it was his decision. I’m surprised you do and it might be worth exploring why).
I’m inclined to offer a bounty for evidence about what caused the change in rules (including nailing down whose decision it actually was). Before I do that, I want to check if this is cruxy for you/if there is information on this specific link that would change your overall beliefs. It might also be worth making a bet on this, although having one of the bet participants administer the information bounty feels suspect.
I’m not worried about you administering the bounty, that seems fine and good and I’d trust you to honestly evaluate the bet at the size range I’d expect it to be. If it was big enough to be ‘real money’ we’d need to use better procedures but I’m assuming that’s not the case here.
I’d also note that if I was provided a ‘bounty fund’ I would at least experiment with using it more generally. Might be a good idea.
If you propose a bet, then I might or might not accept depending on odds, size and terms, I don’t feel it is necessary but have no objection.
There are two distinct questions here where you’re questioning things, as I understand you.
You are questioning whether X → Y, where X = Tweetstorm/Ryan and Y = change in LB stacking rule. We have the timing, and we have the explicit word of the person who made the change. Both seem like strong evidence to me. We also have that the storm was unique in the ways I’ve described, and in ways that reflect a desire to cut the enemy that seems unmistakable and rare to me. Yes, it is possible that it wasn’t causal, but this seems heroic and unlikely. It either has to be a coincidence followed by a decision to lie, or the Tweetstorm was timed to give cover to a decision already made. I don’t know what evidence you’d gather here or what bet you’d want to make, but this seems very much like an isolated demand for rigor to me if it’s more than being curious to explore more to verify/expand the model, which seems fine.
You are questioning Y → Z, where Y = stacking rule and Z = conditions at the port. Here I’m not sure how far apart we are. I’m saying that as long as (1 above) Ryan was causal in making the change, (2) the stacking rule improved conditions in the port somewhat with no downsides that matter, that’s enough for me to think of this as a template. It’s a bottleneck situation, so improving one bottleneck might or might not have a huge impact on the condition of the port, and it was rapidly getting worse (from all reports I know about) before the change. The scale of cost of effort versus impact of port conditions is many zeroes. And to me what matters is the template working to execute a good idea—if it didn’t ‘save the economy’ then that’s fine, never really thought this alone did it. Especially when it was only at LB and not LA! Which loses us more than half the effect. Also, my mechanism for why this matters largely involves ‘momentum builds towards both more physical actions to improve the ports, and more efforts to discover, communicate, amplify and use information, models and solutions for physical problems and places where we can improve’ and for that purpose the threshold of actual effect here is definitely cleared by a moderate improvement.
Others have raised cruxes regarding whether trying to create or use information sharing and attention direction vectors is a good idea for strategic reasons, I don’t know if you share this concern or not but it’s a distinct discussion and would require additional posts to fully address. Don’t think this could be settled by more data gathering, it’s a different kind of thing.
Something else?
Do I want to know magnitude of effect? Of course, and I am happy that you’re considering sending out a bounty to find out. I’d also love better access to better metrics on what’s going on at the ports in general even if we can’t create change.
Anyway, I offer to escalate to video call when timing allows, if that’s something you’re interested in.
That’s definitely a failure mode that exists, but I’m not worried about it here, relative to incorrect models. A short list of things I think we need to nail down before concluding this was a success:
People raised serious questions on the previous thread that you haven’t responded to
I’m kind of shocked to see “they mayor said so” cited as conclusive evidence that a government official acted based on nothing but public outcry, in less than one workday. At a minimum we should expect shipping executives to have raised this issue with the city ahead of time, because of strong incentives to do so.
It’s been 10 days it’s unclear if the backlog has improved? It’s very hard to get data on this- Marinetraffic.com either aggregates by week or doesn’t break out cargo ships, which are a small percentage of Long Beach’s work work, but
This article from 10⁄20 lists “over a hundred” waiting ships
The tweet on 10⁄22 lists 70 waiting ships (but that might be Long Beach only)
This article lists 153 waiting ships as of 10⁄29 (7 days after the tweet)
This article lists ~70 waiting ships as of 11⁄2. That might be big progress, or an inconsistency in definitions. And of course, throughput could increase while keeping backlog the same as ships redirect from other ports. And even then, this wasn’t the only change made in the last month.
I’m all in favor of doing more things and iterating rather than waiting for perfection, but it really seems like we have not squeezed nearly the learning we could out of this one.
Marine Exchange Facebook page has graphs of “container ships at anchor or loitering” for Los Angeles + Long Beach combined. It has been relatively flat in the 70s since mid-October, with an early November dip followed by a bounceback. The peak of 80 was on Oct 24, the current (Nov 8) is 77.
Port of Los Angeles has its own data (current pdf) with POLA Vessels at Anchor; it’s the “historical container vessel activity” which you can get from this page by clicking on the picture that says “Working Container Vessels”. It shows a peak of 40ish from Oct 19-29, which then dropped to 30ish, but is back up to 40 now (Nov 8).
Port of Long Beach has this page listing the container vessels at anchor there. It currently (Nov 9) lists 43 vessels. I can’t find historical data except through internet archive; the most recent archived page shows 31 vessels as of Oct 8.
If I gave the impression that I don’t support additional learning and investigation then I apologize. That was most certainly not my intention. It was more that my default model of what happens is ‘some analysis then basically nothing’ and that seems like the most likely failure mode in worlds where making this work is feasible.
To address the question of the backlog, if I wasn’t clear, I too was unable to find good evidence of the number of ships waiting and how that changed from day to day. I did see claims it had improved somewhat but it’s all confused.
I tried to make it clear here that I am confident it was, in that previous comment’s terms, ‘a first step’ and did at least some good, but that I’m not confident on magnitude. Which, as again I tried to say, I still think is good enough.
I am confused why you think this is good enough, and would like to understand why. I look at the situation and see several extremely important cruxes that are not yet satisfied. I’ve specified some of those cruxes above, albeit not super rigorously and I can’t promise they’re exhaustive.
I think I get the failure mode you’re worried about here- that comments like mine will raise the friction of doing anything so high that nothing is even attempted, and it’s more important to try more things and learn from them than to nail down any particular one. There are versions of that I would very strongly agree with. But I don’t think trying things with the level of rigor and intentionality shown in this post and the last one will generate good data. At best it’s measuring the ability of charisma and emotion to compel action, which is a symmetric weapon, but I don’t think it even does that if you don’t check the causality chain (as stated above, I don’t consider the mayor telling a newspaper he acted in response to constituent demand to be strong evidence, when it’s not even clear it was his decision. I’m surprised you do and it might be worth exploring why).
I’m inclined to offer a bounty for evidence about what caused the change in rules (including nailing down whose decision it actually was). Before I do that, I want to check if this is cruxy for you/if there is information on this specific link that would change your overall beliefs. It might also be worth making a bet on this, although having one of the bet participants administer the information bounty feels suspect.
I’m not worried about you administering the bounty, that seems fine and good and I’d trust you to honestly evaluate the bet at the size range I’d expect it to be. If it was big enough to be ‘real money’ we’d need to use better procedures but I’m assuming that’s not the case here.
I’d also note that if I was provided a ‘bounty fund’ I would at least experiment with using it more generally. Might be a good idea.
If you propose a bet, then I might or might not accept depending on odds, size and terms, I don’t feel it is necessary but have no objection.
There are two distinct questions here where you’re questioning things, as I understand you.
You are questioning whether X → Y, where X = Tweetstorm/Ryan and Y = change in LB stacking rule. We have the timing, and we have the explicit word of the person who made the change. Both seem like strong evidence to me. We also have that the storm was unique in the ways I’ve described, and in ways that reflect a desire to cut the enemy that seems unmistakable and rare to me. Yes, it is possible that it wasn’t causal, but this seems heroic and unlikely. It either has to be a coincidence followed by a decision to lie, or the Tweetstorm was timed to give cover to a decision already made. I don’t know what evidence you’d gather here or what bet you’d want to make, but this seems very much like an isolated demand for rigor to me if it’s more than being curious to explore more to verify/expand the model, which seems fine.
You are questioning Y → Z, where Y = stacking rule and Z = conditions at the port. Here I’m not sure how far apart we are. I’m saying that as long as (1 above) Ryan was causal in making the change, (2) the stacking rule improved conditions in the port somewhat with no downsides that matter, that’s enough for me to think of this as a template. It’s a bottleneck situation, so improving one bottleneck might or might not have a huge impact on the condition of the port, and it was rapidly getting worse (from all reports I know about) before the change. The scale of cost of effort versus impact of port conditions is many zeroes. And to me what matters is the template working to execute a good idea—if it didn’t ‘save the economy’ then that’s fine, never really thought this alone did it. Especially when it was only at LB and not LA! Which loses us more than half the effect. Also, my mechanism for why this matters largely involves ‘momentum builds towards both more physical actions to improve the ports, and more efforts to discover, communicate, amplify and use information, models and solutions for physical problems and places where we can improve’ and for that purpose the threshold of actual effect here is definitely cleared by a moderate improvement.
Others have raised cruxes regarding whether trying to create or use information sharing and attention direction vectors is a good idea for strategic reasons, I don’t know if you share this concern or not but it’s a distinct discussion and would require additional posts to fully address. Don’t think this could be settled by more data gathering, it’s a different kind of thing.
Something else?
Do I want to know magnitude of effect? Of course, and I am happy that you’re considering sending out a bounty to find out. I’d also love better access to better metrics on what’s going on at the ports in general even if we can’t create change.
Anyway, I offer to escalate to video call when timing allows, if that’s something you’re interested in.
Note: have emailed Zvi + Eli Tyre to arrange double cruxing out of band.