The legs are all on shoulder joints with z-axis rotational motion. The front and back legs have enoguh range of motion to mirror what we know of the Mammoth MK II and the ATAT. The center legs contribute to the movement via rotating on that axis.
Ahh. Hadn't factored in the shoulder joints. Man, no wonder they're having problems with bugs.
If you're not using the minimum number of dice to enable one of those projects to complete, you're wasting dice. If you were, for example to put 5 dice on the Station Bay this turn for a 49% completion chance, that would be ignoring the 9% possibility of spending only 4 dice and completing the project for one die less. Doing projects slowly trades expediency for dice efficiency, and since we have a limited number of dice each turn, over multiple turns that enables us to do more things overall. That's Planquesting 101.
Viewing any one project in isolation, your analysis is correct. In a field where we have a large number of
uncorrelated projects running in parallel, such as Military, your analysis is also correct. If I were concerned with only a single project, or a set of unrelated projects in parallel, I would follow your analysis to the letter.
This is a special case.
I am not viewing one project in isolation; I have two projects that I'm balancing against each other. Both projects are being completed at the moment because they are valuable to a third project, which should not be begun until both of the others are done. Each of the two projects is relatively useless until the other is completed, so there is little or no advantage to hurrying one project if it means the other is significantly delayed.
If there weren't a path-dependency here, if I didn't need to complete (A and B) in order to begin (C or D), then I would prioritize either A or B, trickle dice into B, spin out one or the other while allocating most dice to C or to D, and so on. I would, in general ruthlessly optimize
only for dice efficiency. You saw me do this with the moon mines in the last several turns, so I hope you know that I know how.
...
Let us consider a scenario.
If I do as you describe and put four dice on the station bay and two on the shuttle factory... Let us not worry about very unlikely contingencies. Let us instead worry about
maximally likely contingencies. Suppose we get perfectly average outcomes on all dice. Obviously, perfectly average outcomes are quite unlikely, but statistics mean that they are more likely than any other
single outcome (there are more ways to get numbers near 202 on 4d100 than numbers far from 202), and that the great majority of probably outcomes cluster around the mean outcome.
If we look at this possibility, in 2062Q1, the station bay is at, if I remember my arithmetic rightly, 4*50.5 + 4*(26+5), and the shuttle factory is at 2*50.5+2*(26). I believe, to be clear, that the station bay gets Seo's station bonus but the shuttle factory does not. If I'm right, and if I've avoided typos, the relevant two lines on the probability array would look like...
-[] Station Bay 326/400 1 die 20 R 73%, 2 dice 40 R 100%
-[] Leopard II Factory 153/350 2 dice 40 R 26%, 3 dice 60 R 82%, 4 dice 80 R 98.7%
Now, before we go on,
you are correct to say that it is not some kind of disaster if we don't start
Columbia in Q2. At the same time, it has been made very clear along multiple lines of support that in a real sense,
Columbia is the flagship of our overarching space plan. It is what we have been building up to since the mid-2050sstarting such an important and high profile and long-sought project, one that has been called for for several years, is important and the public has an interest in seeing it happen in a timely manner. The clamor grows loud: "Build us
Columbia!" And we may have many other things to do with our orbital dice, some of them gated behind
Columbia itself, if we are trying to meet a "space population" target, since we cannot build other habitat facilities of any real consequence or size until
Columbia is ready.
Thus, while delays are
acceptable, they are not
desirable. I view them as a bad thing that we would prefer to avoid, all else being equal. It is worth some degree of sacrifice to prevent delay, as opposed to a project where delay is insignificant, as was the case with the lunar rare metals mines.
Bearing this in mind, what do we do with our projects? Well, one die on the station bay is the obvious move. It would be nice if the chance of completion were higher than 73%, but delay to Q3 is
acceptable, and it's probably not worth blowing an Orbital die to improve the chances of a 2062Q2 start by just 27% or less. but what about the shuttle factory? Following your approach, we would invest two dice, but the preponderant likely outcome in that case is
non-completion; we are making it more likely than not that we
can't start
Columbia in Q2. Hopefully we'd be able to finish the shuttle factory in Q2 and start
Columbia in Q3, but if we fall prey to Freeze Drying Plant Syndrome, we could have a real problem on our hands with
Columbia's delays stretching out much of the time from plan start to the 2063Q4 elections. Starbound would not be happy, and with each passing turn the temptation would grow to just bite the bullet and start
Columbia without the benefits of the shuttle factory.
At which point it becomes
tempting to spend a third die on the shuttle factory, to get a .73*.82 = 60% chance of being able to start
Columbia in Q2 rather than a .73*.26 = 19% chance. Is it dice-optimal? Not quite. Does it have compensations elsewhere? Very possibly.
So my dice allocation in this contingent scenario would look like
-[] Station Bay 326/400 (1 die, 20 R) (73% chance)
-[] Leopard II Factory 153/350 (3 dice, 60 R) (82% chance)
-[] Orbital Cleanup (Stage 11+12) 32/170 (2 dice, 20 R) (Stage 11, 76% chance Stage 12)
Which is not so bad. I think it's worth it, rather than allocating two dice to the
Leopard II factory and, say, one die to the fusion yard, which would still take quite a while to finish anyway.
...
Now, we can debate the "two or three dice on the
Leopard IIs" question in theory. But the point I want to make is that a large part of the reason we're
in this position, in this scenario, is that the
Leopard II factory is a long way from completion... because we spent two dice on it.
My view is that spending two dice on one of the projects and four on the other is an invitation to end up locked into Freeze Drying Plant syndrome. And that was annoying and unpleasant enough then, when it was just a matter of mechanical effects. It's going to be considerably more frustrating when the project being delayed is a key flagship program for us.
So I'd rather balance spending on the two projects.
If you want an "acceptable and reasonable" chance, instead of a minimal but viable one, you're assuming that low chances of completion won't happen and therefore end up spending more dice overall than you need to, thus losing out on doing more things in total.
The answer is that it depends. Simple optimization strategies become less optimal in the context of more complicated situations.
Is it worth rolling an additional die to have a 10% greater chance of completing a project? Unless you are up against a very hard deadline, no. Is it worth rolling an additional die to have a 60% greater chance of completing a project? Quite possibly, given the high likelihood that you would have to roll that die anyway and just get the benefits of the project a turn later without saving yourself any trouble in the long run.
And again, I want to ask you: Why? Why do you want to start Columbia in Q2 instead of Q3? Why are you willing to spend more dice to get the stations done faster? Why are you treating this like a sprint, when it's an endurance run?
I am not treating this as a sprint. Why are
you trying to get a sliver of a chance of rush-completing one of the projects this turn? You seem to have the idea that I'm planning heavy dice over-expenditure. I am not. In some unusual statistical edge cases (say, where a project is 130 points from completion and thus unlikely to finish with one die but quite likely to finish with two) I
might consider courting some degree of 'over'-expenditure for the reasons you describe. But that is not my preferred outcome.
