Gap Math

I'm somewhat curious about the math behind some of Griffin's comments at the Senate Hearing today:

“we are focusing initially on cargo because, I just want to be clear with everybody, we already have a mechanism for getting crews to the station with the Soyuz system, but unless we can bring some new commercial capabilities online, we really have no cargo resupply. So, actually, of the two, the most important COTS capability to me right now is cargo, and I must be honest about that.”

No cargo resupply capabilities? Now I may not have as many academic degrees as Dr Griffin (I only have two so far--an neither of them in how to Pile it on Higher and Deeper), but at least by my count we have two existing, proven capabilities, and three more in the works. Progress is just as real for cargo delivery as Soyuz. ATV just docked with the station last week. HTV is slated for first flight next year. Plus there are the two COTS competitors, and a possible third based on the SOMD cargo flights RFP. That hardly seems like "no cargo resupply" capabilities to me. If you can count a sole-source foreign crew resupply vehicle that doesn't meet the full needs of a fully-operational ISS, why can't you count existing and near-term cargo resupply vehicles?

I find it ironic that when arguing for adding billions of dollars to an already bloated Ares-I/Orion budget, flying NASA astronauts on Soyuz is an unbearable problem, but when the possibility is raised of using that money to fund a program that actually has a chance of shortening that "gap", the response is that it's a non-issue. Ho-hum really, providing a fourth, fifth, and possibly sixth provider of cargo transport to ISS is a far more pressing matter.

I also found it amusing that Mark Whittington relies on the following statement from Griffin as an attempt to handwave away such a contradiction:

Griffin doubts that “even with their [the COTS companies’] best efforts, even if more money were provided, that COTS crew transportation capability will arrive in time to be available after the shuttle retires or even by the end of the current contract with Russia in 2012."

The problem is that:

1. Exercising COTS Option D does not necessarily imply giving additional money to SpaceX or Orbital for performing Option D. In fact, Orbital has more or less stated that they aren't planning on Option D capabilities. SpaceX might be a reasonable choice (seeing as how they're designing Dragon for manned flight from the get-go), but Option D funding could be recompeted and awarded to any of a number of companies, many of whom would likely be flying on existing vehicles.
2. Ares-I/Orion is not going to be flying by 2012 either, no matter how many billions are thrown at the problem (well, maybe if you threw $10B+ at it over the next three or four years, but that's political unobtanium wishalloy). That's not an apples to apples comparison.
   

The question isn't if COTS Option D could somehow, in spite of not being funded till the last second, eliminate the gap which Griffin's poor decisions have created. The question is if putting more money into COTS is more likely to reduce that gap than putting an equal amount of money into Ares-I/Orion. The GAO seems to think that putting the extra money into Ares-I/Orion that Griffin seeks would most likely only increase the probability of a 2015 IOC for Ares-I/Orion to a more reasonable confidence level (85% or better), rather than the 65% confidence level NASA has been using to make their numbers look less bad. In other words, adding several billion dollars to Ares-I/Orion would only prevent the gap from growing even bigger than the current 4-5 years.

On the other hand, look at COTS. It's total budget through demonstration of capabilities A-C is only $0.5B. If you had $1B from a Mikulski miracle to invest in NASA, which do you think has a higher probability of shortening the gap of time when we have to rely on Soyuz for all manned flights to the ISS? Putting that $1B into a $20-30B program that already only has at best a 65% chance of flying by 2015? Or putting at least some of that into a much smaller program that is intending to fly two cargo vehicles by late 2010?

I'll be the first to admit that COTS isn't a risk free venture. NASA's intentionally avoided picking providers that use existing launch vehicles like Atlas V, in favor of ventures that have to provide both a cargo vehicle and a launcher. There's a non-zero chance that one or both COTS providers won't actually be able to deliver.

But the same can be said of Ares-I/Orion. It too is going with an all-new launch vehicle (which derives little more than the paint scheme from previous Shuttle hardware) that may or may not work right the first time. It too is using a team without a real track record--when was the last time MSFC designed and flew a manned launch vehicle? How many of its engineers were even around for the last time they completed such a program? The Ares-I/Orion program has also been running behind schedule (notice how the schedule seems to slip by at least one calender year per year of development?).

So, if you were a congressman or senator with a limited amount of money available, and you have two risky ventures to pick from to try and reduce the gap, what would you do? Would you place all your money on the one option where your money is going to be a relative drop in the bucket, and that even then has little or no chance of actually reducing the gap? Or would you invest at least part of your money in a much smaller program where it has a much higher probability of actually hastening the day when the US once again has manned spaceflight capabilities--and better yet, commercial manned spaceflight capabilities?

You do the math.

Pay As You Go

One of the cutsie phrases that the current NASA Administration likes bandying about regarding their approach to exploration is "Pay As You Go". Kind of like O'Keefe's "spiral development", this pithy phrase is the latest fad at NASA. The basic philosophy behind "Pay as You Go" being that instead of coming up with an actual cost number, and asking for go ahead on what they know is an extremely expensive and bloated space exploration program, they instead try to live with their budget, and accelerate or decelerate their plans depending on funding. The idea being that if you don't actually specify a cost, it's impossible to have a "cost overrun"...

Gotta love that logic, eh?

The real problem with this approach as I see it, is when you factor in the fact that they're trying to run this as a jobs program to try and make the program as "unkillable" as possible. The idea being that while trying to reduce costs compared to the Shuttle a little bit, they want to keep as many former Shuttle employees employed as they possibly can. Who cares if the ability to attach ceramic tiles to an orbiter isn't really that important for lunar exploration? Who cares if most engineers in their right mind would get rid of the labor intensive, expensive, and dangerous SRBs in any future launch project? Who cares if it would be far cheaper to just buy existing launch services and focus instead on the orbital, translunar, and lunar portions of the architecture? This is about "butts in seats" in Utah, Texas, Florida, Alabama, California, Louisianna, etc.

The problem though is that if NASA doesn't get the big budget increases they want over the next few years, and they decide to "throttle back" the exploration development, something interesting happens...

Basically, if you are trying to keep the Shuttle team continuously employed, you have to keep paying them fulltime paychecks. Whether they're actually flying anything or sitting on their thumbs, they're collecting a fairly hefty paycheck every month. And SRB techs or Shuttle TPS techs can't exactly be moved over to engineering and plopped down in front of a computer and told to start running CFD analyses. These guys have lots of skills in maintaining the shuttle, but for the most part aren't exactly engineers.

So what? What that means is that as you "throttle back development" more and more of the exploration budget goes to just keeping people around, and less and less of it goes to actual bending metal, flying tests, firing engines, etc. As that percentage drops, it actually drags the development program out more than you would think, and can have a major impact on the end cost of the program. Especially after the Shuttle gets retired, when all of the technicians that need to be kept around to support Ares I and Ares V need to get paid out of the Exploration Systems budget every year, regardless of how long it takes to get Ares I and Ares V flying.

Let's look at a simplified example. A company has a project that requires about $2.5M in actual hardware, consumables, test articles, etc. and a fixed monthly cost for payroll and other overhead of about $250k/month. If the company is funded at a rate of $500k per month, the project ends up costing about $5M total and taking about 10 months.

But what if you "throttle things back"? What if the company is having a hard time raising money, and is only able to keep a trickle of $350k per month flowing in? If the company can't layoff employees, or shift them to part time (which may very well be the case if the company is doing something that requires lots of specialized skills), then the project ends up dragging out for over two years. In the end, it ends up costing $8.75M instead of $5M, with only about 28% of the money actually going to hardware.

Admittedly this is oversimplified a bit. If a company were in that situation, it'd probably have its engineers spend more time doing analysis, and low-cost tests, that might decrease the amount of overall tests a bit, and bring things back a little bit under control. But you get the basic point. If you have high fixed payroll and overhead costs, "throttling things back" doesn't eliminate cost overruns, it causes them. At least if you're intellectually honest. A program that ends up costing almost twice as much as it was supposed to because it was stretched out over a longer time has overrun its budget whether you declare that number up front, or play coy games about not knowing the costs.

If NASA were pursuing an exploration plan that didn't have such a large percentage of its funding tied up in maintaining large standing armies, multiple factories, etc. it wouldn't be so bad. If they were able to layoff employees (especially those not doing anything useful) when the money got tight, it also wouldn't be so bad.

Keeping the overall program cost from mushrooming due to delays would imply the ability to layoff excess employees at will. If your funding rate cut in half, and you were able to layoff half of your employees, you could theoretically still get done for the same budget, but just taking twice as long (ignoring all the details). Unfortunately this just isn't a reality, or if it is, it undermines one of the key arguments for Ares-I.

The argument goes that you need a bunch of these Shuttle employees to stick around for Ares V, and also to prevent a dangerous brain-drain from endangering any of the remaining shuttle flights (as people see their job ending and jump early). If you could really lay some of these people off without incurring these effects, the argument is bogus. If the argument is real, then you're stuck paying for them, even if they end up sucking up almost all of the exploration budget.

By going with NASA developed, owned, and operated HLVs for their space exploration plan, NASA has put itself in a situation where any decrease in funding is going to have a disproportionate effect on schedules and overall costs. If they had instead tried to build a more open-transportation architecture that mostly relied on existing and future commercial vehicles with NASA only developing the orbital, translunar, and lunar components, things would be a lot different. The percentage of exploration systems costs that were direct payroll and overhead would be much lower, which would make the project far less affected by temporary budget cuts or lack of budget increases. The overall cost of the project wouldn't go up anywhere near as quick, and the schedule wouldn't be impacted anywhere near as much.

As it is, if NASA can't manage to get something better than the full-year continuing resolution for this next year, the whole return to the moon might get delayed by several months to as much as a full year. With Democrats in charge of the House and Senate, large budget increases (or raiding money from Science or Aeronautics) are less likely to be allowed, which starts really raising questions about the political viability of NASA's exploration plan. The very thing that is being used to make the ESAS approach "unsinkable" (all of the jobs in all the Congress districts) may end up slowing the whole thing down so badly that it ends up putting the whole program at risk of cancelation at some future point.

This also highlights once again how critical COTS is to the survival of the whole ESAS scheme. If the budgetary situation doesn't end up being as rosy as ESAS initially planned (and even with unrealistic budget increases they still were coming in over budget for the timeline they're publically touting), and if COTS doesn't get the funding and support it needs to succeed, NASA may find itself spending almost its entire exploration budget on just maintaining ISS and keeping all those Shuttle people employed while they slowly try to develop Ares V and eventually start working on the other lunar transportation elements. If that were the case, and as the return to the moon date slides further and further into the future, do you really think that nobody is going to care just because they're staying within their budget?

Results obviously aren't the most important thing that Congress expects out of NASA, but they aren't entirely unimportant. At some point somebody is going to question the wisdom of continually blowing as much money as they do every year for so little in return.