16 May, 2010

point estimates and standard deviations

If you could freeze a point in time, any point - it would be a point estimate as you would not possibly be able to find an exact point, seeing as time is a fast moving stream. It's like trying to pick out a certain molecule of water in a river. Almost impossible... unless it is a fixed point in time - as in a fixed event. Like Doctor Who says, some events cannot be changed, or you get a paradox. In that case, it would become a fixed point estimate.

Point estimates can be variable. It's annoying, because then the calculations get more complicated and you have to remember what all the little symbols stand for. The only problem is that point estimates are useful. For example a point estimate of 1945 would be World War 2, a more accurate point estimate of that would be the end of World War 2 - but that doesn't sum up the whole year, because...drumroll... tantatara: it's an estimation. You could also say it's the 'victory of the daleks' or possibly around the time where the boy asked 'are you my mummy?' Plenty more where that came from.

I'm tired of point estimates. However my title mentions standard deviations, so let's talk a little about standard deviations and why the TARDIS has so much trouble pinpointing certain points in time. I can tell you straight up that it's the fault of variance and the standard deviation. Possibly the standard error as well. Yep, all their fault. You see, the Doctor wants to travel to a certain point in time, but because time is always in flux, always moving and has so so so so so so many variables, it's not surprising that the TARDIS sometimes accidentally misses some. Sometimes the ones it misses are the important ones when it does the calculations, so that it makes an error. Technically, it's still correct yet inaccurate, it just might be a few seconds, minutes, hours, days, years off, but it still gets there (although not always in the specified location - that might be because of the standard errors and deviations in it's Earth GPS).

Now if the Doctor specified he wanted to travel to a certain time with a standard deviation of ummm... say one, and a minimal standard error of uh, one, the TARDIS would get him as close to that point in time as it could within those specifications, because you gotta allow for error. Ask any engineer. Ask any rocket scientist. You aim to get the error as small and miniscule as possible, but the error is always still there. Imagine that. You can never be correct. You can only be nearly right.

I suppose that would apply to the Destiny in Stargate Universe too.

Now, factor space into the time complex and you get a complicated mix. Space and time are 2 separate dimensions and are treated as such; however, they are also intricately linked. The people on Destiny must be glad that they don't have to deal with this birds' nest of knotty hair. Good thing Doctor Who is a Timelord that can solve those multiple stories of mathematics in his Timelord-y brain, otherwise he'd forever be turning up at the wrong places in the wrong time and the TARDIS would never get it right. Although, I have to say, he does still make mistakes. Often. Then again, he's only a Timelord. He can't help it.

The standard deviations and errors would act as overlapping targets pointing to a specific place in a specific time, painting the area within which the TARDIS is most likely to land. Interesting. It's like that game in neopets where you have those parachuting bugs that try to land on red and white targets of varying sizes. There is room for error, but your aim is to hit the bullseye.

Have I mentioned that the latest series of Doctor Who is mildly somewhat much more scary than I anticipated? The classic series never really scared me, just cos, you know tin foil and plastic bags...great props that leave plenty of 'scope for imagination'. Note the sarcasm. The last two series of Doctor Who were reasonable. Suitably crazy and nonsensical, yet with great character moments. The current Doctor Who... well, as the 10th Doc said, when he first took over, he's 'still cooking'. Has the 11th Doctor said that? Can't remember. The Weeping Angels were freaky - but twas all good, especially considering the props are so much more realistic than the classical series. Love the accents. Gotta love the accents - bar Captain Jack Harkness. I find him more annoying than anything else. Grins too much. But if he didn't do all those 'cutesy' grins, he wouldn't be him, so I'll forgive him for not being able to help himself.

Stargate Universe on the other hand is starting to pique my interest a bit more than it did in the first half of the seaon. It started off on a flat plane, but since then, it's gained more 'wholeness' or '3-D-ness' (for lack of a better word). It's a good one for character study. Telling the story from a different perspective to what most people are used to. It's good. Saves us from getting bored of the same 'same ole, same old'. Much more interesting in the second half of the season, what with more conflicts and waving arms going 'aaaargh!'.

Anyway, I gotta get back to working out my point estimates, stomping on standard deviations, tugging t-tails, probing for values of p and digging out effect sizes before sampling whiskers from the boxes of my meta-analyses. Why, oh why statistics? Why? I can't even do long division. What am I doing trying to work out what the confidence interval means in both the fixed and variable plots and whether the null hypothesis is indeed null? Tell the numbers to be kind to my uncomprehending cerebrum, please. I'm not River from Firefly, nor Sam Carter from SG-1, nor Rodney McKay and I'm definitely no Eli. I don't play computer games. Were I to have the brain of a Timelord... actually, rather not. Those Timelord-Human metacrises aren't worth it. Look what happened to poor Donna Noble.
Well... back to the lists. Can't be late or so-and-so will strip the wax from my ears (if I don't get there first).

No comments:

Post a Comment