All of roundsquare's Comments + Replies

A question about Bayesian reasoning:

I think one of the things that confused me the most about this is that Bayesian reasoning talks about probabilities. When I start with Pr(My Mom Is On The Phone) = 1/6, its very different from saying Pr(I roll a one on a fair die) = 1/6.

In the first case, my mom is either on the phone or not, but I'm just saying that I'm pretty sure she isn't. In the second, something may or may not happen, but its unlikely to happen.

Am I making any sense... or are they really the same thing and I'm over complicating?

You might be interested in this recent discussion, if you haven't seen it already: []
I think the difference is that one event is a statement about the present which is either presently true or not, and the other is a prediction. So you could illustrate the difference by using the following pairs: P(Mom on phone now) vs. P(Mom on phone tomorrow at 12:00am). In the dice case P(die just rolled but not yet examined is 1) vs. P(die I will roll will come out 1). I do agree with Oscar though, the maths should be the same.
The cases are different in the way that you describe, but the maths of the probability is the same in each case. If you have an unseen die under a cup, and a die that you are about to roll, then one is already determined and the other isn't, but you'd bet at the same odds for each one to come up a six.
In the second case, you either roll one on the die or not, but you are pretty sure that it will be another number.
Remember, probabilities are not inherent facts of the universe, they are statements about how much you know. You don't have perfect knowledge of the universe, so when I ask, "Is your mum on the phone?" you don't have the guaranteed correct answer ready to go. You don't know with complete certainty. But you do have some knowledge of the universe, gained through your earlier observations of seeing your mother on the phone occasionally. So rather than just saying "I have absolutely no idea in the slightest", you are able to say something more useful: "It's possible, but unlikely." Probabilities are simply a way to quantify and make precise our imperfect knowledge, so we can form more accurate expectations of the future, and they allow us to manage and update our beliefs in a more refined way through Bayes' Law.
It looks to me like your confusion with these examples just stems from the fact that one event is in the present and the other in the future. Are you still confused if you make it P(Mom will be on the phone at 4 PM tomorrow)= 1/6. Or conversely, you make it P(I rolled a one on the fair die that is now beneath this cup) =1/6

Thats fine, as long as you lay out the relative importance of different aspects so people can predict what will and won't be important to you.

As long as I'm doing what I decide to do, why would I worry about varied reasons for doing it?

One reason that comes to mind is that you might be avoiding something you should be doing.

Ah, I see what you are saying. Thanks for the explanation. And you are indeed correct.

I'm not sure I see how I"m privileging the hypothesis. Not saying that I'm not, but if you can explain how I'd appreciate it.

Aside from that, I think you are using "god" to mean any of the gods discussed by any popular religion. By this definition, I'd probably agree with you.

I was using the word "god" in a much more general sense... not sure I can define it though, probably something similar to: any "being" that is omnipotent and omniscient, or maybe: any "being" that created reality as we know it. In either de... (read more)

There is no reason to propose such a being - privileging the hypothesis is when you consider a hypothesis before any evidence has forced you to raise that hypothesis to the level of consideration. Unless you have a mountain of evidence (and I'm guessing it'll have to be cosmological to support a god that hasn't visibly intervened in the world) already driving you to argue that there might be a god, don't bother proposing the possibility.

For anyone interested, here is a decent algorithm for getting the "correct" number of lines in your linear regression.

Pages 5 and 6.

Ouch. Comic Sans. Good cookbook, though.
Welcome to LessWrong! Feel free to introduce yourself in the welcome thread. [] That is a very good summary and review for those who want want to brush up on dynamic programming -- it gives several example problems and cost functions to be minimized, and shows how the optimal substructure fits in. I do have to say that the bit for the tradeoff between overfitting and accuracy is not terribly useful for those trying to understand such things. It is a cookbook method, with no justification for why these particular error weightings are terribly useful. EDIT: Of course, almost any regularization will help compared to nothing, and it does show a nice way to do this with dynamic programming, which can greatly speed things up over naive implementations.

You need to make two assumptions for the analogy.

1) You can't re-light the candle.

2) If you do things exactly right, you'll get out with just before starving to death (or dying somehow) otherwise, you are dead.

I think it makes sense, as a poke at atheists.

Think about it this way. You walk into a bar, and you see no bartender. In your mind, you say "anything that is a bar will have a bartender. No bar tender, not a bar." Of course, the best thing to do before revising your assumptions is to wait for a bar tender. Maybe he/she is in the bathroom.

Similarly, if you claim there is no evidence of god that I've seen in my lifetime, you are using the wrong measure. Why should god (if there is one) make itself obvious during the short period that is a human lifetime.

This is almost an "irrationality quote" instead of a rationality quote, but still enlightening.

I was with you up until the "similarly". After that you start privileging the hypothesis [] - you should expect a god to make itself obvious during a human lifetime, by any description of a god ever proposed in history.