A five digit rollover today: number twenty thousand! ... My last birthday was a four digit rollover, base three.

Significance testing is at best unsatisfying
On Deborah Mayo's Error Statistics Philosophy blog there is a discussion of a Bayesian attempt to ridicule frequentism:

‘Suppose we decide that the effect exists; that is, we reject [null hypothesis] H0. Surely, we must also reject probabilities conditional on H0, but then what was the logical justification for the decision? Orthodox logic saws off its own limb.’

In the blog comments, Paul Lawrence Hayes provides the original source, a talk given by Edwin T. Jaynes. His version appears on page 52 (pdf page 10):

"In the orthodox test, the sole basis for decision is probabilities conditional on the null hypothesis H0. Suppose, then, that we reject H0. Surely, we must also reject probabilities conditional on H0; but then what is the a posteriori justification for the decision? Orthodox logic saws off its own limb."

That context makes it clear that Jaynes is joking, and does not claim that this refutes frequentism. He even acknowledges that this is an acceptable proof by contradiction (albeit probabilistic), and states that even Bayesians will find occasion to use this form of reasoning.

But his complete talk, and even the title of the talk itself -- "The Intuitive Inadequacy of Classical Statistics" -- makes his larger point clear: we should prefer constructive proof to proof by contradiction.

But in the discussion on Mayo's blog, no one seems to understand this. Perhaps it is because of blind devotion of the discussants either to frequentism, or to Bayesianism?

Softball 2013
In late July, I managed to play four games of softball.  The bone bruise at the bottom of my left tibia had improved to where I was able to walk 2+ miles and run short distances without a setback.

On Sunday 7/28 I played my 40+ team's final regular season game.  We won, finishing 16-2, in first place two games ahead of the powerhouse team that had won the league for the past few years.  Heh, perhaps thanks to my absence up to that point, it was the team's most successful season in its 16 year history.  More seriously, my two recruits -- a shortstop last year and a pitcher this year -- made the big difference.

I batted last -- in order to run as little as possible -- and played three innings of first base.  ... The shortstop and pitcher had told me that even if I couldn't play outfield, I could still help the team because the regular first baseman had such limited range -- not just fielding, but catching throws.

On Monday 7/29 our team entered a multi-division playoff that included the top two teams from our league, and eight teams from the town's third and fourth strongest leagues.

We won our first game 18-11.  I played 4 innings of first base and went 2 for 3.

We then played our 40+ rivals and lost 9-4.  Both teams were showing signs of fatigue, especially ours.  I played 2 innings of first base and went 1 for 2.  Up to this point, I hit every ball well.  The two outs were line drives to the outfield.

We then played a team that had lost earlier that night to the first team we played.  It was their 4th game of the night, but they were much younger than us.  We lost 14-1, just managing to score our lone run in the last inning.  I was no help on offense, going 0 for 2 with weak grounders.  I had just gotten into the field for one out at 1st base when our left fielder strained his groin.  I replaced him the rest of the game.  I was very busy out there and made a nice shoestring catch to end the top of the last inning.

The next morning, my ankle felt fine.  But then I walked 2+ miles in the evening and it began to hurt.  Through trial and error for the next few days I found I could only walk about half a mile without generating pain the next day.  I began improving from that point, but I've also had a couple of minor setbacks since then, so I'm still limited to less than a mile of walking at a time.

The manager of my baseball team is begging me to play, but I'm not sure if I'll return this season.

Run Run Shaw Prizes
I just read in Significance magazine (published by the Royal Statistical Society and the American Statistical Association) that Hong Kong filmmaker Run Run Shaw is still alive -- 105 years old and has established the Shaw Prize, 'the Nobel Prize of Asia'.

Where were you when you first heard about Stein's paradox?
Normal Deviate (statistics professor Larry Wasserman) asked this question on his blog yesterday.

... I was in the basement toward the end of my first year of college (still living with my parents), reading "Stein's Paradox in Statistics" an article by Bradley Efron and Carl Morris in the May 1977 issue of Scientific American.

The article solved a puzzle that I'd first encountered a decade or so earlier.

After learning multiplication and division, and that a baseball season had 162 games, I would project the season homerun totals for players listed in the newspaper's column of 'league leaders.' Early in each season there would be players projected to surpass Maris's single season record of 61. But invariably, all of the players would fall short, sometimes very far short.

In my senior year of high school, I took a course in statistics, and the puzzle deepened when we learned that the mean was an unbiased estimator.  That made it sound like you couldn't do any better!

Sources of linguistic diversity
OvercomingBias quotes a New Scientist article that claims:

"This [Papua / New Guinea] linguistic diversity is not the result of migration and physical isolation of different populations"

I'm skeptical. Have you ever tried to cross from one valley to another on foot in the tropics? Where my parents grew up in India, it's an incredibly arduous daytrip -- even when there's a trail -- and you invariably find another dialect or language spoken when you arrive.

Spider god
This report claims that the spider is building its own decoy.

But that doesn't make sense: why would you build something that would attract (even more) predators that destroy your web, even if you manage to survive when they do?

Alternative hypothesis: these spiders are weaving images of their deity.

I hope they don't mess it up!

... so it will make for a nice birthday present.

David Deutsch on Artificial Intelligence
Quantum computation physicist David Deutsch recently published an essay on why we haven't been able to develop a general-purpose artifical intelligence. I think it's well worth reading and in particular, making an effort to understand him, for which it will be helpful to read his book _The Beginning of Infinity_, which explains some of the key assertions he makes in the essay.

Some comments:

1. I totally agree with him that the core functionality missing so far in attempts to produce artificial general intelligence ('AGI') is "_creativity_, the ability to produce new explanations." (Note that Deutsch's term 'creativity' is a bit different from common usage.)

2. Deutsch concludes that "it is plausible that just a single idea stands between us and the breakthrough. But it will have to be one of the best ideas ever."

Based on my experience with roadblocks I've encountered in a wide range of disciplines, my guess is Deutsch is wrong: rather, all AGI attempts thus far lack not one, but at least two missing ingredients. I think it's possible that some of what we lack may not even be novel. But the reason for the roadblock is that adding any proper subset of what's missing makes the implementation weaker or intractable.

In this regard, I definitely sympathize with the camp that believes that all we need is more powerful hardware. It seems quite possible that this is one of the missing ingredients. But I agree with Deutsch that we're missing something else, too. Faster computation alone is not going to produce creativity.

3. Despite my agreement with Deutsch, I see a big hole in the logic we share: evolution was able to produce human intelligence -- including in particular the ability to create effective explanations -- without having any ideas at all. Just computation (very broadly defined) and time (billions of years).

How fast can we simulate this process? Do we know how to speed up the simulation, for example, by guiding evolution?

Day game
I've gotten pretty far behind reporting my games this year. Today was our team's 8th game and my 6th game (I missed two games due to quad strains and then a back strain). It was our first day game of the year, and it was our first win of the season. I got on base 4 of 5 times (two walks, two singles). Last time up I hit a high pop foul that the third baseman hardly had to move to catch.

I made a couple good defensive plays but I actually did less interesting stuff than usual.

I've now gotten on base 14 out of 20 times this season, but I still feel very rusty at the plate.

You are viewing backtobaseball