# Smolin book impact in November

## Your expectation of the Smolin/string impact ratio on 1 Dec: closest wins

0 vote(s)
0.0%

0 vote(s)
0.0%

40.0%

0 vote(s)
0.0%

60.0%

0 vote(s)
0.0%
1. Nov 1, 2007

### marcus

The US has an extreme string monopoly situation compared with other countries (as several posters including f-h and John Baez have mentioned elsewhere) in the sense that there is only one non-string QG research group in the whole country.

This contrasts with Europe, Canada, not to mention other places, where there are a dozen or so active non-string QG groups (two or more faculty, typically, with grad students and postdocs). In the US, grad students do not have a choice, except at one university, and new PhDs usually have to go abroad to continue non-string work.

To my knowledge, so far just one book has appeared which not only criticizes this bizarre imbalance, but also describes the various non-string QG approaches, and argues that, since we don't know where the key advances will be made, research bets should be spread. What interests me about the book is its case for positive policy recommendations.

In the belief that books can sometimes bring about change, I want to track the impact of this book---The Trouble with Physics...and What Comes Next. The wider the audience it reaches, the greater its impact, the more hope I have that the string research monopoly will be broken at several other top US universities. Then the balance will be more like it is in the outside world and graduate students will be able to pursue careers in non-string QG and related applied areas.

So I'm tracking Smolin's Amazon sales ranking relative to a kind of benchmark which indicates the size of the problem---the salesranks of the five most popular stringy books.
This gives a ratio. Just to have a regular time, I take a day's reading at noon pacific.

1 October 6.5
1 November 5.2
1 December ?

What do you think this ratio will be at noon on 1st of December? I am going to register my guess---you can if you want, or you can keep it to yourself. Either way, we will see how closely we come to correctly assessing the situation.

To give an example, on 1 November the Smolin book's salesrank was #1771
and the average rank of the top five stringies was #9194.4
The ratio was 9194.4/1771 = 5.2.
That is, judging by salesranks, the Smolin book was selling some 5 times better than topfive stringy average, which serves simply as a benchmark.
As it happens, the five most popular string books on this day were
Randall warped
Greene elegant
Greene fabric
Kaku parallel
Steinhardt endless

2. Nov 1, 2007

### Gokul43201

Staff Emeritus
Marcus, at Barnes and Noble, Trouble has a salesrank of #35,562, while Elegant has a salesrank of #18,249

That's a ratio of about 0.5!

Last edited: Nov 1, 2007
3. Nov 1, 2007

### marcus

Great! I am glad to have your idea of another possible index. I have gotten into the habit of relying entirely on the PHYSICS bestseller list at Amazon.com. It is very convenient because all the bestselling physics books are together in one list, in order by salesrank.

I will have a look at what Barnes and Noble offers. Thanks for the idea.

EDIT: I found the B&N physics bestseller list
http://search.barnesandnoble.com/booksearch/results.asp?CAT=806716&srt=S&z=y&cds2Pid=17801
(I must say it does not look as if they carry a very complete line of popular physics books. Maybe their business is mostly not in science books. I will check it out further...)

MORE EDIT: I found the B&N page for the paperback Smolin
http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?z=y&EAN=9780618918683&itm=2
they are selling it for 15.95 (or 14.95 for "members", but I think that costs something or commits one to buy more from them in order to qualify for "membership"). I will compare the 15.95 with Amazon price.

Hmmm. Amazon price for same book is $10.85, and it is selling at salesrank #1178 among all books at Amazon. Please correct me if I am wrong but I think no person in right mind would buy that book at Barnes Noble where they must pay$5 more. So I doubt that the Barnes Noble sales figure has much connection with the reality of the demand, sales, and impact of the book.
That is just a first-impression guess based on the obvious clues.
========================

As of noon 2 November Smolin book's standing was #1363 and the stringy topfive average was #10,135.0
so the Smolin/string ratio was 10,135.0/1363 = 7.4
As it happened the five most popular stringy books today were
Greene elegant
Randall warped
Greene fabric
Kaku parallel
Kaku hyperspace

what i am seeing is that string is constantly getting new titles, but they don't catch on. Like there is Steinhardt's "Endless Universe" about clashing branes, which 5 years ago might have gone over big but now seems stalled.
And there is Susskind's "Cosmic Landscape, String Theory and the Illusion of Intelligent Design" which is about string philosophy providing for millions of different universes and we live in a habitable one, which looks like it was planned, but he says no. It sounds topical and as if it would take a bite of the current market. but didnt.

Last edited: Nov 2, 2007
4. Nov 3, 2007

### marcus

The way the month is starting out there doesn't seem to be any clear downwards trend
1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
1 December ?

One might have expected a down-trend, given that the book has been out for over a year now, fourteen months actually.

At noon on 3 November the Smolin book's salesrank was #1118
and the average rank of the top five stringies was #8118.8
The ratio was 8118.8/1118 = 7.3
Judging by its standing, TWP was selling some 7 times better than topfive stringy average. The five most popular string books on this day were
elegant 2434
fabric 3364
warped 5530
elegant hardbound 12,484
parallel 16,782

It was suggested that one could also watch the Barnes and Noble salesranks of Smolin and string books.
But it turns out that their price is about 50 percent higher than the Amazon price of \$10.85 (for the Smolin book) which indicates to me that purchases of the book must be predominantly going thru Amazon. The Barnes Noble sales volume is unlikely to be sufficient for significant comparison.

I'm astonished that the salesrank ratio is so high at this point. In September I was predicting it would be 2 by the end of the month (but it turned out to be 6.)

In October I guessed it would be 3 by month-end, and instead it was 5.

Now I am predicting it will be 4 by the end of the month. But who knows?

The importance of this book, its readership, and the impact it will have, are all still difficult to assess even after it has been out in bookstores for a year.

5. Nov 4, 2007

### Chronos

The rankings comparison is somewhat deceiving. A book with an Amazon sales rank of 1000 is selling on average about 100 copies per week whereas a book with a sales rank of 4000 is selling about 50 copies per week, and a book ranked 20,000 is selling about 20 copies a week. December sales are more significant than other months because, like most retailers, it is the 'month of many profits' for book stores. So, the January 2008 numbers will be more revealing. My source: http://www.fonerbooks.com/surfing.htm. This is not a criticism of the honest effort by marcus to stimulate discussion, merely putting the numbers in perspective for the curious.

6. Nov 4, 2007

### marcus

Hi Chronos, assuming Morris Rosenthal's figures are reliable, you have succeeded in providing some information which I tried strenuously to find last year without success. Amazon is, or at least was then, rather secretive about how to translate salesrank into sales.

I hope this graph, which he says is based on 2006 data, is correct! I will try to verify it, by again searching for actual sales figures.

First of all, thanks for your efforts! If you have any more links to where I can get sales data, please PM me or post them here. Either it has become a lot easier to find out, or you deserve kudos as a web information-hound.
===============

Morris Rosenthal seems to be marketing his own book written for the small publisher with advice about how to sell on Amazon, and other business tips. He describes his method for estimating the sales volume corresponding to ranks as "reverse engineering", but I can get no further clue. As he says, Amazon keeps the actual numbers secret, so I'm still in the dark as to how he thinks he inferred them!
It could be he has inside information.

he says:
"...I used to include a table giving some equivalent sales numbers for some distinct ranks, but I dropped it with the recent update. Nobody outside of Amazon knows EXACTLY how many copies of a given title are sold in a given time period, and since ranks are relative to each other, it's a constantly moving target. The idea behind my reverse-engineering the ranking system was always to give rough idea of how a title was selling, not an exact number. So, don't read an average rank of 10,000 to mean you sold exactly 22 books that week,..."

================

whether or not I manage to confirm that his plot is reliable, it is nice to have:
it is a log-log straight line over the interesting range
#1000 corresponds to 100 per week
#4000 means 50 per week
#20,000 means 20 per week

and as you point out, ranks don't really translate into sales volume in such a simple way. at best they indicate SHARE of total volume, which can fluctuate all over the place depending, for one thing, on season. The school year. Holiday season, etc.

==========================
==========================
UPDATE
1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
1 December ?

At noon on 4 November the Smolin book's salesrank was #1333
and the average rank of the top five stringies was #9091.0
The ratio was 9091.0/1333= 6.8
The five most popular string books on this day were
fabric 5028
elegant 5714
parallel 10,663
elegant hardbound 10,822
warped 13,228

Last edited: Nov 4, 2007
7. Nov 4, 2007

### marcus

Chronos raised the interesting side issue of how to interpret Amazon salesrank in terms of sales VOLUME so I did a bit of looking around.

First off I should say that what i care about is just having SOME index of the the impact of this book that I can watch. It makes a difference to me if it stays the same, or if it dwindles away.

So I came up with this particular index, which is comparatively steady (more so than individual book salesranks) and watched it daily for four months Feb-May 2007 to get an idea of what was normal. And each month the average was about 2.5.

The fact that the book wasn't a mere flash in the pan, appearing briefly and then disappearing from the scene, seems significant to me---HOWEVER you gauge the impact.

I think on the whole I prefer to track the book's impact using salesrank because that is clear and easy---it just takes a minute to look up. And also that's what I have the longterm record of, to compare with, from earlier this year.
========================

But there is still this interesting side issue of how would you turn a salesrank ratio into a sales VOLUME ratio!

I hunted around with google for a while last year and I found that a great horde of people (authors, small book publishers, market analysts...) seem to track Amazon standings, but that NONE of them appear to know how to translate them into sales volume. The only person who hazards a guess is the one Chronos found---Morris Rosenthal---and I can find no evidence to support his graph. But just for fun, let's suppose that it is right!

The gist of Rosenthal's graph is that to convert a rank ratio into a volume ratio, roughly all you need to do is TAKE THE SQUARE ROOT! The relevant section of his curve between ranks 1000 and 10,000 is just a log-log straight line with slope 1/2!

So we can do an example.

I just looked and the Smolin rank was #1303 and the Lisa Randall rank was #10,904.
Suppose I want to translate that into a ratio of booksales volumes---to get an approximate figure in accord with Rosenthal's plot I just calculate
sqrt(10,904/1303) = 2.9
So if we believe Rosenthal, all that says is that for every copy of "Warped Passages" that gets sold there are around 2.9 copies of "Trouble...and What Comes Next" that get sold.

I don't want to try to be arithmetically rigorous and detailed about it because this log-log plot seems to be a very rough suppositious thing that he drew back in summer 2006 to put in his book for home publishers and the like. But the general thrust of it is that to the extent one can guesstimate sales volume ratios, you can do it by taking the square root.

What I'm predicting for 1st December is a salesrank ratio of around 4 (between Smolin and an "average" stringy book)
so that would translate to a sales volume ratio of around 2----just taking the square root. But I think the main lesson is the rank ratio is just an abstract index of how the Smolin book is doing and what impact it may be having on the way people think.
It doesnt have any rigorously defined meaning by itself. All that matters is how it changes. (like the "consumer price index" if it goes up, you know there's inflation and if it doesn't go up, there isn't)

Maybe someone else has a different perspective on this, in which case please comment! I would be glad to hear other views.

Last edited: Nov 4, 2007
8. Nov 4, 2007

### Chronos

My point is the number of books sold by either camp is not large. Few laymen have our abiding interest in cosmology. Most such books are purchased as gifts [I have a book shelf full of them courtesy of well intended gifters]. I anticipate the December numbers will spike, but not be meaningful.

9. Nov 5, 2007

### marcus

Hard to generalize, Chronos.
the Smolin book is not a typical science popularization----certainly not the ga-ga type, on the contrary rather sobering although admittedly passionate. He's concerned about the theory community.

the book describes several main approaches, including string, discusses their strong and weak points----often by recounting personal experiences---and it discusses policy issues (how research communities can get off track, how to keep them on track, ways to spread the bets and encourage independence...too many issues for me to list easily)

seems designed to communicate to other physicists, students considering work in quantum gravity, people close to the field who are interested observers, people conversant with research funding policy.

====================
I think it would be very difficult to translate Amazon sales volume numbers (even if we could get them) into nationwide bookstore sales volume numbers. My guess would be that most of the copies sold (in this case) have been sold in campus bookstores. And even if one knew the book's nationwide sales since September 2006, it would be hard to guess the significant readership---the gross numbers and the percent coverage of target groups etc etc. Too hard to guess.

what I do find it possible to keep track of are RELATIVE PERFORMANCE indicators.
A very important one is "how is the book doing now relative to how it did a year ago?"

I can't tell you gross numbers, you see, because Amazon is only a fraction of the sales picture. But it's a cross-section microcosm, I'd wager, so it lets us guess some things about relative performance.

Compared with other books, and over time compared with its own past performance. Those relative indicators won't tell the whole story of the book's impact, but they are important dimensions of it, worth keeping track of, I think.

===================
Anyway, just to keep us current here is an
UPDATE

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
1 December ?

I took an early reading to see how things are going: as of 9AM pacific on 5 November the Smolin book's salesrank was #911
and the average rank of the top five stringies was #7341.2
The ratio was 7341.2/911 = 8.1

At noon, the usual time, the ratio hadn't changed much. Smolin salesrank was #1326 and stringy topfive average was #10,150.8. So the ratio was 10,150.8/1326 = 7.7.

this illustrates the stability of the ratio---something I've often noticed. The standings of individual books can shift around during the day, but the things I'm comparing tend to shift in concert, so that their ratio doesnt bounce around as much. Means it's less bother to keep track of.

In case anyone is curious, the five most popular stringy books at the moment are:

elegant 4615
warped 5167
fabric 11,505
parallel 13,585
fabric hardbound 15,882

Last edited: Nov 5, 2007
10. Nov 7, 2007

### marcus

As of now two predictions are registered in the poll, dpackard's and my own.

Just to keep us current on how things are going this month, here is an UPDATE

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
...
...
1 December ?

At noon pacific on 7 November the Smolin book's salesrank was #2232
and the average rank of the top five stringies was #10,022.6
The ratio was 10,022.6/2232 = 4.5

Same time on 8 November it was #2715 and #9295.2, for a ratio of 9295.2/2712 = 3.4

In case anyone is curious, the five most popular stringy books on the two successive days were:
elegant 3780
parallel 9398
fabric 9800
endless 12,766
warped 14,369

elegant 4025
warped 7121
parallel 11,272
fabric 11,856
endless 12,202

Last edited: Nov 8, 2007
11. Nov 8, 2007

### Demystifier

Marcus, I never met a person so interested in sale ranks of scientific books. Is it somehow related to your job?

12. Nov 9, 2007

### Chronos

I'm guessing the ratio will decline due to the 'Thanksgiving effect' - 2.4 is my official SWAG for Dec 1.

13. Nov 9, 2007

### marcus

We now have three predictions registered! Packard and Chronos anticipate that the ratio will be 2.

What makes this extra likely is what Chronos called the "Thanksgiving effect"---specifically that parents buy Elegant at this time, to give to teenage children. It is the season when the giftworthy hardbound edition sells, as well as the less expensive paperback.

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
9 November 5.3
10 November 5.9
...
...
1 December ?

At noon pacific on 9 November the Smolin book's salesrank was #1520
and the average rank of the top five stringies was #8057.6
The ratio was 8057.6/1520 = 5.3

At noon pacific on 10 November the Smolin book's salesrank was #1148
and the average rank of the top five stringies was #6821.8
The ratio was 6821.8/1148 = 5.9
========EDIT TO REPLY TO NEXT POST========

Hi Chronos, I am glad you are scrutinizing these indicators and thinking about possible shortcomings. Many of your comments make sense (although I can't think of any easy way to to improve on what we are using.) Amazon says its ranking is based on number of copies sold but it does not tell us the raw data, namely it does not publish how many copies it has sold of any given book.

There was something you said that I didn't understand: "One hardcover book sale offsets more than one softback sale."
I don't see how that could be.

Last edited: Nov 10, 2007
14. Nov 10, 2007

### Chronos

Another deceiving factor in sales ranks is they are polled by accountants. One hardcover book sale offsets more than one softback sale. This contributes significantly to the 'Thanksgiving effect'. Buyers of expensive hardcover books during the holiday season are mainly gifting people with little awareness or interest in content, but a keen eye for book covers with fetching reviews - e.g., 'Seeing Red' - a masterpiece of well marketed bad science. I'm not suggesting Smolin's book is pop trash, quite the contrary. He is an outstanding scientist. I'm merely attempting to put sales figures into perspective.

15. Nov 12, 2007

### marcus

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
9 November 5.3
10 November 5.9
11 November 3.7
12 November 3.5
...
...
1 December ?

At noon pacific on 11 November the Smolin book's salesrank was #1629
and the average rank of the top five stringies was #6041.0
The ratio was 6041.0/1629 = 3.7

At noon pacific on 12 November the Smolin book's salesrank was #1898
and the average rank of the top five stringies was #6649.8
The ratio was 6649.8/1898 = 3.5

Hi Chronos, as I said in my previous post I think you may be mistaken about how salesrank based. I think it is based on number of copies sold and not on amount of money paid. So I am dubious of your assertion that "One hardcover book sale offsets more than one softback sale." However I'm glad someone else is thinking about how we can keep track of a book's impact.
The indicator that I'm using at least has the merit that it doesn't jump around as much as individual salesranks do, and that it shows trends.
Right now we can see a possible downtrend, which I was actually expecting to happen much earlier (in late September already) but which did not happen back then.

16. Nov 15, 2007

### marcus

very busy for past couple of days and missed the noon check, here's today's

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
9 November 5.3
10 November 5.9
11 November 3.7
12 November 3.5
...
15 November 3.0
16 November 2.5
...
1 December ?

At noon pacific on 15 November the Smolin book's salesrank was #2405
and the average rank of the top five stringies was #7330.0
The ratio was 7330.0/2405 = 3.0

We can see a downtrend, which I was expecting to happen in late September already but did not materialize.

Based on watching this index for four months earlier this year (February thru May) when the monthly average was a fairly steady 2.5, I have the impression that 2.5 would be a "normal" place for this ratio to be. That is, with the Smolin book doing 2.5 times better, salesrank-wise, than the five most popular string books' average. Now it's looking like the index might be returning to that 2-3 range, which would make Chronos and Packard's guess right on the money.

Here's the amazon physics bestseller list
https://www.amazon.com/gp/bestsellers/books/14545/
in case anyone would like to check out and see what physics-related books US readers are buying.

Last edited: Nov 16, 2007
17. Nov 15, 2007

### staf9

I don't see how his sales are going to be too high. Partially because of the physics (as a side note, I was always skeptical of string theory, and now even moreso after reading Garrett Lisi's paper), and partially because Smolin doesn't have that charisma in his writing like Greene does.

People who weren't interested in physics at all bought Greene's Elegant Universe because he could explain it in a level where most people could understand the theory, and that's where you get the majority of your sales from.

18. Nov 15, 2007

### marcus

I think you probably hit the nail on the head, as far as explaining the great popularity of Brian Greene books. Smolin by comparison has a fairly plain clear style---not much poetry but the verbal analogies tend to be accurate. Sober (no New Age mystique, very little by way of flourish) experiential (where went, who talked to, what learned, how the struggle with ideas went etc.) and directly about the policy issues he cares about (freedom for grad students to study wider range of approaches to gravity and unification, spreading the research bets, non-string postdoc career paths).
It's not strictly speaking a science popularization book, simplifying for the masses---it is more a book that is from the heart to whoever is prepared to listen and understand.
Given that enormous difference in style and intent, I find it truly amazing how well the book has done on the market and how many people seem interested in its content.

BTW thanks for registering your forecast! I like seeing what other people expect to happen in the future.

Last edited: Nov 15, 2007
19. Nov 18, 2007

### marcus

Well that big September October spike does seem to be subsiding. Chronos guess and the others (Packard, Staf9) are looking pretty good.

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
9 November 5.3
10 November 5.9
11 November 3.7
12 November 3.5
...
15 November 3.0
16 November 2.5
17 November 3.9
18 November 2.5
19 November 2.2
...
1 December ?

At noon pacific on 18 November the Smolin book's salesrank was #2064
and the average rank of the top five stringies was #5220.4
The ratio was 5220.4/2064 = 2.5

At this point the index is behaving just about how it did during the four month period from February thru May, earlier this year.
In case anyone wants to check out salesranks and stuff for themselves, here's the amazon physics bestseller list
https://www.amazon.com/gp/bestsellers/books/14545/

Last edited: Nov 19, 2007
20. Nov 21, 2007

### marcus

That big September thru October spike has definitely subsided. Looks like Chronos, Packard, Staf9 have won this round, but I'll wait until the first of the month before I concede.

1 October 6.5
1 November 5.2
2 November 7.4
3 November 7.3
4 November 6.8
5 November 7.7
6 November 4.6
7 November 4.5
8 November 3.4
9 November 5.3
10 November 5.9
11 November 3.7
12 November 3.5
...
15 November 3.0
16 November 2.5
17 November 3.9
18 November 2.5
19 November 2.2
20 November 3.7
21 November 2.5
22 November 2.2
...
1 December ?

The index is behaving as it did during four months from February thru May, earlier this year, when it fluctuated around 2.5.

In case anyone wonders, the five most popular stringy books on 21 November were elegant, fabric, warped, parallel, endless.
Their average rank was #6134.6, and TWP's rank was #2490, so the ratio was 6134.6/2490 = 2.5. As Chronos anticipated, there are signs that Greene books are getting some lift as gift-time approaches, being considered good presents for teens.

Last edited: Nov 22, 2007