Why Learn Basic Programming Skills?

[raybo]

Originally Posted by traynor

Good advice, with one disagreement. I think a novice will get further faster with the CS106a from Stanford than the equivalent from MIT (which is great if you are already familiar with some of the basic information). Jumping straight into the fine points of sort algorithms might be a bit arcane to a newbie.

David Malans is far and away the best instructor, but the CS50 from Harvard places too much emphasis on the trendy web-related material (JavaScript, etc.) rather than the basic foundation novices need.

That basic foundation is covered elegantly, engagingly, and thoroughly by Mehran Sahami in CS106a, and expanded in CS106b. One thing it won't do--that so many dry, boring presentations like that from MIT manage to do--is completely turn off a novice from pursuing further studies in computer science.

The "talking heads" model of instruction is seriously deficient. It is also pervasive. One should be able to distinguish good instruction from bad instruction, benefit from the former and soundly reject the latter. I did a lot of research on the consortium attempting to fill the gaping deficiencies in computer science education (Monash, BlueJ, and others). The world needs more David Malans' and Mehran Sahami's. Fortunately, both Stanford and Harvard seem to realize fully that teaching is more than just talking and reading prepared notes.

For once, we agree! I know I'm extremely late to this thread, but the quoted post from you explains exactly why most people can't, and will never, learn to program, or to do statistics, or physics, or engineering, or anything else of that complexity. It's because the teachers of that stuff have progressed so far in their education and usage that they don't want to be bothered with how to add 1 and 1. They can't possibly teach it to a "novice", meaning someone who has absolutely no knowledge of the subject matter. And, this is why I dropped my physics 101 course, after one week, the same day of the first hour test, in which all except 1 of the 300+ students in that course failed the test (I made a 46 on that test, and I was in the top 10 of my graduating class in high school), at the Univ. of Texas. That instructor was exactly like the one in DL's MIT video.

[raybo]

Originally Posted by badcompany

I found this book in the basement of my building. Is there anything in there that is still relevant, today?

VBA is an easy way to automate Excel, Access, etc.. The ability to write VBA code makes Excel a very powerful data analysis tool, because you can spend much more time with other aspects, like wager structuring and wager amounts.

Much of this thread is just a pissing match between 2 or 3 people over what is going to be around in 20 years. Heck, I probably own't be around in 20 years, so why should I, or most others here care?

Bottom line is, use what works for you, whatever that is. Becoming a computer programmer will not make you a profitable player, but it might make your job so much easier that you can actually spend more time on things you didn't have time for before, and that might be the difference between losing and winning.

[raybo]

Originally Posted by badcompany

I actually have a lot of experience working with spreadsheets for basic functions, primarily adding up rows and columns of numbers, but I never did any programming with them.

What exactly would a spreadsheet program do?

Anything you want them to do, you just need to know what you want them to do first.

[raybo]

Originally Posted by MJC922

Still somewhat relevant, VBA probably hasn't changed much since then, VBA is essentially VB6 for Office apps. Excel has changed a bit, the ribbon UI, there's no longer a 65000 row limit but a million rows and things like that. A book like this which generally started with high marks typically sees a refresh with each new version of Excel and in so doing they will cover any new features and put new screenshots in it to capture the UI changes etc. I probably wouldn't want to get serious about Excel 2007 or newer and then want to spend a lot of time with a book that dated but I still have a lot on my shelf which is nearly as old.

VBA doesn't care what version of Excel you use. I have 2002 (and had other versions before that), and 2010 (2007 before that, and the same exact VBA macros work in all of them.

[MJC922]

Originally Posted by raybo

VBA doesn't care what version of Excel you use. I have 2002 (and had other versions before that), and 2010 (2007 before that, and the same exact VBA macros work in all of them.

Still if someone is going to get serious about learning Excel VBA it would be wise to obtain a newer edition of the book. The 2000 edition of the book may show examples that code around the 65000 row limit which no longer exists. Additionally some functions are now deprecated, even for a commonly used function like RANK Microsoft has indicated that RANK.AVG and RANK.EQ should be used moving forward. MS has been good about allowing for backward compatibility however it would be wise to use a newer edition of the book if for no other reason than to understand the current limitations of the software.

[raybo]

Originally Posted by MJC922

Still if someone is going to get serious about learning Excel VBA it would be wise to obtain a newer edition of the book. The 2000 edition of the book may show examples that code around the 65000 row limit which no longer exists. Additionally some functions are now deprecated, even for a commonly used function like RANK Microsoft has indicated that RANK.AVG and RANK.EQ should be used moving forward. MS has been good about allowing for backward compatibility however it would be wise to use a newer edition of the book if for no other reason than to understand the current limitations of the software.

Of course, but that wasn't the question. He asked if there was anything in that book that is still relevant today. I would say that virtually 100% of the VBA used in 2000 can still be used today in the later versions. If he can get a foothold into VBA programming with that book, then he will know if it is worth it to continue, in which case I would then recommend his buying a recent book.

[raybo]

The great thing about VB, VBA, Excel, Access, etc., is the enormity of free support on the internet. If you need help or don't understand something, all you have to do is Google it and you'll get your answer, or go to one of the hundreds of forums dedicated to anything Microsoft, and post a question. One doesn't have to be a VBA programmer to use VBA, others will write the code for you (and many will add the comments telling you exactly what the code does, so you learn even if you don't try to learn, simply by reading the code). It's almost like stealing, legally!

I use "MrExcel.com" all the time, and there are literally hundreds of sites like that online.

[traynor]

http://ipython.org/notebook.html

For technical reasons that would be pointless to explain, a new user will find it easier/quicker/simpler to download Anaconda with the 2.7 version of Python rather than the "latest, greatest, spiffiest, most up-to-the-moment" version.

At:
http://continuum.io/downloads

choose the Python 2.7 version rather than the 3.4. That completely bypasses all the "multiple app dependency issues" that make Python and related apps seem so frightening to Windows users.

[DJofSD]

http://www.wired.com/2014/12/obama-b...puter-program/

[DJofSD]

http://oleb.net/blog/2014/12/program...anguages-mooc/

I just completed the University of Washington’s Programming Languages course on Coursera, and it was one of the best learning experiences I have ever had. This class is not your typical introduction to a particular language. Rather, instructor Dan Grossman aims to teach the fundamental concepts that underlie all programming languages, highlighting important (conceptual, not syntactical) similarities and differences between them.

https://www.coursera.org/course/proglang

[traynor]

https://www.youtube.com/watch?v=iopD...6EB79A149EC727

Short, lucid, no experience necessary. Intended for pragmatic users--not career programmers. In short, how to get things done that you need to get done, without the typical "hooray for our side" nonsense typical of most "open-source" offerings. A refreshing emphasis on teaching the core basics necessary, rather than fostering an "us vs them" dichotomy that has no place in learning.

[raybo]

Originally Posted by traynor

https://www.youtube.com/watch?v=iopD...6EB79A149EC727

Short, lucid, no experience necessary. Intended for pragmatic users--not career programmers. In short, how to get things done that you need to get done, without the typical "hooray for our side" nonsense typical of most "open-source" offerings. A refreshing emphasis on teaching the core basics necessary, rather than fostering an "us vs them" dichotomy that has no place in learning.

Thanks Traynor, very good example of efficient, and understandable teaching. One doesn't need to be proficient in the "tool" to understand, basically, what the tool encompasses, and get an idea of what might be possible to accomplish by using the tool. Once that is understood, one knows whether or not they want to proceed to learn to use the tool, or not.

What I now understand is that, if I decide to pursue this, from what has now been explained to me, I know that chances are pretty good that I have the ability, with some study and work, to learn to program in Python. That's more than I knew in the past.

[traynor]

https://www.youtube.com/watch?v=b7z1RybudG4

From the ground up. No previous knowledge assumed or required. Take special note of the use of the IPython Notebook. As I mentioned above, it is a game changer. So is Greg Wilson and the Software Carpentry project.

Best suggestion for those interested in applying basic programming skills to handicapping horse races (at this point): Watch the five (short) videos at the link above, download Anaconda 3 for Python 2.7 and use the Spyder IDE (included with Anaconda, and it already has the IPython Notebook).

Anaconda has every analysis and data mining tool you are likely to need for quite awhile. At least until you learn enough to make those greedy whales work a whole lot harder for their money, to make up for that which you are siphoning off.

There is also a decent (if a bit slow) series of tutorials on "Learning Python with Spyder" on YouTube.

[traynor]

My mistake. Wilson is using IPython, not IPython Notebook. IPython Notebook has way more functionality, while being as easy to use as the IPython in the videos. Another error--there are more than five videos, but they are all short and easily assimilated. No profound, ponderous pontification by hard-core academics intent on demonstrating their "importance" and "elevated status" to hapless students (rather than teaching).

[traynor]

i think it is important to understand the "why" as well as the "how." Virtually every "handicapping application" publicly available (meaning someone offers it to you, for free or for a fee, or "free" because they get subscription kickbacks from data providers, or whatever) is based on crunching the same set of numbers (Equibase for US/Canadian races), and reflects the bias(es) of the developer(s).

That means that you have to accept the "wisdom" of the developer as superior to your own. If you have an idea you want to try, or test, you can ONLY do it within the constraints imposed by the developer, or by the data providers in conjunction with the developers. That may often dilute your idea(s) so much that they produce skewed results, no results, or poor results. There may be workarounds in some applications, but they generally take more skill to pry out than it would take to code a simple app to test your idea(s) on your own.

The idea that you need a computer science degre and a million race database is utter nonsense. Initial testing takes only a handful of races at your local track, and may even be worth entering data by hand for those races. No data subscriptions needed. No big investment of time or resources needed.

An example. When I was in graduate school, a group of students formed a "handicapping group." One of the most useful ideas was "what happens if we ignore time completely?" It is not a new idea--Sartin and others have advocated it at various times under various labels. In essence, actual race times are ignored--all "times" are entered as defaults, and only the running positions and beaten lengths are considered as meaningful.

It took very little effort to code a simplistic app in the TI-84 calculators that all students were required to have. It took very little effort to pick "a good recent race for comparison" for each horse, and 3-4 seconds to enter the data for that horse into the TI-84. Contender selection was simple--take the top 5 morning line, or--if less adventurous--the top 4. No rocket science, no agonizing over woulda coulda shoulda. And it picked a LOT of winners.

A few more lines of code, and it generated "pace ratings" based primarily on each horse's performance, rather than arbitrarily as a factor of race times. As much as the concept might make "pace handicappers" or "speed handicappers" cringe, it worked just about as well as the number crunchers using data downloads. Additionally, it was not restricted to Equibase data--it could be used on "foreign" races just as easily, and often with better results. Not every race at every track every day--but more than enough to be quite rewarding.

So what's the point? An idea (one of many, BTW) was implemented with a few lines of code, tested on a handful of races, and the results were promising enough to devote more time and effort to improving it. You never know until you try whether one of those ideas you have been kicking around in the back of your mind for years might be the keys to the candy strore. It makes no difference whatsoever how it performs "in the long run" on a million race database of old races from tracks you have no interest in whatsoever. It only matters how it performs on the races you actually bet on.