Morbid use of Google Maps

Australians are obsessed with road deaths. Newspapers display "Road Deaths this year" on the front page as if they were lottery numbers. Therefore this use of Google Maps doesn't surprise me. It shows every road death in Melbourne so far this year, detailing how the person died.

Nobody cares that you "rewrote from scratch"

I read this product announcement:

"As part of our ongoing commitment to our customers to provide the best software we can, we've rewritten [...name omitted...] entirely from scratch!"

My opinion: users don't care that you rewrote your software from scratch. That's not a feature. It's not a benefit. Nobody thinks, "this software I'm thinking of buying...they rewrote it from scratch so now I'm definitely buying it."

Better is to announce the new features you've added:

"Because so many people requested it, we've added interactive charts."

or

"The new release of XYZ now supports in-situ wiki editing."

These are things that might make one curious and consider updating.

The technical details for those who really care about such things, I post in my blog. But I don't think the fact that Poker Copilot was written in Java has swayed a single person for or against buying it.

Be Yourself

A tip I read on taking advantage of being a one-person software company:

Differentiate yourself from bigger companies by being personal and friendly
A lot of small companies make the mistake of trying to act big. It's as if they perceive their size as a weakness that needs to be covered up. Too bad. Being small can actually be a huge advantage, especially when it comes to communication.

SEO is ...

A strong and pithy opinion from Eoghan McCabe on search engine optimisation (SEO):

http://www.contrast.ie/blog/seo-is-bullshit/

Maybe he's says it a bit too strong. I think SEO has a place as a marginal exercise in tweaking and tuning a website. But that's all.

Do Big Shiny Download Buttons Work?

A couple of days ago I uploaded a new version of the Poker Copilot home page. The new version has a enormous, bright, in-your-face download button that you can't ignore if you try.

I've heard that this completely unsubtle approach maximises downloads. The idea is that when people look for a product - let's say a poker hand history analysis tool for Mac OS X - you should make it as easy and tempting as possible to get the product on their computers. Did it work for me?

Let's take a look at a chart:

The spike at the end of the graph says it all. Spike is not even an adequate word here. It's a kangaroo leap to a new level.

I've not yet determined if there were side-effects, such as a drop in downloads from the download page or an increase in sales.

Why Poker Copilot stats differ from Poker Tracker stats

Some people compared the statistics Poker Copilot generates with the equivalent Poker Tracker statistics and have spotted differences. Why the difference?

Fair question. It's matter of definition. Unfortunately there is no official organisation to issue precise definitions of terms such as "Blind Steal Attempts" and "Aggression Frequency".

Consider "Blind Steal Attempts". The current Poker Copilot definition is this: a raise on the button in the first round of betting when every player except the blinds have folded. But maybe the small blind can only make a blind steal attempt if the button folds. Or the player one before the button, who make be extremely courageous or have observed that the player on the button doesn't like stealing. These situations are hard to detect pragmatically. However I feel the current definition is acceptable for making comparisons between players.

I added tool tips to each dashboard statistic defining Poker Copilot's terms. There is also an explanation here. In both cases I've tried to use clear and concise terms.

The Future
I understand differing from Poker Tracker definitions makes it hard when comparing statistics on forums. Therefore at some point I intend to move my definitions closer to Poker Tracker, as it seems to be a de facto standard.

Poker Copilot 1.34 released

After a marathon debugging and testing session last night, I've now released Poker Copilot 1.34. The last changes were

• a fix to the HUD grey bar problem
  
• a correction to "Aggression Frequency"
  
• a change in the definition of "Won at Showdown"

There are still a couple of known issues, but these only seem to affect two specific players. While I'm still investigating them, I didn't want to hold up the release any longer for the wider public.

Many thanks to all who helped with suggestions and testing. Double thanks to Jerome and Matt who both seem to have a talent for finding bugs and tracking them down to specific hands where the bugs could be reproduced.

Download New Poker Copilot beta 1.33

Yesterday you saw the bankroll chart. Today you can download it:

Barring any further bug reports, I've got one issue left to deal with before I can move this from beta to full release. Thanks to all who have been patiently re-downloading the rapidly-changing beta.

Chief Community Officer

In the signature of an e-mail I received from Get Satisfaction:

Amy, Chief Community Officer, Get Satisfaction

First time I've heard of one of those!

New Chart: Bankroll by Hands Played

Every Poker Copilot user seems to want to see graphs by hands played. Currently I've only been graphing by date. There's certain performance issues with the excellent charting component I use that was making this tough. But recently, that component's developer supplied a nice patch to solve these issues. So I now added a bankroll chart:


But please, please, please don't expect all the charts to get this new x-axis. It will be quite a bit of rework. For now, all I can offer is a one-chart taste. Consider it an appetizer for some delicious new charting morsels coming in a couple of months time.

Version 1.3 beta Updated

There were a couple of calculation bugs in the Poker Copilot version I linked to in the previous post. Also, the HUD performance was not good with lots of hands - because I forgot GUI basics (for the techies: do long-running calculations in a background thread, update the screen in the event-handling thread).

So here's an updated version 1.3 beta:
http://pokercopilot.com/downloads/pokercopilot1.32.dmg

New Update with HUD Stats

Today I've put together an alternative HUD view, showing a bunch of stats in tabular form. Here's the new view:


And as it was (and still is, if you choose):

I'm misusing standard Apple buttons. But as I intend to dramatically enhance the HUD over the next month or two, I don't want to spend too much time on this interim solution.

You can download this Poker Copilot update now.

Using Mac OS X System Images in Java

Here's a Java feature in Leopard: You can use the system images, found on Apple toolbars and applications, in Java. They exist in a virtual file system called "NSImage". Here's an example to show the image for "Icon View".

import javax.swing.*;
import java.awt.*;
import java.awt.event.WindowAdapter;
import java.awt.event.WindowEvent;

public class ScratchSpace {

public static void main(String[] args) {
JFrame frame = new JFrame();
final Container pane = frame.getContentPane();
pane.setLayout(new FlowLayout());

/***********/
Image image = Toolkit.getDefaultToolkit().getImage("NSImage://NSIconViewTemplate");
pane.add(new JLabel("Icon", new ImageIcon(image), SwingConstants.LEFT));
/***********/

frame.pack();

frame.setDefaultCloseOperation(JFrame.DISPOSE_ON_CLOSE);
frame.addWindowListener(new WindowAdapter() {
public void windowClosing(WindowEvent e) {
System.exit(0);
}
});
frame.setVisible(true);
}
}

The full list of system icons can be found here. Scroll down the page to see examples of each icon.

Alas, this doesn't work on Tiger. To get around this, you can use Leopard to create a copy of the image as a PNG. Then you can use the PNG. This code creates a PNG file for a specified image template.

import javax.imageio.ImageIO;
import java.awt.*;
import java.awt.image.RenderedImage;
import java.io.File;
import java.io.IOException;

public class ScratchSpace {

public static void main(String[] args) throws IOException {
writeImage("NSListViewTemplate");
}

private static void writeImage(String imageName) throws IOException {
Image image = Toolkit.getDefaultToolkit().getImage("NSImage://" + imageName);
ImageIO.write((RenderedImage) image, "png", new File(imageName + ".png"));
}
}

Incremental Improvements: Every day, every way

I'm big on continuous improvement in software development. Every week, every day, I hope to find something I can do marginally better. If that change makes my one-person show only a fraction of a fraction better, it beats a fraction of a fraction worse.

Some marginal things I have done to make things better for users and to make more profit for me include:

• Using Typinator to store oft-typed e-mail phrases. This gives me the ability to respond to e-mails quicker, with less proof-reading. The second time I find myself typing a certain paragraph, I stop writing, enter the text into Typinator, then use the new shortcut.
  
• Using Get Satisfaction. This reduces my e-mail load and increases the openness of my development, while keeping a healthy level of communication with my best customers. (My best customers are the ones who tell me what they like, what they don't like, and what they want to see next).
• Automating the license key generation system. At first this was a semi-automated process, but still needed me to "flick the switch" to send out a new license key. To do this I needed to be awake and at the computer. This meant that the first thing I did on waking in the morning was groggily sitting in front of the computer to find if anyone had ordered while I was sleeping and was awaiting a license key. Not ideal. But now new customers receive a license key within minutes of ordering. New users are happier and I can sleep restfully.
• Always looking for ways to tweak Poker Copilot for more speed or for easier use. For example, did you notice the newly-added red section in some of the graphs in version 1.3? Suggested by a user, by the way. Or did those of you with tens of thousands of hand histories notice the screen updates are snappier?
  
• Implementing an "empty in-box" policy, systematically handling e-mails to save me time. I check my inbox twice a day, deciding instantly for each piece of e-mail to delete it, to respond immediately, or to move to the "handle later" folder if I can't address it yet.
  
• Every time someone reports a bug, I add a unit test that will fail until I fix that bug. From then on, the unit test ensures the bug stays fixed.
  

There's still plenty of improvements to make, to the processes, my systems, and of course, to my software. Continuous improvement is never a completed task. I can't make everything perfect at once. But I can make things a little bit better tomorrow than they were yesterday.

Every week, every day, I aim to do something marginally better.

Poker Copilot 1.3 beta Available

To my loyal blog readers:

You can download the beta release of Poker Copilot 1.3 (aka the Funky Dashboard version) now.

What's new:

• the dashboard, of course!
• more graphs
• Stake Level summary
• Position summary
• filtering by table size (heads up, 6 max, 9 max, 10 max)
• loads of new statistics
• tooltip help to explain the statistics
• a "Casino Detector" to make getting started easier
• some performance improvements
• some basic help pages
• several bug fixes
  

Known issues:

• The tournament stats for total winnings, and winnings/100 hands are meaningless because they are based on chips rather than buy-in and prize pool.

If you discover any irregularities in statistics, please let me know.

Full Tilt Poker's Matrix

Ay, ay ay. I don't play poker for a week or so while on holiday and Full Tilt Poker goes and releases an update and a new innovation - the Matrix tournament - on me. New features, new ways to break Poker Copilot.

I just gave the Matrix tournament a shot. Lots of fun. Six players play simultaneously on four tables, with points awarded for knocking other players out and for winning a table. I came first on my initial attempt, which naturally made it even more fun!

Unfortunately the Matrix tournament seems to confuse Poker Copilot a little. I'll need to spend some time tweaking the parser.

Back from Albania and in the Black

While I was on my trip to a little-explored corner of Europe, not only did I see some breathtaking Balkan scenery and fascinating Ottoman-style towns, I also passed a significant milestone with Poker Copilot. I have now earned back every cent I put into Poker Copilot. That's right, I've now broken even - and in fact, started earning a tidy profit - on my one-man, boot-strapped, personally-funded show.

Dreaming of Starting a One-Person Software Company?

Peldi wrote a detailed post about the hows and whys of starting up Balsamiq, his one-person software company with a cool name. It's good reading for anyone dreaming of throwing in the day job and starting up shop. My opinion: Peldi had the benefit of skill and luck: skill, in that he seems to be blessed with a broad range of skills one needs for such an endeavour; and luck, in that the niche he chose for his first product has turned out to be a very good one.

My (Java) Competitive Advantage

Matt asked:

what Java IDE do you use? (and recommend?)

IntelliJ IDEA. It costs money ($249 for a personal license), but if you are a professional Java developer it is more than worth it. It is far ahead of the competition in quality and features. They release a new version once every year or so and I always purchase an upgrade immediately. I feel it is my competitive advantage over the hordes of Java developers who use Eclipse or NetBeans simply because they are free.

Albanian Wireless; Poker Copilot Guide

If you find yourself heading to Tirana, the capital city of Albania (and I can't really think of why you would be), and you want free wireless, head to the cafe of the UFO University. Yes, it's really called the UFO University. It probably stands for something poignant in Albanian. It's just off the the central plaza, nearby the Opera. Not only does it have free wireless, but a cafe espresso costs way less than a euro.

Now for some news that has left me gobsmacked, flabbergasted, and astounded - in a positive way. MacPokerTools.com has gone to the trouble of writing a "Guide to Reading Poker Copilot". It's nice indeed to have an ecosystem forming around Poker Copilot.

Blogging from Albania

I arrived in Albania this morning after a long train and ferry journey from Cologne, Germany. As a true geek, I have my MacBook Pro with me. On the journey I alternated between being enchanted by Mediterranean views, watching old South Park episodes, and improving the Poker Copilot unit tests.

I'm yet to find wifi in this unique and chaotic country so I'll sometimes be slow in replying to Poker Copilot questions and issues for the next week or so.

The Casino Detector

The #1 support issue with Poker Copilot is making sure the customer has enabled hand history recording in PokerStars or Full Tilt Poker. Almost daily, someone contacts me because they have this problem. It's causing many customers (and potential customers) unneeded frustration.

Therefore the coming Poker Copilot update has a practical solution: the casino detector. This is shown for all new users. It can also be summoned at any time using Window -> Detect Casinos.


If PokerStars or Full Tilt Poker are not found, the Casino Detector shows download links. If hand history is not enabled, it shows links to instructions on what to do.

I've got to write some docs...

...or help files, or some nice HTML pages explaining how to use Poker Copilot. I foolishly - even arrogantly, perhaps - claimed that Poker Copilot would be so easy to use, there would be no need for documentation.

Then I noticed some of the Google search queries that send people to my site. Queries like "how to use Poker Copilot", "Poker Copilot help", "Problems with Poker Copilot". In a way, I feel a little chastised. :-(

So over the next couple of weeks, I'm going to resolve this, hopefully in time for the release of Poker Copilot version 1.3 (the one with the funky dashboard).

Poker Copilot 1.3 Code Freeze

As of ..... now, I've frozen the code for the next update of Poker Copilot. The only development now is testing and bug fixing. This means I'm on track to release update 1.3 in mid-October.

Naturally all existing licenses will work with the new update.

Announcing the Poker Copilot Dashboard

There's a new screen in the upcoming Poker Copilot update I've particularly enjoyed working on: the dashboard. It's given me a chance to try something creative and new, as well as a chance to apply and hone the skills from my professional work with data warehouse user interfaces.

Hopefully it will be invaluable to online poker players, and will make Poker Copilot something quite unique amongst poker tracking software (other than simply being the only actively-maintained Mac OS X product).

Enough words, time for pictures. Here 'tis:


The released product may be a little different, in terms of layout and statistics shown. Possible future enhancements include customisable dashboards and HUD-style mini-dashboards.

Looking for Auto-type Software

I type similar phrases or paragraphs often in support e-mails. There's software that will auto-type repeated fragments for you, to make e-mail support easier. I found Mac OS X software to do this: Typinator. Great name, I'm almost want to buy it because of the name alone.

Here's an example of how I use it. In an e-mail I type:

rgds

Typinator instantly replaces it with:

Regards,

Steve McLeod
Poker Copilot

In fact, it just did it now in this post. Twice.

Pretty neat. Before I buy this software, please let me know if you are aware of a better product. Free would be good, too.