the DC Way

Over a year ago the first episode of "The Eightfold Path to Poker Enlightenment" appeared on DeucesCracked and since then it has become one of the most popular and beloved training series in online poker. If the word "classic" can be used to describe a poker coaching video series, the EPTPE would have to be considered one.

If you haven't watched it yet, The Eightfold Path is not a traditional strategy series. It's a series aimed at fundamentally altering your mental and emotional approach to poker, and tilt, as a whole. It's narrated by Tommy Angelo and Wayne Lively and draws from Tommy's book, "Elements of Poker" as well as the the Noble Eightfold Path of Buddhism. Each episode deals with a different aspect of how to mentally approach poker in a calm and balanced way, and is intercut with soothing piano music played by the man Tommy Angelo himself. .

But piano music alone does not a video series make. So how come The Eightfold Path remains so popular amongst DC members?

Firstly, The Eightfold Path can be useful to poker players of all games and stakes.  A midstakes HU PLO video series could be the most comprehensive and brilliant HU PLO series on the internet, but if you're exclusively a 6max NL grinder, it's not going to do you much good.  The Eightfold Path, on the other hand, can help anyone who has ever tilted, and that audience encompasses pretty much anyone who's ever so much as touched a playing card.

The Eightfold Path stands apart as a video series because it just feels different. There are a lot of video series on DC and you'll find a lot with the familiar poker table backdrops and discussions of ranges and equity and math, but only one that links poker and Buddhism. When you think about it, it's hard to think of two concepts more antithetical. Buddhism is a devotion to the severance from attachment and material things and poker is about tricking people into giving you money so you can buy an iPad. But that's exactly why the series works. For a lot of players, it provides a completely new way to think about the game. It's like exercising a brand new muscle.

Most studious poker players spend their time in a certain mindset that becomes familiar through repetition: grinding thousands of hands, highlighting hand histories in a tracker and posting them to a strat forum, IMing stats and reads to a poker buddy on Gchat. Videos, conversations, and days can start to bleed together in a soup of nomenclature and numbers, a left-brained lifestyle that starts to perpetuate itself.  Unchecked, that approach can become a little unbalanced. The Eightfold Path is a chance to step back from all of that, from the EV calculations and the flush draws and the endless repetition, and see the entire experience from a completely new perspective.  Where other videos champion immutable math, the Eightfold Path touts the ephemeral benefits of breathing. Where others talk about mericlessly applying will-crushing aggression, the Eightfold Path reminds you to be calm. Where other series show you how to focus, the Eightfold Path shows you how to see.  

Then there are the artistic flourishes that set it apart. Tommy and Wayne, in a Doonesbury-esque touch, appear onscreen only as icons, the former a baseball cap with a curved brim, the latter a cowboy hat. Wayne the cowboy represents the brash, bold, tempestuous gambler and Tommy's well-worn ballcap perfectly symbolizes a peaceful familiarity with the game. Not to go all college sophomore on this, but everything, including their diametrically opposed personalities, the dynamic of student-teacher, even songs played forwards and then backwards, serves to impress ideas of balance, duality and completeness, which is of course what the entire series is about. And just when it seems like everything is as rigidly organized as an origami swan, Wayne will say something to cut through it all, something like, "There should be a poker movie called 'The Nutcracker'."

When I started out watching The Eightfold Path for the first time, I had it in a window on one half of my screen while I browsed forums and read e-mail on the other half. Yeah, yeah, I know I should always devote my full attention to videos, but as we all know attention sometimes wanders.  But slowly, as time went on, I found myself focusing more and more on the video.  Eventually, I had the video full-screened and I just listened to the words and music as I breathed slowly and deliberately. Just watching the videos became almost meditative. When the music stopped and the video ended, I felt like I had just come out of a relaxing steam or something. It was the first time anything related to poker had produced anything even resembling a state of calm.

As Tommy points out, "Since no one is tiltless, every one can tilt less". Very true. And to help us tilt less, we can all watch the Eightfold Path a little more. 

-JMac

We have had a few threads recently asking about the future of DeucesCracked.

First off: thank you.

All of you have helped contribute to making this an amazing poker training community. It genuinely makes all of us happy to see how much people care and that we are all looking out for each other.

Secondly: we aren't going anywhere.

In the short term, some of our US coaches will likely be focusing more time around coaching-based videos (hand reviews, member-submitted video reviews, etc.) than live play videos. Fortunately for us, we have a superstar roster of coaches centered in Europe and we will be working with them as well. We expect to have more content that relates to live play, including Bart Hanson's upcoming series and more theory-based content as well. We got into this business because we love teaching people to play poker, this will never change.

Here's where we could use your help: when it comes to content, we'd love to hear your thoughts. This is new ground for everyone, so getting feedback from you on what you like and what you don't is more important than ever. Additionally, we have plans to make it easier than ever to subscribe, so signing up for offers like our PartyPoker, Kickoff Poker (Entraction), Betfair and other offers not only should be great for you (we'll be offering more details on these offers soon - but we'll have Free DC, free HEM, and more), but will also help us continue producing the excellent content you've grown accustomed to.

Feel free to post in this thread with any questions you've got. We'll do our best to answer them.

So you're in the middle of yet another frustrating stretch of poker hands. Maybe you've been stuck at the same limit for a lifetime. Maybe you feel your game has stagnated. And you think, "I know! I'll get a poker coach! HE'LL tell me all the things I'm doing wrong! HE'LL show me how to crush! Then I can move up, make fat stacks and buy a Lexus! THEN Sarah will return my e-mails!"

...or maybe you're just an ambitious, dedicated player who thinks a coach could really help you learn.

Either way, you're thinking about getting a poker coach. Excellent! It can be a huge step forward. But there are a few things you should consider first to make sure you get the most out of your coaching experience.

PREPARATION:

-Study Poker Theory. More of it.

Yeah. It's hard. It can be boring. It's math sometimes. Doesn't matter. Suck it up. If this game were easy, it would be the only job in the world.

Some students come to a coach looking for the "What" of poker (raise this spot, fold this hand, punch this opponent, etc.) when really they should be looking for the "Why" of poker. The stronger your understanding of poker theory, the more creative and effective your play can be. Poker theory is probably going to be you and your coach's common language. Get good at it.

Some series worth watching (and rewatching):

-The Mathematics of NL Hold'Em

-The +EVolution of a Poker Player

-The Eightfold Path to Poker Enlightenment

-Calibrate Your Expectations

Coaching is not going to magically fix everything. It is not going to plug all your leaks, prevent you from having downswings ever again, or get you to finally start going to the gym regularly. Coaching is just one tool out of many in your poker development. Make no mistake, it can be a very effective tool (perhaps the single most effective tool), but you will still need to post and discuss hand histories with players you respect, study theory, watch videos and play smart if you want to really grow.

-Make a video of yourself playing

Here's a cool idea: record yourself playing a session using Camtasia or other screen-grabbing software and record your commentary as if you were making your own strategy video for DC. Show it to your coach. By actually articulating your thoughts out loud you make it easier for your coach to understand where you are coming from and spot possible leaks in your thinking. Additionally, the very act of verbalizing your thought process can make you conscious of things you didn't even realize you were doing. If you don't know why you're doing something, there's a good chance you're making at least a small mistake. Here's a tip: record yourself playing for about three hours instead of just one, then pick the one hour that has the most interesting situations to show to your coach. Watch Episode 6 of "The Haj School" for more information.

PICKING:

Now you're ready to choose a coach. So how do you know who to pick?  DC has [quite an arsenal] of coaches available. Firstly, you should know you can sort coaches by game type, stakes and hourly rate, so if you just need midstakes PLO coaching, you can sort for that.  Here are some other things you can do to make sure you get the right coach

-Have some goals

Do you want to move up in stakes or are you just looking for a bigger winrate at your current limit?  If it's the former, maybe you should look for a coach who has been successful at higher stakes.  If it's the latter, maybe you'd prefer someone who plays closer to your own limit.

-Check out the Coaching Forum

There is a [whole forum] on DC dedicated to coaching. Spend some time looking over it, there's lots of valuable information there. In particular, check out the thread: [ASK TEAM DC ANYTHING ABOUT POKER COACHING]

-Talk to your poker friends

Have any of your friends had coaching? If so, chances are they can provide some recommendations to guide you towards a coach that could work for you. Don't have any poker friends? Get some. Seriously. As FoxwoodsFiend points out, having peers that you can talk poker with is HUGE.

-Consider group coaching

One-on-one is not the only coaching option available on DC. Group coaching is available, which can be a great option for students looking to find students with similar goals. Sometimes it's easier to  motivate yourself to keep going if you're part of a group, and it's a great way to meet other like-minded poker friends you can talk strategy with later. Also, group coaching can be cheaper!  Check out the 'Secret HQ' forum to learn more.

-If possible, watch videos made by your potential coach

If you've narrowed your list of coaches down to a few names, check to see if they've made a video series for DC and watch their videos. In addition to being a preview of the caliber of player they are, watching videos can be a good way to predict coach-student compatibility. Pay attention to their style of teaching. Does what they say immediately make good sense or are you having a hard time following their logic?  After all, everybody learns differently and what is helpful to someone else might not be helpful to you. WiltonTilt is an excellent poker mind and teacher, but there's no point in hiring him at $500 an hour if everything he says flies over your head.

All that should help you pick out a coach who's right for you. Take the time to do it right. As jk3a said, "It's a major purchase. DO YOUR HOMEWORK" 

WORKING:

Once you've prepared and picked your coach the real work begins.  Here are some things to keep in mind to make sure you get the most out of your actual coaching sessions.

-Be prepared

If your coach is going to be sweating you online, be on the tables ten minutes before your coaching session begins so you don't waste the first couple minutes of your session on waitlists.  Make sure Skype, Hold'em Manager, or whatever other software you need is already up and running. Have a notepad or a text editor ready to take notes. Have a few hand histories or theory questions ready in case few interesting situations come up during your session. Remember, you're paying for every minute your coach is there. Make each one count.

-Ask questions!

Many students think the simple act of listening to a coach for an hour a week will automatically make them a better player. Unfortunately, it just doesn't work that way. It needs to be an active, engaged process. Ask questions! Coaches complain all the time that their students don't ask enough! Believe it or not, most coaches got into coaching because they like coaching! Helping students improve is a tremendously satisfying thing. If you don't understand something your coach says, say so!  That's why he's there!  You're not doing anybody any favors by nodding and not learning.

-Do Your Homework

Now that you're actively thinking about poker and (hopefully) not playing on auto-pilot, it's time to keep your poker mind engaged.  Post hands, watch videos and keep track of spots where you've had trouble to discuss with your coach during your next session. Maybe after a few sessions with your coach you can make another video of yourself playing. Keep yourself actively engaged in all aspects of poker learning. Think of your poker brain as a body and do all the exercises you need to work all your muscle groups.

Coaching can be a richly rewarding experience that helps you grow as a player and propels you to greater success - or it can be a disappointing waste of time and money. It's important to make sure that you pick the right coach and can use coaching effectively.  There are a lot of great poker minds out there willing to teach, and frankly, it's tough to reach your full potential without one. After all, "even Michael Jordan had a coach."

For more information on getting the most out of your coaching experience, check out Episode 7 of "The Haj School" by WiltonTilt.

By JMac

It's been a tough few days in DeucesCracked land.

Over the last several days, we've been getting hit regularly by a fairly large DDoS (Distributed Denial of Service) attack. We don't know who is responsible for the attack or why they launched it. The first appearance of it was two weeks ago - the attack lasted a few hours. It was the first attack we had ever been hit with, so after rolling out an initial set of fixes, everything seemed ok.

Then last weekend came. We got hit again - hard. Just like the first time, we started getting hit with hundreds of thousands of connection requests from computers across the globe. We couldn't effectively firewall the traffic as every time we shut one network down, another would start up. After multiple days of playing cat and mouse - coming up with a defense for the DDoS for a few hours, only to be hit by a new vector of attack a few hours later - we implemented a more permanent solution. I'm not going to go into details about the specific solution here, but the unfortunate side effect of it is that it's required some performance tuning on our end. That took us to this morning.

And now we're here. Before we got hit with the DDoS, we had scheduled one of the largest overall site updates that we've ever attempted to undergo. We've been planning for this for a few months and have been working round-the-clock over the last 6 weeks on it. I detailed what went into this in my last blog post (which I thought would be my last technical post for a while - I guess not). We had tested it on development servers extensively, we have automated tests for it, and tested it on a staging server as well.

So today - with things looking better - we launched Rails 3. And watched the site come to a screeching halt. It turns out that one of our dependencies, memcached (which we use as a session store), segfaults when combined with some of the dependency upgrades required as part of the launch to Rails 3. After digging around and doing some pretty impressive sleuthing, the guys at RailsMachine discovered this, helped us upgrade to the latest version of REE (1.8.7 patchlevel 334), and fetch a new version of memcached. Unfortunately, that wasn't the end of things - after deploying and fixing everything, we were still left with a server that was spending lots of time in garbage collection, eating up all of our memory and spare resources. An hour later (basically now), we'ved rolled back all of the changes from the past day.

So you're thinking - great - things are much faster now and I'm glad I can watch videos again. But what is DC doing to address the downtime?

For those of you with current subscriptions (or were in a free trial period as of today), we'll be adding 3 days to the end of your subscription. For those of you who are on the basic plan and were sad you couldn't get to podcasts or post on the forums, we'll have something in store as well. We'll be sending an email about this shortly, once we're 100% convinced that all our problems are behind us.

Again, we really appreciate your patience over the last few days. We hope it's something we never have to go through again. 

In the past two posts I've talked a bunch about what we do on a daily basis. For the final installment, I'm going to talk mostly about what we've been up to lately in the development front, with a small bonus chat about how our dashboard works.

The Journey to Rails 3

Over the past month and a half or so, we've been grooming our entire codebase for a massive overhaul and update. As of posting this, we are running on Rails (a framework for Ruby) 2.3.11. As of Monday, we'll (hopefully) be running on Rails 3. What has this meant for DC?

Well, the biggest thing is that we've had to review, refactor, and sometimes remove tens of thousands of lines of code. Including the Rails plugins (pieces of code that help us do things like upload & share videos on Amazon's Cloudfront streaming servers), our codebase is over 100k lines of code, which means that migrating all of it over to an updated framework has been a LOT of work. I'm sure some of you have seen me post that we can't focus on a new feature or idea because "it's not a priority" - I haven't said this lightly. While the overhaul has been necessary and will help us future-proof our codebase to a degree, it's certainly been painful having to prioritize a large-scale upgrade like this when the outward appearances aren't so visible. It's always tough to have the true hallmark of whether you upgraded your site or not be that no one notices - if no one notices, we did our job very, very well. So why upgrade?

1) Inevitability

While we've generated quite a bit of code from scratch, a lot of the code we rely on is open source. Over time, the people who contribute to the plugins that we use (active merchant for billing, for example) tend to move away from deprecated code, which in turn means that it's best for us to do so eventually as well. There's a careful balance that you have to walk between upgrading too soon and upgrading too late, but regardless, upgrading is inevitable.

2) SEO

One of the biggest areas that Rails 3 should have a positive impact for us is in its SEO-friendly URL structure. Google is an ever-changing beast, and we've constantly been fighting a battle to make sure that the majority of our content (2000+ videos, 530k+ forum posts, 65k+ topics and thousands of blog entries) gets indexed. Moving to Rails 3 will help us get a uniform URL structure that we can set now and stick with in the future to help guarantee the indexing of this content, which helps our performance on the 'long tail' of search queries and incoming traffic to DC in that regard.

3) Performance

While initially a lot of reports coming in from people running Rails 3 were that they were running into performance issues, this seems to have changed over time. The biggest areas where Rails 3 has improved performance are in the generation of SQL queries through ActiveRecord and Arel. For a site that is pretty database intensive, that's a good thing.

Reads if you're interested in more geeky details: 

http://weblog.rubyonrails.org/2010/11/15/rails-3-0-3-faster-active-record-plus-plenty-of-fixes and http://www.williambharding.com/blog/uncategorized/rails-3-performance-abysmal-to-good-to-great/

4) Bonus content! The dashboard.

So - I've never really talked about how the dashboard works before but I've seen some misconceptions (based on people guessing how it works, I guess?) and figured I'd post about it now.

The dashboard is largely powered by Redis. If you don't know about Redis yet and you're a techie of any form, you should. It's a memory-based Key Value store (or, if you prefer, a data structure server). It's very flexible, and while it doesn't have some of the abilities of a relational database, it's incredibly fast and incredibly powerful. Here's how we use it:

On the dashboard, we have Latest Videos, which is cached data. It's lightning fast since it only needs to be updated once a day or so - whenever a new video is published, it gets cached again.

My Threads, however, is a lot more complex from a data perspective. We take a list of the 15 most recent posts you've made (we'll be changing this shortly to make it more useful). This is generated from a simple hash (list of 15 topic IDs scored by the time your most recent post was created), then sort those topics based on which was updated most recently. If you have alerts, we append the 10 most recently updated alerts (these are RSS or email-tracked threads) to that collection as well. If you're a coach, we also add videos that you were the author on, so that coaches always see their recent videos on the dashboard as well (making it easier to answer questions).

That alone is a fair amount of data - but that's not the area where Redis is really truly necessary. Where it becomes necessary is in figuring out which of those posts was most recently read by you. To do that, we store the maximum-post-read ID for every topic you've read on DC (when logged in, of course), so we can calculate the # of unread posts you have in any thread. Then, as if that wasn't enough, we need to load the most recent unread post into memory (so you can hover over the title and read the body of the post) and the URL of that unread post, so when you click the thread you either jump to the most recent unread post or to the end of the thread (if you've read all the posts in the thread). Latest Threads is done similarly (in terms of generating links to the posts and the unread counts), though it's much easier to track all of the recent public threads than it is to track all threads an individual is interested in.

Of course, sometimes things go wrong with that setup - as some of you have noticed, there are occasionally missing topics from the "My Threads" section. The tradeoff with the lightning-fast speed of Redis is that we're on the bleeding edge of technology - so simple things like "SQL browsers" that exist for troubleshooting relational databases don't exist yet for Redis. With 1.5 million keys being stored in our database, that makes things kinda tricky to troubleshoot and fix - but we're getting better at it. In the next few weeks we're hoping that updating from our current version of Redis to Redis 2.2.2 should help significantly, and refactoring some of the code (reusing code from our Reader) to continually refine the dashboard and make it more and more useful.

Ok - that's enough for A Day in the Life for now. It's been fun blogging about this and if you guys want more, post in the comments here and the next time I've got something cool to post about I definitely will. The @deucescracked blog is going to be a place for posts like this, posts about coaching and occasionally humorous/other blog posts, and I've been dominating it with techgeek stuff for the past week, so it's time to give someone else a try for a while.

peace,
Rob

Last week I talked a bit about the process DC uses to manage workflow. This week I'm going to go into a bit more detail about how we determine our priorities for what needs to be fed into our workflow.

I've said this before - I'll say it again. Our #1 feature is speed (Fred Wilson Agrees). We keep a close eye on our system performance in a few different ways. From a general (and very subjective) standpoint, we eat our own dogfood - that is to say, we use the site daily. That helps us keep fairly alerted to threads that are alerting us to performance problems and bugs, and also lets us see which areas of the site are holding up well and which might require more resources or possibly refactoring.

On the objective end of things, we monitor our site's performance with Scout and with NewRelic. I've mentioned both of these before but I'll go into them a bit more in detail now.

Scout gives us access to a ton of server monitoring and server performance data. Our primary use for it is in application monitoring - to make sure that the site is up and running and that the servers we rely on (MySQL, Rails, or Redis) are all up and running smoothly. It also contains some basic data that we can use to analyze slow elements of the site, like slow database queries or pages that are performing poorly.

In general, Scout is step #1 whenever we're about to dig into application performance. Once we see something that stands out - for example, a higher-than-acceptable standard deviation of page response on our Forums (TopicsController#Show)...

...we move to other tools to work on understanding the root cause of the performance issue and to see what we can do to improve it. This is where New Relic shines.

New Relic is indispensable for digging deep into your application. It provides a general application overview so that you can see how individual elements of our application are performing (in this case, MySQL, Ruby, and Memcache). Some of the real utility of New Relic comes in its ability to individually investigate page actions and report back on the individual traces within pages.

Rather than just knowing what pages are slow, you can look into the individual components of a page and dig from there. For example - based on our earlier example of a high stdev in page response time (which would indicate a page that is alternating between very fast and very slow), we can take a look at one slow trace (New Relic automatically logs each situation where a page takes longer than 2s to load) here:

What does this tell me when I look at it? Well, it looks like Redis (which we use as a data-caching store, keeping track of all of the latest posts people have read, etc) is occasionally taking a LONG time to respond to this specific HGET query. That's odd behavior given the way that Redis works, and is something that I'll make a note (in Pivotal this time around) to investigate further. The page is fast almost all the time, but occasionally is very very slow, and is definitely something that warrants further investigation.

In general, we aim to have DC pages served up in anywhere between 1/5 and 1/2 of a second, depending on the complexity of the page. Our average response time has historically been about a third of a second - which would be slow for a company like Google, but is pretty good given two things:

1. Our size. In the past we've served everything off of 1 server - we've bumped this up recently to 3 servers to accommodate our growth over time.
   
2. The amount of customized data we send. Customized data is the bane of an application - it's very easy to just spit out 25 pages of recent forum topics, but the moment that you commit to tracking things like "last post read" (which is different for each of the users who use the site), your application becomes orders of magnitude more complex. Managing this complexity is a full time job, but is something that we're committed to, given how much it helps our users contextualize and follow the information that they're reading.

I was going to end this post here, but currently we're under a bit of a DDoS (distributed denial of service) attack which is trying to bring our site down. This is where the guys at RailsMachine deserve a shout out - they help us with server monitoring 24x7, and when things like this come up, they're immediately on call to help out. They also help with other server configuration issues and monitoring to help DC as it grows. Thanks, guys.

So - that's it on the performance management and tracking mode. Next week I'll be back one last time to talk about what we've been up to lately and what's coming up from us.

If you frequent the DeucesCracked forums, you've probably seen me post something about "the list" or our "development queue" before in threads about feature suggestions. I'm sure some of you have been curious about how we manage that, and what things look like behind the scenes for the DC crew. This one's for you.

Agile Project Management

We're a small team - 1 core developer (@wiseleyb), 1 UX-pert (@aarongerdes), 1 jack-of-all-trades (@arussell), and @deathdonkey and myself (@robcole) for some small cleanup tasks (DeathDonkey also handles the majority of the hand converter updates and coding). We do our best to follow general agile practices (http://en.wikipedia.org/wiki/Agile_software_development), releasing updates to DeucesCracked very often (often multiple updates daily), and we coordinate this via a few tools. If you're already yawning, it's probably best to skip the rest of this one.

Pivotal from Pivotal Labs

Pivotal is our go-to day-to-day management for development. It's a good blend of bug tracking, project development, and to some extent, project scheduling. Each of the scenarios we write in Pivotal tells a story to help communicate to our team what the intent behind the feature/bug/release is. This also helps us write automated tests around the story to make sure it doesn't break in the future (this is TDD - for those of you curious: http://en.wikipedia.org/wiki/Test-driven_development).

On a fairly regular basis (every week or two), we go through an iteration plan, which basically means that we set a priority for what will be worked on in the next week. The priorities usually are performance-based issues first (i.e. - keep the site running quickly), followed by site improvements that impact large portions of the site, and then experimental feature releases/changes. For each improvement we make at DC, there are a fair amount of things that have to be done behind the scenes first to make sure that it's implementation doesn't mess with other critical systems.

We couldn't manage all of this easily without Pivotal. There are other systems out there and I've tested a few of them, and each of them have their individual strengths and weaknesses. Pivotal's software has been great for us, always reliable, and has helped us keep the team moving consistently.

Campfire from 37signals

While we occasionally use AIM/Gtalk/other IM programs to communicate, the majority of our communication is done using Campfire from 37signals. A few of you have probably seen this before, when I've opened up a chatroom to discuss features/bugs/experiments that we're running when I need real-time communication. That's Campfire's bread and butter for us. It's also a nice place for us to be able to paste near-realtime notes for other people to look at in the future, and to break out individual features/stories in Pivotal so we all understand what the requirements of pieces of code are.

As with any other real-time chat tool, Campfire can be a huge distraction at times, but that's also part of the fun of it. You can always see which of us is at the end of our work day by the amount of Reddit links being reposted in there. I'd strongly recommend anyone who is looking to manage a remote team check it out.

Coming Soon

In the next installation of A Day in the Life, I'm going to detail a few of the applications we use to keep things running smoothly behind the scenes at DC, including performance monitoring tools and a little more about our code and behind-the-scenes server ops, and how some of the key portions of the site work... (to be continued)

 -Rob

“For organizations, the why is born out of the very reason or purpose the organization was founded. The problem it set forth to solve or the change it wanted to make.”