A Collection of Updates: Life, Poker, and Exciting News

This post was originally published on October 15, 2021, on my personal website, Lukich.io. I have since consolidated all of my poker-related content by reposting it onto Solver School. Please note that the analysis spreadsheet discussed in this article has been retired and is no longer publicly available. If you are interested in discussing custom analysis or ways to replicate for your own work, reach out to me at [email protected].

And just like that, it’s October. This past year has dragged on and flown by at the same time…if that makes any sense.

I decided to try something new today. My normal writing process is as follows…I typically start by outlining some of the key topics I want to explore. Then I fill out some of the details underneath and attempt to weave it all together into a cohesive blog post. It may not be the most efficient process, but it has typically worked well for me.

However, when I started writing this post a few weeks ago, I immediately encountered some challenges getting into a steady flow.

It’s been six months since my last post, so when I built an outline, I realized I had much to share. After a few days of false starts, here is the general list of things I came up with to write about:

• An overall update of 2021, how focusing on taking care of myself has greatly improved my mental and physical health, and an exciting new project I’m launching this winter

• Some thoughts I had about using comparisons as they relate to poker data analysis

• The latest update I made to the Flop Analysis Workbook that I sell within my product store

Just looking at that outline…it’s obviously a disparate list of topics. I tried to plow ahead with my normal process, and I struggled to come up with transitions from one topic to the next.

After a couple of days of banging my head against the wall, I gave up on trying to tie everything together. Instead, I decided to separate each bullet point into its sub-section below. Call it a collection of mini-blog posts instead of one long update. So, while I hope you read the whole thing, I realize it’s a long post. So if you’d like to navigate to any individual section, you can easily do so by clicking on the links in the outline. Enjoy!

An overall update of 2021

When I last posted in March, I was caring for my mental and physical health. And while I’ve been up to a lot with my family and work over the year, I’ve tried to prioritize taking care of myself as much as possible.

Since January, I’ve developed a series of consistent habits that have included:

• Lifting heavy weights 4x per week — I bought a squat rack, bar, and bumper plates from Rogue and carved out an area in my garage to work out. It’s one of the best investments I’ve ever made.

• A daily stretching and meditation routine when I first wake up — I added my stretching routine in July because my muscles often felt tight, and it has increased my range of motion considerably. I still feel like I struggle somewhat with meditation. I’ve tried Headspace and currently use Calm. I’ve recently been recommended Sam Harris’ Waking Up, which I plan to try.

• Drinking more water — I started drinking a gallon of water a day while doing a 75-Hard challenge in January. The first week was miserable, but I got used to it after that. When the challenge ended, I kept the habit of filling my gallon water bottle up and drinking it daily.

• Fixing my diet and mostly eating non-processed foods — I always ate well for the most part, but I’ve tried especially hard to stick to meats, vegetables, fruits, whole grains, and good fats while cutting out junk food and sugar.

• Taking long walks daily — Again, another addition from the 75-Hard challenge that stuck. I needed to get two 45-minute workouts per day as a part of the challenge, so I used a 3-mile walk (2 loops around my neighborhood) as one. I found that I enjoyed the time outside by myself. That time feels more meditative than my morning session, to be honest. I do my best thinking on that walk and often come into the house with a few ideas to write down.

• Spending time outside — The climate in Michigan suits me perfectly. I never wanted to go outside during summers when I lived in the DC area — it’s so hot and humid. It always felt gross to leave the house and move around. The weather is a little cooler and far less humid up here. I don’t mind the cold and enjoy playing with the kids in the snow, so winter is fun, too.

• Sleeping at least 7 hours per night — I’m not as tired during the days now and don’t have to drink coffee all day. I have 2 cups in the morning and am good for the day.

• Talking with a therapist regularly — I started talking with a therapist last year, and it’s just been nice to be able to dive into my thoughts and feelings with someone. I think that I understand more about myself. I used to think someone needed to feel “broken” to see a therapist. But now I realize how beneficial that time can be. I think everyone would benefit from regularly seeing one.

• Drastically limiting my social media time — This is obvious, but I was doom-scrolling too much in 2020. I deleted Facebook and have limited my Twitter and Instagram time to early morning while drinking my coffee or late evening before bedtime when I’m winding down on the couch. A+ decision.

I know these are all seemingly obvious and repeated ad nauseam by everyone on “self-optimization Twitter"” But they do work. None of these have been a “magic pill” to make everything perfect. And yes, they take time and effort to include in my day. However, the cumulative effect of consistently prioritizing the things on the above list has drastically improved the quality of my life.

To start, I feel great. I turned 40 this summer, and I think I’m in better shape now than when I was 20. When I was in college, I skipped gym sessions regularly, viewing it as a chore. Now, I rarely miss a session, and never when I’m at home. When I do have to skip a workout due to travel, I feel sluggish and low-energy for a couple of days.

My mental clarity has improved significantly. The lifting sessions, long walks, therapy, and (I assume) meditation have all helped me to bring things back into focus. 2020 fucked with a lot of us. Finding time to sort things out in my head has helped me put things in perspective and prioritize the things that I care about most: spending time with my family and my health and happiness.

It has also given me the confidence to make a significant life decision. I recently gave my current company notice to resign. I don’t officially close out my work responsibilities until November, but the wheels are in motion to leave my job. After working in the corporate world for years, my heart isn’t in it. I was bored, unmotivated, and needed a new challenge.

I’m not ready to announce the specifics, but starting in November, when I do leave my job, I will be devoting 100% of my time to launching a new project within poker that I’ve been thinking about for the past year.

The communal response has been awesome since launching this site in January 2020. I didn’t think many people would read a personal blog about data analysis and poker. To my surprise, it has reached many people from around the world. Many of you have reached out to me to discuss some of my ideas, ask questions about the data analysis and insights, and share some of the work you are doing to dig into the game's mechanics.

A consistent theme that I’ve noticed is that many people from all around the world are interested in exploring poker concepts through data. And I think there’s an opportunity to dig into that further.

I have the website up and running. It’s a shell, mainly there as a placeholder to build an initial interest email list. I also have some branding and a logo designed (as demonstrated to the right), which I will be showcasing on some hoodies and t-shirts while I’m in Vegas next week (more on that below). There’s not much more to say than that I plan to launch in January 2022 and am very excited to share my 20 years of data analysis expertise with the poker world.

As I have more updates in the coming weeks, I’ll share them. But eventually, all the work for that project will stay on that platform, keeping this site separate. So make sure you sign up for the mailing list over there to get all the latest timing on its launch as soon as it comes out. You can also follow my new social media channels on Facebook, Instagram, and Twitter. There’s nothing there now, but more to come very soon.

Comparisons

I also want to write about poker in this post. I suppose it’s more of a data analysis method, but I’ll use a poker example to demonstrate it. And that is the method of comparison.

There are many sophisticated ways to analyze data. Comparison is not one of them. It’s a simple concept to understand because we do it every day in our lives. When buying something, we compare different product features and prices of various options. When dating, we compare our compatibility with someone to how we have connected to someone else, such as in a former relationship. When playing a game, we compare our performance to a competitor or a previous high score (an example of a benchmark).

To back it up a bit, I’ll define the methodology. The comparison involves looking at two or more data points to evaluate against each other or a defined benchmark (usually an aggregate descriptor, such as the average). We typically use it to evaluate points side-by-side to understand differences, rank several items in a category, list key segments against one another, identify patterns or outliers in data, and in many other ways.

Back in March, Berkey invited me to lead the S4Y Mastermind about studying poker using solvers. I explored how I framed part of my solver-related study in that session. Without talking about the concept specifically, I spent much of the 2-hour course (and subsequent 1-hour Q&A session) demonstrating the concepts of using comparisons of solver data to infer insights and derive strategies.

As I prepped for the WSOP, I have been working with Floptimal to familiarize myself with ranges at various stack sizes. I believe Floptimal’s tool facilitates comparison analyses and any commercial poker product I’ve seen to date. I understand the challenges involved in presenting a lot of complex data in a simple-to-consume format. Floptimal’s UI for doing this is just outstanding, allowing for easy comparison of different data points and the ability to pivot to multiple dimensions. I highly recommend it as an excellent study tool for preflop ranges.

When comparison is used well, it’s an effective way to make data analysis more accessible. As a result, when helping individuals start utilizing data-driven tactics to get better at poker, I often point to comparison as an initial methodology. By starting with a data point and comparing it to one or more reference points, we ask, “In relation to what? The insights, when compared to an anchor point, can often drive home some important ideas.

The challenge with comparison is that it’s better used as a visual methodology. Or said better, it’s much easier to compare things when visualizing the differences and/or similarities of two or more items as opposed to simply looking at raw data points. As a result, we often have to categorize, organize, and/or translate the data into a visualization to get it in a format where we can make good comparisons. I’ll demonstrate with an example.

I created a simple aggregate report to analyze a flop scenario, focusing on a single-raised, UTG vs BB pot in a 6-max, 100BB cash game. I ran an aggregate report for 25 flops within PioSolver. I put the output into a table and gave it some minimal formatting to make it somewhat readable, but I have otherwise done no manipulation to the data. Based on the chart below, how easy would it be to derive some insights?

Pretty hard, right? It isn’t easy to sort through and understand anything about the data set. I think most people’s eyes gloss over when they see a sea of raw data like this, mine included! But with a little bit of simple visualization work, we should be able to turn this table into a way to utilize comparison to learn new findings.

For instance, we could quickly sort a column by a metric we may want to use as a baseline value to start the comparisons. We can also average all boards to set a benchmark to which all other points can be compared. That baseline value will give us an idea of the entire formation across all 25 boards, giving us a better idea of how “good” or “bad” different boards are for our range in relation to the whole group.

Here’s that same chart from above with a bit of cleansing:

I didn’t do much, but that little bit has helped us compare some data points.

• First, I sorted by the IP EV. I was always taught to start with the most impactful metric (or KPI in the business world), and I can’t think of a more important one in poker than expectation value. I’m assuming we’re examining from the UTG perspective, although we can (and should!) certainly conduct this analysis for defensive positions.

• I added some conditional formatting to a second set of data points — in this case, our strategic action frequencies (bet 2/3 pot, bet 1/3 pot, and check). These three metrics will always add up to 100%, so grouping them for analysis is beneficial. I switched the column orders from the first table and added data bars to identify the relative percentages to one another quickly.

• I included gradient shading on the three main success metrics that Pio outputs — EV, Equity, and EQR. For each column, the average value is white, with the above-average values shaded green and the below-average values shaded red. Since EV sorts the table, that column will have a continuous green-to-red flow. But the other two metrics don’t look as similar. Usually, the three correlate somewhat with one another. Where they don’t can be an interesting point to investigate further and figure out why.

• I also added data bars on the Global % metric. This represents how frequently we get to this game node. In this case, it’s also the measure of how frequent the BB checks. In most cases, the BB checks almost all of its range. But some instances have lower values, meaning the BB presumably donks a higher frequency. Insights like this are important to understand when analyzing because they assume something about an opponent’s strategy.

• I created an average row as a benchmark to compare data points against. Now, technically, I would have to do a bit more math to give a true “average” of these boards since they appear at different frequencies relative to one another (they’re also only a subset of flops in the game), but for this demonstration, this pure average is fine.

• Finally, I dropped all number significance to 1 decimal point. We don’t need to get more granular than that in this view. The extra precision isn’t helping anything. It just adds unnecessary complexity.

And these changes have helped! We can quickly look at the various boards and compare them to one another much more effectively. But I will introduce one more addition that can help power this even more.

Comparisons are extremely valuable when paired with segmentation or attributes to call out insights. I’ve added a column that includes the flop texture for each board:

Adding that one bit of segmentation can help demonstrate some insights. Notice there are more Rainbow boards towards the chart's top and Two-Tone boards towards the bottom. We can start seeing some interesting patterns in the data. Translating findings terms that we can understand and apply at the table, like board textures, is how we can start connecting the dots between theory and execution.

There are so many different routes you can go with data analysis when you start layering in segmentation using attributes. Data can be powerful when grouped effectively. The magic point of study and strategy development is when you can identify those groupings that will be easily identifiable and memorable when you’re in-game at the table.

But once you get that magic point, the analysis can go in many directions from there. Think of all the possible segments you can explore:

• Paired, unpaired, and trips boards

• Rainbow, two-tone, and monotone boards

• A-high, K-high, Q-high, etc. boards

• Specific descriptors of boards (e.g. Ace-wheel-wheel, KQ-low, middling connecting boards, etc.)

There are many different ways to categorize things, and all can yield interesting insights. The point is to identify those segments that will yield interesting insights AND are accessible in your memory to translate into heuristics you can implement at the table. It’s a delicate balancing act and will require a lot of trial and error, but this is how anyone can start digging into the data to analyze poker and develop in-game strategies.

Update to my Spreadsheet

I’ve been thinking about the comparison note above for a few weeks. Using Floptimal regularly will do that to you, repeating that methodology repeatedly. But through that repetition, I developed an idea to incorporate additional comparison views into my flop analysis workbook, which I think is a decent improvement.

There’s a methodology of comparative analysis that Floptimal facilitates well, and that is an outstanding way to learn. I don’t know the actual term for it, so I’m just going to call it a pivot.

The image to the right has a few awesome examples of it. Suppose I’m starting by studying my 50 BB BTN open ranges. I can enter the parameter options on the left and study the grid in the center of the image.

With many analysis tools and other resources, the study ends there. At this point, after internalizing the range, you might move on to look at CO open ranges or BB defense ranges, or an entirely different spot.

But Floptimal lets you pivot your focus into a different dimension. We can now click on a hand — in the example, I selected JTs. By doing that, we can look at the second graph on the top right. Now, our focus is not on the 50 BB button open but instead on JTs and how that hand plays at every position and stack size.

The pivot lets you shift your perspective momentarily and look at the data from another perspective. These shifts are incredibly valuable, especially when studying a game as complex as poker.

As I was using Floptimal, I had a bit of an epiphany about a gap in my own workbook and decided to add a couple of views to create a solution. All the data in my workbook starts at the formation level and lets the user drill down. In other words, I required the end user to choose a formation first (e.g. UTG vs BTN, single raised pot). All of the underlying metrics supported analysis within that top-level context.

While you could analyze the data within all formations in myriad ways, you previously could not analyze data across formations very easily — that exercise would have been quite tedious. As a result, I decided to build two additional tabs to solve the issue.

Both views are identical in interface to the existing Formation analysis tab. However, they provide additional levels of segmentation, opening up the possibility for much deeper analysis across formations. The first view presents the data for a chosen board heuristic (e.g. monotone, A-wheel-wheel, paired rainbow, etc.). The second view presents the data for a chosen individual board. Both let you pivot from an individual heuristic or board to look at the equilibrium frequencies across all formations.

I’ve created a brief video that demonstrates the functionality below. I’ve since retired the workbook, but if you are interested in discussing the development of a custom one, reach out to me here.

If you have any comments or thoughts, please feel free to leave any comments below. You can also contact me at [email protected] or on Twitter or YouTube through the links in the footer below.

-Lukich