2026 From Online Cash Games to Live

yes UTG PFR vs maniac CO who loves to jam 99%

You're right! I need to add the functionality to my equity calculator that I get as output the hands that are above a treshold. Also need to add a formula for rake.

Without rake, if we raise 99% from UTG we are slightly +EV as we have 50:50 against CO and we share 1.5 bb from the blinds. This leads to a winrate of 75bb/100hh in that spot. Under the asumption BU doesn't want to cold call and blinds not to defend.

As a small poc for my last task.

Maybe an analysis like that can help to judge if a session was good or not. We didn't win this time as much because the table was tighter. We had a sun run because the table was loose.

No clue where I end but I keep just tinkering.

Look into how to write proper formulas. Now that I have a workflow I have no idea what I should write first here as nothing actually needs a written formula.

So I have this: profit is the sum of each possible hand's omega multiplied by cards frequency to play it...

I don't want to study poker. Instead I only want to write the best possible tracker/calculator without using a coding agent. The goal is to work on it for 3 hours a week.

22 Apr 26 Is AI making us become fish?
13 Apr 26 All the major LLMs are losing poker players

I won't make a joke about that GTOWozard said their AI (they are selling) is better than a language model.
But on March 16, 2026 Harvard Business Review published an article about the bias those LLM have. They used seven trade-off weights. For example long-term vs short-term. Dencentralization vs centralization or competition vs collaboration. In total seven pairs. They found that all moderls (ChatGPT, Claude, DeepSeek, GPT-5, Gemini, Grok, Mistral) have a bias towards the same solution. Even when the context changes. They coined gthe word trendslop. So if reddit thinks you should raise and bet all the the time then the LLM will trend toward raise all the time.

It's important to understand what an AI does. Maybe if you gave the agent full access to your system and allowing it to access odds oracle it would come up with better solutions.

Summary Apr
1726 Hands
$+154.52
$2347.51

Yes so far their ads are just two letters: "A I" leading to a prompt where I can buy more prompts... :/

Right now, AI companies start to price out more and more people with the price hikes. I have no idea to which result this will lead.

Pokerstrategy glossary: Risk of Ruin

Works for cash games (for sit and goes and tournaments you should use the finishing places to calculate the risk).

To remind ourself, negative exponent means the 1 divided by the whole expression. Exponent gets smaller when we increasre varinace. Exponent gets bigger when we play better or increase the bankroll.

Also looked at how to write tests in C++. I was a busy last June and did good work. Won't need to learn all from scratch. Buggest headache is still how to design the database so I won't need to parse everything each time I start the tracker.

And then I need to learn how to write game theory-thingy to have as nice graphs and tables as the formulas are.

Thanks! So these software hypers and new software companies have started to outperform my gamedev company... such great news

Yes, to Ralph Wiggum a game for people without development skills will be a thing of the past again. Or they need deep pockets. There are the first places where AI is more expensive than an actual human being. The cost of AI got so much more expensive that Uber spent their planned AI budget already in April.

Setting up all - or at least planning the setup takes more time than expected. Not due to the difficulty more to the available time. And my desire to make it clean and robust to avoid all traps I fall into in the past.

On github I have an idle "game" that I started to write when I wanted to learn javascript. After a week I noticed I do not care about JS only about the idle game. But the idle game was not enough of motivation to continue javascript (or truescript). It would be perfect to share it free to the world.

Couple of months later I am back again to the idea of an idle game. Obviously it's poker themed. But no clue so far if it should only be fun or also teach some poker...

Maybe I will use this library to write my tracker / calculator. So what I might learn here could be useful later.

Massgrinders, did you ever wonder how to plot your hands into a graph?

If you play 1000 hands a day it's 30k a month. But the but a 4k screen only has 4000 pixels.

To draw a line you need to have the points apart. Maybe five pixels.

Which brings you to the task of reducing the population of 30000 points down to a set of 800 points.

Now the fun begins! What's most important to a poker player? Maybe the shape? Or is it the outliers and the biggest win/loss to show off how big your variance is? Or are you reasonably and pick the graph that shows the points that represent most other points.

There are algorithms that have the complexity of n^3. For each point you measure the distance between the line start to end. The biggest point becomes a new sub point that recursively is end and start point for the two segments.

Faster is to calculate buckets. There you can pick min and max values.

And the most fun observation is that you should use different algorithms depending on how much you zoom in. If someone outthere should be that crazy and implement such a plotter instead of using a well-tested library would need to implement multiple techniques. I hope no one would ever come up with the idea to do it yourself for a simple poker tracker!

Keep your own tracker in MS Excel 2010 sheet and go for tracking months, you can update monthly profit daily but when doing data analysis focus on months/years, not days.

It is almost as good as Pokertracker 4, but doesn't show smaller stuff, such as "reg had Ax on the River". The thing with Pokertracker 4 is, it has many dependency conflicts, such as: "partypoker or ipoker could ban you by their t&c if it is open at the same time with poker client".

I still have trouble to make the window resizable without stretching the content shown but I already know my game needs mini games. A lot of mini games! To solve the issue of resize I simply can make the application full screen. Which then of course - at least in my opinion - defeats the purpose of an idle game. I wan't a little bit of story but not as much that a full screen game would need. Or I would need to pick another genere. But I finally managed to redraw the window when resize. It's a short flicker but it works. But I already learned more so I might achieve a smooth resize. An ideae would be also like poker sites do with their tables. They always adjust to the correct aspect ratio if you don't change the size of the window by the handle.

But I also learned how I want to design my poker tracker / calculatot / studythingy or whatever it will become in my fullscreen excursion. I can easily arrange information around the screen. There obviously is only one goal that I have to strive for. It has to be fully customaziable. With it's own script language. And all has to be able to be used only by keyboard.

I did a small simulation to see if I am already capable to create outputs.

I can now draw a chart within a pixel area and save it. Adding X/Y-axes should be easy...

2 December it's EPT in Prague.

Last 24 June I started gradually walk more and more till it was excessive. Which I will pick up this Sunday again. This way I can be ahead one month to my past mr. I have no clue where I am but much better than last year. The urge isn't to change my life but to compete against me with my collected data. A 2nd goal will be to achieve the activity badge from my phone 25 days each month.

Also this month the season ticket for public transportation is on sale. 3 June to 2 September I will be on track - pun intended. 3 September to 2 December I once try again to book the other season ticket. Which then perfectly would align with Prague...

I made some predefined commands for the AI. It's funny how one little improvement using those tools gives you almost the feeling you understood now all. Everyone else has skill issues despite the fact AI is rubbish and far away from what they are selling you.

I learned that I use neovim wrong. It's buffer -> window -> tab.

I coded a button! It's not much further in 10 days. But I didn't manage the 3 hours a week.

102 days journey starting Sunday.
The 91 days encore hopefully straight afterwards.

Day zero. I left my home after seven and was home before two o'clock to run errands. When I was home I managed the make all the steps for the upcoming activity badges. Was even 20% above the target in the evening. Eventhough I managed it I'm in worse shape than I thought I was. Started to need little breaks when I was going back. But on the other hand, I was packed with additional weight in my backpack. To the way through the nice park and came across those two little fellows (and a lot more of them).

Learned that in neovim ctrl-c is also the esc key. Those small things I need to learn but all make more confortable. Buffers are :b :bn next :bp previous. Use a number to jump to that index directly. Use ctrl+6 to toggle between two. Then there is also :bl(ast) and :bf(irst). :buffers to overwiev all. To open a new buffer use :enew. That's enough for me to practice...

In my head I try to solve how to calculate equity for split pots elegantly. But I guess It's looping arrays for pot and each side pot...

Day one. Don't know if it was smart or stupid to already add the next walking session. The start was easy but then I hit a wall. Really thought at the beginning till 11 o'clock I would have achieved that goal. But then I ended the day at 99%. Hopefully it was a smart decision as this practice will continue for a week till it counts.

Those little fellows caught the eye of everyone around.

Was beautiful to watch how they walked from the pond to the grass. Not minding anyone. If someone was in their way they sat down and waited till that person stepped aside. It's their path.

Played a very short session of PLO. Received the best of the best cards and got dealt the most beautiful boards.

Need to learn how fonts work. Just copying them side by side is wider than a individual string. But at least I can use different colors to write the whole cards.

In neovim I learned to use :mks(session) so I can store which files were open. Very useful on a project by project basis. Also allows to switch easy between projects.

Originally posted by Rhodriguez
Day one. Don't know if it was smart or stupid to already add the next walking session. The start was easy

Yup same here, i have this running schedule of 2km run on Mon,Wed,Friday mornings but sometimes kids make me so tired that i wish to quit. But always feel super good to run again through that forest, especially now when the grass has appeared and many paths are more narrow lol.

Absolutely! Was for a short period in the forest as well. The scent alone in the air makes you feel more healthy! I managed to achieve my activity on day two. FInished at 11:15. There are three ways to improve. Less breaks, shorter brakes, faster pace.

PLO continued going well. Won the hand where I had 17% equity. The big mistake would be to think the decision there was wrong. 17%, half of the time opponent fold. I win 50% a pot. Win 17% 2x pots and lose 84% a pot. It's break even. As it was all-in the bet wasn't full bet. Let's say 17% 1.5x pots and lose 84% 0.5x pots. Would be 0.335 pots. The decision was perfectly fine.

RIght now I use my project as one single canvas. Everything I want to experiment on gets in the same source code. I need to create for each single idea a new example. Create a tutorial for future me. Because now It gets too big to keep track what I was doing and where I was doing it.

Day three and it feels already like eternity. The legs seem to weigh tons right after waking up in the morning. But at least I don't feel exhausted in any other way.

If there is a flop that you can imagine to connect in any way shape or form to any four cards it's a good idea to bet. The professional expression is somewhere in the area of "the flop misses opponents range". Unfortunately I only had the courage once to bet with absolute air.

i should not use the sit out next big blind option. Because when that check box is ticked on I feel the pulse... Then it is risky - in that situation is danger!

3 days left to finish boring office work and then begin my journey as mail carrier next Sunday till the end of the month.

Summary May
18 tourneys 227 cash game hands
$+51.43
$2398.94

Yes some are born to do that while others sneak away before high school to drink whisky. Good luck!

I was sneaking back in. For four months that work was laying around and now it was done in an hour plus an hour of delivery.

Read a little bit about the value of different chip types.
The first simple relation is:

From that I will made in my tracker

Which looks like a lot of monnies!