Stackoffski vs Stackoffski

Originally posted by calimanpool
Ouch, that's a tough way to go out. Losing a 70/30 like TT vs ATo right at the finish line always stings. Still, another final table is a huge achievement! Consistency is what matters in the long run, and you're clearly hitting those deep runs. Congrats on the score!

Thanks, today I won a smaller one :f_thumbsup::f_thumbsup: :

And almost completely out of the hole I was digging:

Wow, and another 2nd place finish :f_thumbsup::f_thumbsup:

I feel like I'm becoming a black belt

YEARLY RESULTS:

- MTTs:

- Rush Cashgames:
I am probably still not a very good cashgame player, as I ended 187K hands with a EV loss 30 buyins. With rakeback from the cashgames only though I am down $50 EV. Altough I have to say I was pretty unlucky with hitting sets and straights according to the bell-curve, even after 187K hands. Who knows, maybe I would have been up a little bit without the unluck there.

Hi Stackoffski,
which Cash Game Limit do you play?
I hope you have more luck next here!

best regards,
kingdollar

Originally posted by kingdollar
Hi Stackoffski,
which Cash Game Limit do you play?
I hope you have more luck next here!

best regards,
kingdollar

Hi kingdollar, that is 2nl to 25nl rush, depending on my total bankroll. Here is the same graph in bbs:

So before and after that big dive down at the start I was kind of breaking even in EV bbs for a long time (about 150K hands), but recently started to lose a bit again.

That's why I won't give you specific advice about cbetting, I know you asked the question in your blog.

2026

Have a prosperous 2026 every one!

--

I am starting this year with $687 in my pockets (did he lose the rest of it? :f_grin:):

--

After trying the rush cashgame tables last year and not being very succesfull I decided this year I will try the normal tables. I never played those because rush is so convenient. But, after playing a few sessions at the normal tables, I remembered how much metagame is part of that game. Much more than rush where no one knows whether you've been raising 3 hands in a row already etc. I think I will fare well here! You can really throw these people of their game with selected aggression.

However, I must keep in mind that I myself are also thrown off my game more easily at the non-rush tables. If the same guy keeps beating you with shit it can be hard to have a healthy focus. In rush, you just carry on if you lose some pots ...

--

Have 1 on me (of whatever you're having)

Wasting 1000th post

1000 is still short term. I'd say you need at least 750K to really have a very meaningfull sample.

Speaking of longterm, how many hyper mtts do I have to play in order to have a reliable sample to see if I am a winning player? And how many to be able to trust your winrate?

I have a nice ROI% of 128.23%, but only 2K mtts (I don't know what is considered good, but I bet it's nice).

MTTs = 2164
Profit = 611 buyins
ROI% = (611+2164)/2164*100 = 128.23% (hope it is the right formula ... return over investment, right?).

My graph is pretty steadily:

The real $$ graph is less nice because I go and play more expensive mtts after a win (and go back when I am feeling I am loosing too much).

I am pretty confident in my microstakes mtt play (not so much in my cash game skills). I hope to build a larger bankroll to be able to play the $5 (or better) $10 games on the regular. I have about $830 now as I cashed out a lot in the past. Should be enough to play the $2/$3 hypers?

Ah lol I didn't even know you have to subtract the 100, so my roi% is 28.23% haha. I know nothing about mtts. Only studying I did was for cashgames.
Well, besides reading Dan Harringtons mtt books a few decades ago ...

I played a bit with that primedope variance calculator. Interesting.

A high ROI definitely will give you a much bigger chance to win some nice money. I don't think you should sacrifice too much ROI by playing too many tables. I still only play 1 at the time (not the most profitable, but who cares about money? :f_tongue:) and I guess that's where my edge and ROI comes from.

I also programmed an evcalculator in javascript (nodejs). It is still a bit limited, but it works! The results are the same as fxxxzilla's. Only thing is it takes about 50 seconds to calculate for 2 hands vs each other. And that makes it pretty much useless ... I wonder how 'they' program those things.

Originally posted by Stackoffski
I am pretty confident in my microstakes mtt play (not so much in my cash game skills). I hope to build a larger bankroll to be able to play the $5 (or better) $10 games on the regular. I have about $830 now as I cashed out a lot in the past. Should be enough to play the $2/$3 hypers?

Guess my wish came true ... I final tabled, we made a deal (first time for me) when I was 3rd out of 4.
BR += $725. I have $1500 now, which should be good for starting to play the $5 mtts regularly.

Originally posted by Stackoffski

Originally posted by Stackoffski
I am pretty confident in my microstakes mtt play (not so much in my cash game skills). I hope to build a larger bankroll to be able to play the $5 (or better) $10 games on the regular. I have about $830 now as I cashed out a lot in the past. Should be enough to play the $2/$3 hypers?

Guess my wish came true ... I final tabled, we made a deal (first time for me) when I was 3rd out of 4.
BR += $725. I have $1500 now, which should be good for starting to play the $5 mtts regularly.

Wow! Very nice!

Originally posted by Stackoffski
I also programmed an evcalculator in javascript (nodejs). It is still a bit limited, but it works! The results are the same as fxxxzilla's. Only thing is it takes about 50 seconds to calculate for 2 hands vs each other. And that makes it pretty much useless ... I wonder how 'they' program those things.

Down to 3.3 sec for hand vs hand evaluation for winning chances, with a data file of 10Mb. Still too long, but getting there!

Originally posted by Stackoffski
Down to 3.3 sec for hand vs hand evaluation for winning chances, with a data file of 10Mb. Still too long, but getting there!

It's a very nice coding exercise. 10MB sounds very big - for that performance. But of course it depends what you finally intend to achieve.

And when you done with that you can write one for PLO without looping all combinations!

Originally posted by Rhodriguez
It's a very nice coding exercise. 10MB sounds very big - for that performance. But of course it depends what you finally intend to achieve.

And when you done with that you can write one for PLO without looping all combinations!

Well, it determines winning chances for 2 holecards vs 2 holecards by going through all possiblities. That are a lot of possibilities, takes way too much time to simulate without having a lot of it precalculated. I am still calculating too much because it's still too slow.

I had all 5-card combos ranked with a number, but still have to determine the best 5-card hand from the 7-card combos (holecards + board) 'manually'. And that still takes too much time.

You ever did something like this? I wonder what the best method is.

Well, so I ranked all possible 7-card combinations (133,784,560) and 5-card combinations (2,598,960). Saved that ranking together with a more specific (numerical) hand-decription in a data-file.

The data-file is 250mb, but the results finally are in just below 1sec, as it is only lookup work now.

Originally posted by Stackoffski
Well, so I ranked all possible 7-card combinations (133,784,560) and 5-card combinations (2,598,960). Saved that ranking together with a more specific (numerical) hand-decription in a data-file.

The data-file is 250mb, but the results finally are in just below 1sec, as it is only lookup work now.

* well, my computer did

Perfect!

Next steps are reduce sorting, reduce comparisons.

And if you want to get the used memory down, you could split between flush and non flush hands. (As explained in the link below.)

A good read could be: Algorithm of a PHE

I have one more idea of an implementation for myself. After that a full solver!

Hi Rhodriguez!

I did split the flush and non-flush hands! Because I ranked all possible 7 card combos, the only comparison done is which combo has a higher ranking.

This was not my idea, but: I used prime numbers for the ranks of the cards. If you multiply the prime numbers of a hand, you get a number that matches the hand regardless of suit and order of the cards.
The ranking is just saved in a map with those prime products as keys and the hand-rank as values. (product of a prime numbers can not be achieved with other prime numbers, so in that sense it's unique)

Not every combo could be saved like this though because sometimes the suits of the cards do matter.

Thanks for the link! I'll take a look at it.