Training update: Week 1
******************
Quick report on my first week...I've been collecting all my digital training material up to now and trying to filter out everything that is still relevant since I got a solver. Very useful.
One of the things that is fascinating me lately is the possibility of being able to construct and write out a very basic, primitive strategy - using various 'example' solves - and then committing this central skeleton to memory, like a
gto safety blanket. From Michael Acevedo's Modern Theory of Poker, he recommends categorising flops as follows:
A = Ace
H = High (K,Q,J,T)
M = Mid (9,8,7,6)
L = Low (5,4,3,2)
So, for example, a flop of A72 would translate to AML. Or a flop of KT9 would be HHM.
Clearly we're going to play a board like KT9 differently than we would say JT6, even though they are both HHM. But by grouping them all together and training ourselves to recognise the flop texture quickly we will at least be able to narrow in on our general strategy for these *types* of spots and adjust accordingly.
So, taking these 'whole game' boards as our basis, I've started solving my ranges down for every positional spot. Pio does apparently produce reports like this in the fully paid version but I just have basic which is fine for me. Plus, by processing manually I get to customise my reports into this easier-to-remember (imo) format of ascending flop categories.
Here is an example of how I go about it, the spot is BU 3bets CO, CO checks flop:
*KEY -> starting on the left here we have a list of all the flop categories descending A-high, then the EV of each position, the difference between those EVs, and then lastly the coloured bar is Pios visual approximation of its strategy across its whole range. In this example: red = bet 66%pot, cream = bet 33%pot, and green = check. On the left the solved boards are all two-tone boards (the majority of flops) and on the right are the same nominal textures but rainbow boards. Below these two categories I've included some paired boards and monotone boards, just to get a general idea.
Obviously solving one board for one spot - e.g. HMM for CO v BB - and then declaring
that as
the best way to play every board in that category, would be a disaster. No. We're looking for BASELINE gto. We start off by distilling the game to it's essence so that we can grow in complexity later. We're also looking for patterns.
I can never hope to memorise everything so I'm focusing on balancing that which is most retrievable with that which is most important and common.
So far I've solved about 4 spots using this method and it's been really helpful as a starting block in trying to conceptualise my whole strategy. But now beginning to wonder if it's a little too clumsy and I'm over-simplifying by, say, including only one board for one spot, when that particular texture might occur far more often and therefore require much greater examination to play well.
Not all flop categories are created equal. As Acevedo breaks down in MPT, the most likely flop is HML, which is actually 18.5% of flops. So I could be wasting my time learning boards like AMM when they only occur 2% of the time. But I also think it's easier to memorize spot patterns when you can see their strategy laid out in a linear fashion.
So - I'm stuck on that bit at the moment. I think the system needs modifying...