Skip to forum
Exclusive AMA with ...
 
Notifications
Clear all

Exclusive AMA with Adam 'coffeeyay' Sobolewski

25 Posts
19 Users
16 Reactions
8,158 Views
Head Admin
BarryCarter
Joined: 13.01.2011

As part of our partnership with Spin & Go staking site SpinLegends we have an exclusive opportunity for our readers to ask our head coach anything.

As you can see from our interview last week, Adam coffeeyay Sobolewski is an expert on math, HUDs, Spin & Go, HUSNGs and much more.

On February 15 at 21:00 CET he will be live on our Twitch channel with host Collin Moshman. Adam will be playing some live Spin & Go's and talking strategy with Collin, as well as answering your questions.

You can start asking them right now. Adam will be choosing the questions so there is no guarantee yours will be picked, but the sooner you ask the more chance you have of getting it answered.

Suggested topics could include:

*Spin & Go strategy
*Mental game
*HUSNGs
*Using maths in poker
*How to optimise your HUD
*Why should you get staked at SpinLegends?
*GTO
*What his favourite coffee is
*Anything else that you want to ask

You have until midnight on February 14 to get your questions in, so hurry up as this is a rare opportunity.


Reply
Quote
24 replies
Head Admin
BarryCarter
Joined: 13.01.2011

Ok my question:

What's your favourite type of coffee and how much do you drink (enough to make it your screenname)?


Reply
Quote
Head Admin
BarryCarter
Joined: 13.01.2011

Also, do you agree with this rather crude description of Spin & Go - it's basically a HUSNG where you can also get lucky and play for a small fortune.


Reply
Quote
zeris
Joined: 18.04.2015

Do you think there is a chance to learn spin strategy by yourself? If yes, then what sources would you recommend?


Reply
Quote
Youthanasiaa
Joined: 23.07.2010

1.
Let's say we have a player that has 40cEV/tournament in HU Hypers. Is he going to have the same cEV in Spins when playing Heads-up? (assuming playing vs the same quality opponents) My bet would be that in hypers we play more hands with a deeper stacks so we probably should have higher cEV/game there. Is that true, and if yes how big of a difference?

2.
What is your take on multitalbling?
I don't know if that's a good indicator, but at what level on spins and on what level of cEV/game should we start adding more tables?
One tabling you can be more focused on every detail and decision --> quicker progress --> higher cEV --> less swings
When to give up on that a little bit?

3.
Maybe you could talk a little bit about sizings, especially flop sizing. Where is the starting point for value, bluff and with draws. How much to increase having a drawy flop?
The same with check raising. How much with a value on safe flop? What to do with middle top pair? What about draws? Is shoving over cbet (25bb deep) generally a worse (more rare) play than smaller raises?
I know this is a wide topic to cover but a general idea would be greatly appreciated!

4.
When we play versus an opponent that we know is trying to think during play and is considered as a good player should we look for cbet frequencies that makes us harder to exploit and we lean towards more optimal strategy? Playing this type of opponent should we have one cbet sizing or still, vary this depending on the type and strength of a hand we hold?

Thank you.


Reply
Quote
tonypmm
Joined: 11.01.2009

Thanks for agreeing to hold a free Twitch coaching, Adam - I did enjoy your previous guest appearances on other streams (PyhaKaru's and Pbogz1114's) in 2016, which I've just rewatched and learned a ton about the advantages of limping 5-7 bb deep vs weak players, and I'm definitely looking forward to the upcoming stream with you and Collin :f_love:

As you've already resolved my previous doubts about limping, let me ask you technical questions about the calculation of exploitative strategies.

In their preflop solutions, all the well-known 'GTO' tools are not doing calculations for all the possible flops but using the same small standard weighted sets of flops, called 'the kuba subsets', designed to be representative, in a sense, of all the possible flops, e.g.

25 flops
49 flops
74 flops
95 flops
184 flops

(here's a zip file consisting of these and other kuba subsets). These subsets are the closest to the set of all flops in a special metric that accounts for EVs obtained by solving for all flops in the 100 bb cash game setting, as well as equities of hands vs predefined cash game ranges. The kuba subsets are believed to perform better than Will Tipton's original 103-flop subset.

There's a trade-off between the precision of the betting abstraction and the one of the card abstraction, i.e. between the size of the subset of flops (and thus the speed of calculation and the number of possible bet sizings that can be studied within the same timeframe) and the accuracy of the approximation of the EV (over all the possible flops).

In your opinion, when studying preflop strategies in 2-3-max shallow stack (<=25 bb, <=12 bb or, most interesting to me, 4-7 bb) spots, how many flops should be included into the analysis in order to get the most accurate solution within a given reasonable timeframe (say, 8 hours with 8 GB RAM @ 2.2 GHz, 4 cores, 8 threads)? Is a standard kuba set (tailored to 100 bb cash games) still enough for decent accuracy, or would you advise to write code to create a special set of flops tailored to Spin & Gos?


Reply
Quote
IvicaIliev77
Joined: 31.05.2012

Very much look forward to this one.

Questions:

1) Expected win rates for SB 3way by stakes (particularly interested from 15$+) 2 - 5 tabling. Your expectations from good regs as well as your win rate.
2) Do you use CG type strategy on SB vs BTN (never CC, always 3bnai for example) or do you like on most lower stake games have CC range and 3bnai and 3bs range? Elaborate more on this by stakes please (expect on higher that its better to enter with 3bnai due to BB being able to squeeze more vs our CC...)
3) CREV shows that BVB strategy from SB should be mostly MR tight (30-35%) and folding rest of the hands, considering that BB will defend 60ish% vs MR with both 3bnai, 3bs and CC and will realize equity more due to IP advantage. Whats your take on SB strategy BVB 18bb+, 14-18bb, 10-14bb, 10bb or less? Particularly interested on strategy suggestions for 15$+ limits. The worse the opponent is , not defending enough (which is super rare in 2017!) we should open more hands to win pre flop but that is not realistic expectation so I am looking for your strategy vs a) Read less (on 15s+) b) Aggro reg (who 3bets and isoAI and naI pre flop) c) balanced/winning reg d) recreational
My take : Saw in CREV study that it suggests 14bb lower we play our range with shove vs rc or l/t lf rf because IP player will realize equity more and to maximize our EV we have to play jam fold strategy pre flop. Very intrested for over 14bb strategy, in particular 18bb+ cause that's where my win rate right now BVB OOP is negative due to post flop OOP play and not balanced pre flop ranges (r/c, r/f, jam, l/t, l/f, l/c iso, l/c ai)

Have tons of other questions like HU Iso nai for shallow stacks ratio value to bluff and how wide do you use for value (expect bluff iso is 52s and such hands that play post flop easy when hit and overall have no sdv) or defend vs 3bnai for 13bb+ on 15s+ and post flop play (bluff catching on different textures that don't hit us well like 552 664 J42 Q62 A94 KQ2 mono etc) - but I will settle for previous question (BVB strategy from SB 18bb+ in particular and 14-18bb).

Appreciate it and wish you all the best in the Twitch stream along with Spin Legends staking and coaching program!


Reply
Quote
Beatlemanic
Joined: 14.10.2010

Hi, I'm interested in staking program at some spins stable. Could you briefly describe SpinLegends - what are their competetive advantages against competitors (another stables) in your opinion, i.e. why should a poker player choose this stable?


Reply
Quote
Super Moderator
Zeezout
Joined: 16.11.2007

I really wonder: What would be the best approach to build up knowledge about Spinners from the beginning on (assuming poker as a game itself has no secrets).

How can a semi recreational player (like most of us are) build up a solid Spin & Go game. Could you share a solid basic strategy to build on, without suffering devastating losses?


Reply
Quote
Dagalaz
Joined: 08.11.2011

What and how do you calculate when you make strategies for the spins?


Reply
Quote
vikash62004
Joined: 11.05.2010

i want to know everything you want to tell me about your:
*Spin & Go strategy
*Mental game
*HUSNGs
*Using maths in poker

and as i to play in spin n go regularly in 1$,3$,7,15 and upto 30$(few) but as i noticed as i got beaten up by small hands or kind of random hands which makes no sense and they just shove with it and i try to take them out by my hands like AA,AK,KK,AQ,KQ,AJ,KJ,QQ,JJ,QJ, AND OR SUITED OR CONNECTERS but got beaten sbout 65 to 80% times and got frustrated but i controlled my mind but after finishing the games i feel tired and more frustrated so give me advice to make my situation better and better. and one more thing i won 10k in 1$ spin n go.

and i have a question regarding the hand which i missed on the flop as opp. to called me preflop but i missed the card on flop and i c-bet him and he called this time too so tell me what is the best option in this situation in all the senario's what if he is tight or loose player , as i loose most in these kind of situations.

and what strategy should i apply as i use push or fold in heads up in spin n go as i lost majority of hands to opponent whenever i tried to let him limp in when i had something call-able thats makes it compleate waste but since i applied this strategy become profitable in long run but i am not satisfied, please help me out.


Reply
Quote

+1


Reply
Quote
Pictor89
Joined: 10.06.2011

I would have only a question: with what worthy to start? (Mental game,maths, HUD optimize or other?) if somebody starts this game now.


Reply
Quote
olympiolf
Joined: 28.12.2016

Q: Whats the main advantages of spin and go if compared to cashgame? Today im a nl2 player with 8/100 profit on 100k hand sample and im about to moving on to nl5. If i was a spin n go player, playing equivalent on spin and go, where i would be more profitable? Thank you.


Reply
Quote
kucu2014
Joined: 24.08.2014

GL ALL:f_drink:


Reply
Quote
bartix
Joined: 02.08.2014

how much ROI I could make easly on 15s ?


Reply
Quote
KTU
KTU
Joined: 24.01.2007

Can you proof this in 2 hours without looking up the proof? Do you use the axiom of choice personally? Do you like our definition of cardinality or does it still feel somehow wrong to you.


Reply
Quote
Troja2013
Joined: 02.02.2017

Bankroll and ROI for 1$ - 3 $ and 7 $


Reply
Quote
JooZeeF
Joined: 14.08.2010

Do you like Placki ? :f_cool:


Reply
Quote
tonypmm
Joined: 11.01.2009

Originally posted by KTU

Can you proof this in 2 hours without looking up the proof?

Lol. I'm sure that pretty any MSc or better in maths, Adam included, knows the proof more or less by heart and can reproduce it within 2 hours, though the length of the proof depends on which definitions you use, which lemmas you consider known and which ones you ask to prove.

Challenge accepted :challenge: (see my proof plan under the spoiler). So Adam doesn't need to spend the precious stream time on outlining the proof.

Spoiler

Basically, the proof boils down to a few steps that are explained in any textbook of complex and real analysis.

1. The series sum_{n>=0} z^n/n! converges for any complex z (proving that is an easy game with the ratio test for series convergence aka D'Alembert's criterion). By definition, exp(z) is the sum of this series. (I haven't heard of any other popular definition of exp(i*pi), to begin with.)

2. The derivative of sin(x) is cos(x), and the one of cos(x) is -sin(x) (we only need them for real x). The proof of this is based on the formula sin(x+y) = sin(x)cos(y) + cos(x)sin(y) (note that you only have to prove it for x and y whose absolute values are less than \pi / 4 because you only need to evaluate the derivatives in a small neighbourhood of pi), its (and Pythagoras' theorem's*) direct corollary 1 - cos(y) = 1 - sin((pi/2-y/2)-y/2) = 1 - cos^2(y/2) + sin^2(y/2) = 2 * sin^2(y/2), and the fact that sin(y)/y -> 1 as y->0 because, for y strictly between -pi/2 and pi/2, as a corollary of the formal definition of the length of a curve (a circular arc in our case) as the supremum of the lengths of inscribed polygonal chains,|sin(y)| <= |y| and |tan(y)| >= |y| implying |sin(y)| >= |y| * |cos(y)| and sin(x+y)-sin(x) = cos(x)*sin(y) + sin(x) * 2 * sin^2(y/2) = cos(x)*sin(y) + 2 * sin(x) * O(y/2)^2 = cos(x)*y*(1+o(1)) + O(y^2) (y->0).

Somehow I found it difficult to recall the proof of the formula for sin(x+y) (assuming that you meant the geometric definitions of sin and cos). I ended up doing a proof involving areas of triangles; I know it's difficult to prove the correctness of the definition of the area (the Jordan measure) and its basic properties, but that's the easiest way that I've been able to come up with. In the below drawing:

Assume |OG| = 1.

|OF| = cos(x), |CF| = sin(x) => tan(x) := |DG| = sin(x)/cos(x)
[by the similarity of the triangles OCF and ODG]
tan(y) := |BG| = sin(y)/cos(y)

S(OGD) = |OG|*|DG|/2 = tan(x)/2
S(OGB) = tan(y)/2
S(OBD) = S(OGB) + S(OGD) = (tan(x)+tan(y))/2 = (sin(x)cos(y) + cos(x)sin(y)) / (2*cos(x)*cos(y))

Otoh

S(OAC) = 1 * sin(x+y) / 2 [the triangle has side 1 and height sin(x+y)]
|OD|/|OC| =|OG|/|OF| = 1/cos(x) [by the same triangle similarity]
S(OAD) = S(OAC) * |OD| / |OC| = sin(x+y) / (2*cos(x))
S(OBD) = S(OAD) * |OB| / |OA| = sin(x+y) / (2*cos(x)*cos(y))

Ergo 2 * S(OBD) * cos(x) * cos(y) = sin(x)cos(y) + cos(x)sin(y) = sin(x+y), QED.

* Just in case, Pythagoras' theorem is obvious once the theory of areas of triangles is developed:

c^2 = (a+b)^2 - 4 * (a*b/2) = a^2 + b^2, QED. Ergo cos^2(y/2) + sin^2(y/2) = 1 for all y.

3. Knowing the derivatives of sin and cos, we can now expand sin and cos into Taylor series with the Cauchy remainders at pi (which is just an exercise in applying Lagrange's mean value theorem), which are convergent by the same D'Alembert criterion. Comparing the series for exp(i*pi) and cos(pi) + i*sin(pi) finishes the proof of Euler's formula.


Reply
Quote