Not really sure why you say this. What is easier format for a bot (other than husngs)? here you only have 3 positions and very shallow stack sizes, and very limited 3 way action ?That is basically botter paradise

It`s not that we don`t have GTO solution(s) on the game theoretical side of things. It is the fact that we are talking about pure poker game where we have imperfect information and agents cant co-operate. Actually it is pretty interesting how people beat the game in practical sense, and imo it has to do with that all practical solutions used must have came existence with very similar algorithms. Let me clarify, so if only 2 players (3-handed) play under the same strategy produced by the same algorithm it must be the case then that those two players know each others ranges. Hence, the ev loser on the table is then the 3rd uninformed player and in game theoretical sense the two other players are then co-operating against the 3rd player. This is mostly in line with intuition also, if you play enough i`m sure you are familiar with the feeling you get when you sit down on the table with 1reg and 1fish. You know that you are pretty much sharing the cake with the other reg you sat down with.

And which format is then better than this? Well just take 1 player out of the game and remove random prize pools. Pretty sure that kind of game exist

There are players who plays really close to GTO after huge amount of hands (on winamax, on party, on other sites also). They are also playing lot of tables, grinding lot of hours per day.
Basically, they are unbeatable by humans..

Definitely true... its easier for humans too. And why they play spins against 2 players and random prize pools?
Bots comes there, where money are.
And yeah.. heads up is full of bots too. Sadly.

Sure, kind of depends on the 2-handed games. If it is the case that they are indeed full of bots then I def see some botters taking their shot on lottery games. Kind of an red alert sign for any site if there is botter(s) on these games. Tells that the environment is reaaaally bad

Yep and you play with lot less fish. Any players/bots winrate hugely depends on fish.

If you write a bot and it crushes best players in a given stake you for sure could just write a bot that destroys fish more easily.

The lobby is blind (at least for most players) in spin formats so you dont really have to fear from cartels trying to force you to play elsewhere.

Also while your chipev depends on how you play in ANY prize pool, someone can write a bot which notifies the owner when a huge prize pool was hit and a human can put the bot in semi-bot (advisor) mode, making missclick / bad screenreading errors less taxing in real dollars.


Furthermore the lottery games have bigger player pool so you can go for more time before ppl start reporting you.

Well, you kind of make it sound like every site is like pstars where you actually have big pool of players and cartels on HU. This is def not the case. And from individual botters perspective you are gonna have to keep alot more dollars on your account in order to keep the bot rolling. Thus risks increase alot compared to a more stable format.

i mean spin bots have been openly advertised in skype groups, it's not really about whether or not they exist

basically just like playing vs a solver with a ton of abstractions, 3 way pseudo preflop gto solutions exist

heads up obviously has the same problem but on stars bots need to play divisions without getting reported and banned

Ok, so for a couple of days I've done some investigation of my own and seized all play on partypoker. I recently desided to have a go for it due to the rakeback offer being so lucrative. I have now changed my initial view of things and believe indeed that bots are taking over at party sng heros. I did had some suspicions all the way though but I've now confirmed the worst case scenario. Partypoker is indeed infested with bots.

And what you had to say about 3-way solutions you dont know what you are talking about. Go through the theory and you will eventually converge to the same conclusions that I did. I have never said there is no GTO solution for 3-way. What I have insisted is that there is indeed multiple (x amount) of solutions and since you are playing poker and cannot co-operate with players you will very rarely be playing under same solution with others. BUT, the fact that most of the solutions are indeed calculated with same algorithms makes it in practice possible that 2players are co-operating against the uninformed player. Again, please have a read on academic papers on game theory / microecon and you will understand this.

im not saying that the 3 way preflop solutions are gto but will have very little exploitability in any situation that goes heads up postflop

'gto' bots are still a massive problem ánd impossible for humans to beat even if they arent playing a true gto strategy

this isn't academia, the problem doesn't come once the game is fully solved. the problem comes when ai because superhuman.

lol,
Of course it is academics. You seriously think that the current methods of solving became existence and were developed by poker players? Srsly, it is ongoing debate on the academics world which means that there is no superhuman on practical (nl poker) game play on games where there are more than two players.

And how come you solve postflop without first solving pre flop? You can't just assume some random range without actually calculating all of the streets. Thus your whole model where you just assume some range on postflop is biased in gto sense.

This take is just extremely ignorant. Two people playing pseudo-gto based on the same model won't be a collusive consideration vs. a 3rd player playing with a marginally different model. A collusive consideration would be if those two players were actively sharing information in-hand (eg. their specific holecards). Put another way, if there were 2 of me (and so I knew my other self's exact strategy) and a bot at the table but I couldn't share my exact holecards with my other self, both copies of me would lose to the bot despite attempting to play at a potentially different equilibrium (and literally perfectly knowing the strategy of my copy). Try it yourself -- run two different pre equilibria against each other across the game tree and tell us how much of an ev difference there is.

You could say the same thing about husngs -- if I'm playing skier gto pre and a bot is playing instagto pre, the postflop models won't be perfectly applicable. But that doesn't mean I'll be able to beat the bot, and even if I were replaced (by another bot using skier ranges) neither bot would be able to beat the other past rake despite imperfect strategy pairs.

There are bots that have been winning at midstakes+ deepstack cash for many years, and you're somehow trying to say that in 3h super shallow play (muuuuuch easier to abstract for a variety of reasons) bot play isn't at the level humans will be unable to win against? This seems delusional.

You're correct that there could be multiple equilibria due to the multi-handed nature of the game, but I just don't see how that's relevant to 3h bots being unexploitable at a human level and certainly not exploitable for chips that will bite substantially into rake. If you want to say optimal play can win a few chips per 100 vs a pseudo-gto bot...maybe. If you want to say a human can win a few chips per 100 vs a pseudo-gto bot....I'm very skeptical about that, but okay. Even in that case, 3h bot play is a massive problem for spins, not to mention the bots come with unbeatable hu play attached these days.

TLDR even if what you're saying is true, it's a terrible argument and doesn't get around the problems with bots in spins, even just restricting the discussion to 3h play.

Yes. I agree with you on many points here, and I tried to make the problem more practical to explain it in here. I wont drill it any further. Just going to point out that the example where you have 2 agents playing same equilibrium and one that does not is untestable due to the solver only knowing one algorithm. It just so seems that many players are treating these multihanded solutions like they are some absolute thruth which is just absolute bs. We can all have our own opinions on this subject. It is still an academic issue and ongoing debate. If you can solve multihanded unperfect information games where agents cant co-operate you should contribute to academics since it has variety of applications on other fields also.

GTOplayer

In a very theoritical "classroom" sense you are maybe right.

Practically you are in the wrong.

Are they quite different? I've never tried InstaGTO.

In most spots they're not super different though there are some minor differences. In a handful of spots they're quite different because the different sizing restrictions more heavily alter the strategy.

Hi guys,

So I'm looking for a spin and go hud (5$ buy-in). Anyone has recommendations?

Could anyone please share a recent cEV figure for Stars.com $0.25s over 1K+ games?

If not, I might soon provide my own EV graph as a benchmark if I'm unlucky enough to stay there for too long I've just restarted grinding and merely have a cEV of 110 over lol 250 games, but I've been on life-induced tilt for the majority of the sample. I may extend it to 500-600 games this weekend.

tbh i doubt anybody even put over a 1k sample in at 25cs, the rake is insanely bad you might aswell just start with a smaller BR at 7s, also 1k sample is too small to take anything from anyway, probably like 3k+ you can get an idea and even then its still small

8% isn't insanely bad, and I'm not gonna play 7s with a 1.5 BI BR anyway , especially not with the setup that I yet have on the new laptop after I failed to back up - most $1 regs and their dogs know the 3-max p/f Nash charts, and I'm yet to memorise them better, as the starting point from where adjustments can be made, and yet to rewatch Adam's appearances on Paul's stream, etc.

Because of the Russian law, money can no longer be uploaded to Skrill directly for gambling purposes (but can be uploaded through Bitcoin), and I'd now have to learn another way of depositing, and besides, true crushers never reload, even after life expenses, do they?

You see, I need a confidence boost, and nothing does it better than crushing reckless donks (well, not quite crushing yet, but the work is in progress).

You can probably expect something at least 100 chips / game in the 25c, if you're a winner in higher ones. Also agree that the rake of 8% is definitely beatable due to the quality of our opponents.

ofcourse its beatable that was never a doubt, the question is whats better to be playing? you can probably get 90+ cev at $7s and the players there are not much different to 25c players, it would take a loooooong time to move up from 25c to $7s because of rake, youd be better of scraping together some money and starting at 7s

nah man.

What do you think is reasonable for $7s? I always thought 60-80Cev was decent.

Anyone who makes 60 chips at $7s is at least as strong as most of those who earn 100 chips at $0.25s, just because regs tend to use 50+ BI BRM in order not to have to move down all the time, so moving up 3 limits is usually correlated with ample experience and hence skill improvement.

Besides, the population tendencies of $0.25s and $7s are different, and someone with a strong fish game but a weak reg game would do fine at $0.25s but would have to watch zzzTilt videos or something to adapt to $7s regs. (Skier's DB is probably an overkill for $7s.)

I'm seeing the voyage to $0.25s as an opportunity to hone the fish game

Wut?

Variance reduction pools don't serve $0.25 players but obviously serve $3+ regs. From my very limited knowledge of staking offers, some stables accept applicants who're proven winners at $3+ but don't accept $0.25 regs unless the latter agree for a contract at least a year long, so $7 games must be more abundant in staked regs than $0.25s that are abundant in Eastern Europeans but mostly very disinhibited recreational ones.

OK, I think I've just been levelled, gg.

And to clarify, as a Leatherass disciple (but a ton lazier than him), I always focus on the edges that I can have right now in my current development state, not longterm prospects, as I'm still rather seeing myself as a Spin Max reg in the long run, just I want to avoid $7 and possibly $15 Spin Max because their payouts are way too often winner-take-all and so leaving little edge for me, and am tempted to substitute them for €5-10 HexaPro when the time is right, so I still need some HU skill.