Welcome, Guest. Please login or register.
June 27, 2022, 04:24:10 AM
Home Help Login Register
News:

Shanky Technologies Poker Bot Support Forum  |  General Category  |  PPL Support (Moderator: sngbot)  |  Topic: cake 0 Members and 1 Guest are viewing this topic.
Pages: 1 2 3 [4] 5 Print
Author Topic: cake  (Read 38019 times)
Botopotamus
Confident Newbie
**

Karma: 5
Posts: 77



« Reply #60 on: October 10, 2011, 07:21:36 PM »

My proposal for a solution on the part 2 thread:

http://bonusbots.com/support/index.php?topic=4332.msg42895#msg42895
Logged
vilvil
Extreme Botter
*****

Karma: 68
Posts: 1357


53 68 61 6E 6B 79 75 73 65 72


« Reply #61 on: October 10, 2011, 07:25:51 PM »

StackSizeInChips is a good idea and I have requested this

 thanks

I have suggested this a week ago:

A new stacksize variable who could identify the real stacksize (not blinds) would be nice.
Logged

int main () { cout << "Hello Shanky!"; return 0; } 38911 BASIC BYTES FREE SYS 49152
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #62 on: October 10, 2011, 07:26:39 PM »

StackSizeInChips is a good idea and I have requested this

 thanks

Cool  wink
Logged

SethUnter
Extreme Botter
*****

Karma: 87
Posts: 1016



« Reply #63 on: October 10, 2011, 07:53:32 PM »

Quote
preflop

when (StackSizeInChips = 8.88 or StackSizeInChips = 88.00 or StackSizeInChips = 888.00) userstacksize8

when userstacksize8 and not (hand = AA or hand = KK) fold force

The stacksize misread happens very frequently, but has an undesired effect much less frequently, so this code would make us fold much too often I think.
Logged

Have you tried turning it off, then on again?
Botopotamus
Confident Newbie
**

Karma: 5
Posts: 77



« Reply #64 on: October 10, 2011, 08:14:17 PM »

The stacksize misread happens very frequently, but has an undesired effect much less frequently, so this code would make us fold much too often I think.

I don't bot Cake very much because of this bug.  I tried it yesterday so I have one log.  Maybe someone with some more logs can give us more info.

In 84 hands, it happened 5 times, all 5 times it was preflop.  It only caused me problems on the hand that I posted.  It seems that it's only a problem when you have a marginal hand and the blind size gets large enough that it makes your stacksize small.  Then you are pushing with marginal hands when you really don't need to.  It can knock you out of tournaments quickly.

A better solution would be to use the user variable at the points in your profile where you are pushing with marginal hands.  That will be a lot more work, but a better overall fix.
Logged
Aenima773
Sellers
Experienced Botter
*****

Karma: 45
Posts: 452


« Reply #65 on: October 10, 2011, 08:21:44 PM »

The stacksize misread happens very frequently, but has an undesired effect much less frequently, so this code would make us fold much too often I think.

Agreed. I was under the impression that the misread caused us to do odd things. Looking at others' codes, it looks like this is only an issue in tourneys where the blinds are really high. Obviously this triggers the bot to push with a marginal hand in this spot. If the code lines are acted on properly, just the stack-size is wrong, then this will do nothing for cash games. It holds little value for tourneys as well because you can't simply fold hands everytime you get below a stack of less than 4 bbs, you would just get blinded off.

I am not quite sure the value of any of this, but I figured I would at least try and code it properly. The new variable may be somewhat valuable to tourney players though.
Logged
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #66 on: October 10, 2011, 08:43:56 PM »

Agreed. I was under the impression that the misread caused us to do odd things. Looking at others' codes, it looks like this is only an issue in tourneys where the blinds are really high. Obviously this triggers the bot to push with a marginal hand in this spot. If the code lines are acted on properly, just the stack-size is wrong, then this will do nothing for cash games. It holds little value for tourneys as well because you can't simply fold hands everytime you get below a stack of less than 4 bbs, you would just get blinded off.

I am not quite sure the value of any of this, but I figured I would at least try and code it properly. The new variable may be somewhat valuable to tourney players though.

My attempt was aimed at cash games
and i think it will work as intended when we get this new variable (StackSizeInChips)

the idea behind my code is to fold all hands except the strong ones everytime the bug appears
Logged

Aenima773
Sellers
Experienced Botter
*****

Karma: 45
Posts: 452


« Reply #67 on: October 10, 2011, 08:58:36 PM »

If you look at any of the logs that were posted, hands were executed correctly even with the stacksize misread. What seems to be happening is a stacksize of 888.00 is short stacked in a tourney with a big blind size of 200. This causes the push code to come into effect with a marginal hand. I don't believe this misread is currently causing many problems in a cash game. It makes no real difference how most profiles play if the stack is say, 8.88 or 9.50  on a NL10 game. This does factor into tourney with large big blinds though.

I think you are attacking this wrong. The issues I have seen that cause the bot to go nuts in cash games come in the pot size. A -4734283492 sized pot really messes things ups. Having 88 bbs is not going to effect a single code line in most cash game profiles.
Logged
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #68 on: October 10, 2011, 09:05:46 PM »

I have been playing cash games at cake and i think it will solve the very bad situations where we call weak draws or with bad kicker etc.. when the bug appears

but ill just leave it be if you guys dont think its working  beer
Logged

Aenima773
Sellers
Experienced Botter
*****

Karma: 45
Posts: 452


« Reply #69 on: October 11, 2011, 12:02:15 AM »

I certainly see the need for the new variable. It would make a lot more sense to fold a hand with that exact stacksize as opposed to any stacksize that is close. It is just not feasible to always fold between 3-4 bbs in a tourney with a big blind size of 200.

I am not sure what the misread is doing, but as some of the posted logs show, the correct lines are still being executed. It just has the wrong stack size being shown. It won't come into play much if you fold all with a big blind size of 25 or less, I just have not seen a posted log where this issue has caused a bad play except in the tourneys where it makes you appear short-stacked.
Logged
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #70 on: October 11, 2011, 12:49:16 AM »

the point of my coding attempt is not to call when the bug appears unless you got a strong hand , so postflop we will fold everytime the bug appears unless we got

(havequads or havefullhouse or havestraight or haveflush or
      haveset or havetrips or havetoptwopair or havetwopair or haveoverpair or havetoppair)

preflop we can just add the specific hands we dont want to fold here

when (userstacksizepreflop8 or userstacksizepreflop88) and not (hand = AA or hand = KK) fold force


and when it comes to toppair and overpair i have put in a few more details

when userstacksizeflop8 and haveoverpair  and not ( hand = AA or hand = KK or hand = QQ or hand = JJ or havetoppair) fold force

when userstacksizeflop8 and haveoverpair  and (hand = AA or hand = KK or hand = QQ or hand = JJ ) and not havetoppair  and opponents > 2 fold force

when userstacksizeflop8 and havetoppair and not (HaveBestKicker or Have2ndBestKicker or haveoverpair) fold force

when userstacksizeflop8 and havetoppair and (HaveBestKicker or Have2ndBestKicker) and
(flushpossible or straightpossible or paironboard) and not (haveflush or havestraight or haveoverpair) fold force

I think my code will work for cash games with or without the new variable but without ill have to code a little more and then it will as you mention not only effect the 8.88 etc... but also when stacksize is 8 or 9 etc..
Logged

Aenima773
Sellers
Experienced Botter
*****

Karma: 45
Posts: 452


« Reply #71 on: October 11, 2011, 02:28:59 PM »

If created, then obviously every body can choose whether or not to use it, so my objections matter very little. The point I am trying to make however, is although the stack size is misread, the correct code line gets executed.

This is only really a problem when a shove code comes into play for a tourney profile as outlined in the second example. Since shove codes in cash games don't factor in when you have 88bbs, there just seems to be little point is all.

Now if a single person can come forward with a log that shows this misread is causing any problem besides the shove code issue, then I would love to see it. It just has not been posted yet. This really needs to be an effort to code against bots shoving when they really have something like 10+ bbs left, instead of the 4 or 5 that this misread can cause the bot to think it has. I think people are mistaking this for the all in problem, which I believe is different. Again, if not, I would love to at least see 1 log in which a code like this for a cash game would do anything productive at all.

Why would anybody want to fold QQ or JJ preflop, or even a flushdraw or straightdraw post-flop, because the bot thinks you have $8.88 instead of $9.50 or $7.50 or $18.50 or whatever the real stack is? If you code this detailed enough to include folding all hands that are not of value when this happens, then what is the point? 95% of those hands should be folded anyway with normal coding unless a bluff or C-bet is being used.

I am not saying this is wasted effort, I just have not seen a log on here yet that this would help in any way. I have seen a couple of logs that it could hurt though by folding a semi-premium hand. Now obviously if when an opponent goes all in our stack changes to 8.88, then we have something. But I don't think that is the case.

Shanky Technologies Holdem Version 6.1.1: Hand: 2s Qs
Players = 8 have Q7-2 suited
If calls = 3 or more and raises = 0, call but fold for 2 raises cold
hands played in session = 66
in timer: players = 8, folds in game= 6, raises = 1, checks = 0, calls = 0, bets = 0, folds = 6
casino = RS 10 Max, game = No Limit Tourny, BB = 200.00, Pot = 3 BBs, Stack = 888.00, TI = 0 BBs
In Small Blind
Matching user defined condition at line 927, Action = fold force
clicking Fold
Dealer: xxx wins 500


Shanky Technologies Holdem Version 6.1.0: Hand: 8h 7h
Players = 8 have 87 suited
If calls = 3 or more and raises = 0, call but fold for 2 raises cold
raise if opponents left <= 3 and no calls and first hand cards is black
hands played in session = 48
in timer: players = 8, folds in game= 6, raises = 0, checks = 0, calls = 2, bets = 0, folds = 6
casino = RS 10 Max, game = No Limit Tourny, BB = 250.00, Pot = 3 BBs, Stack = 888.00, TI = 0 BBs
In Big Blind
Matching user defined condition at line xxx, Action = raisemax force
clicking MAX
clicking Raise
Flop: 6h 9d Ah
Board: 2 suits no possible straight
have flush draw and straight draw (including inside straight draws)
Turn: 6h 9d Ah Th
Board: No Pair possible flush three of one suit, including possible straights needing two cards
have any flush
River: 6h 9d Ah Th 5d
Board: No Pair possible flush three of one suit, including possible straights needing two cards
have any flush
Dealer: xxxxx wins 8,885 with Flush, Ace

Logged
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #72 on: October 11, 2011, 02:59:53 PM »

Why would anybody want to fold QQ or JJ preflop, or even a flushdraw or straightdraw post-flop, because the bot thinks you have $8.88 instead of $9.50 or $7.50 or $18.50 or whatever the real stack is? Grin

the point of the code is to see if we could find a specific thing in the bot log everytime the bug appears and then code a fold , i have then added some details to the code so we dont fold the nuts or good hands and i am sure more can be added but i really need to see first if the principle is working before spending a few hours adding more details

and reason why i made the code for cash games is that there is an issue with cake and cash games its not something i just imagine  violin

Any bot log is welcome and if posted i will have a look  wink
Logged

Aenima773
Sellers
Experienced Botter
*****

Karma: 45
Posts: 452


« Reply #73 on: October 11, 2011, 04:16:03 PM »

Obviously we are on two different wavelengths with this issue. I will hold back my comments until I see any log that confirms that the stack size issue does anything besides make the bot think it has a different amount of chips than it really does.

There are 3 bugs I know of on Cake.
1) The bot cannot read the size of the bet when an opponent goes all in. I am not sure if this makes the bot call or fold, but it is obviously a problem.

2) Sometimes the stack size is misread. I have seen the havoc this plays on tourneys, but am waiting to see a log that says it makes the profile do anything other than what it would normally do if you had 88bbs in chips.

3) Sometimes the pot goes -45945498548. This causes the bot to do crazy things and should be easily fixed with a potsize < 1 type variable. I actually think this is the bug that wreaks the most havoc on Cake cash games, but I have not played it in a while and do not have logs to corroborate this.

Those are all the issues that I know about and any log shedding any light on any of this would be appreciated. Without logs it is simply guesses and speculation and the only issue we have noticed so far is enacting the push code.

Now that I have given all of the information that I have on the topic I will sit back and watch whatever comes of this happen. I will be more than happy to release any type of code, or use something created by another, if it solves a known issue. But looking on any of these threads, this does not solve any problem that I have seen in any bot log for even 1 single hand. But hopefully somebody will come forward with information proving me wrong as a fix like this would benefit me as well. But until then, I remain skeptical.

I am also not sure of the value of any code past the preflop as once you have folded all but AA and KK, who really cares what happens. If something slips past this because you are in the big blind, simply do a when Userwhatevervariableyouwant and not (hand = AA or hand = KK or havequads or havefullhouse or havehandyouwaitedalongtimefor......) fold force.

It takes like 6 seconds to code, but I have not done it because the value has not been shown in logs just yet. The problem with this is I feel as if the entire forum is waiting for some magical code that is going to fix Cake. I know that I am. I just hope somebody can produce a log that shows the misread in stack size does anything to the profile so we can move forward productively instead of feeling around in the dark as we currently are.

So I wait............ popcorn
Logged
zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #74 on: October 11, 2011, 04:31:32 PM »

check the date from when thread was started  popcorn

i still am pretty sure when we get the new stacksize variable my solution will be ok
and any hand can be added to the solution.
Logged

zulu
Extreme Botter
*****

Karma: 80
Posts: 2893



« Reply #75 on: October 11, 2011, 04:48:59 PM »

There are 3 bugs I know of on Cake.
1) The bot cannot read the size of the bet when an opponent goes all in. I am not sure if this makes the bot call or fold, but it is obviously a problem.


this is the main problem i am trying to solve and if i am right its not only when opponent is all in , and as mentioned if we can pinpoint something the botlog in that situation has in common we can use my code

I am not sure if its even possible to show a botlog that display the whole problem but hopefully the bot produce a log where the stacksize is shown as missread as well in the specific situation.

2) Sometimes the stack size is misread. I have seen the havoc this plays on tourneys, but am waiting to see a log that says it makes the profile do anything other than what it would normally do if you had 88bbs in chips.

Hopefully 1 and 2 has the same thing happening in the bot log when the bug appears



« Last Edit: December 01, 2011, 09:38:21 PM by zulu » Logged

Botopotamus
Confident Newbie
**

Karma: 5
Posts: 77



« Reply #76 on: October 11, 2011, 04:55:43 PM »

3) Sometimes the pot goes -45945498548. This causes the bot to do crazy things and should be easily fixed with a potsize < 1 type variable. I actually think this is the bug that wreaks the most havoc on Cake cash games, but I have not played it in a while and do not have logs to corroborate this.

There is already a PotSize variable in PPL.  Code it to fold the hands you don't want to play when PotSize < 1.  This one should be simple to work around.
Logged
Botopotamus
Confident Newbie
**

Karma: 5
Posts: 77



« Reply #77 on: December 01, 2011, 08:43:38 PM »

StackSizeInChips is a good idea and I have requested this

 thanks

Egor, because US players now only have Cake and Minted, is there any way you can put a priority on this so MTT players can play at Cake while we wait for Bodog to be resuporrted?
Logged
whoremonger
Sellers
Serious Botter
*****

Karma: 33
Posts: 577


unexpected havenuts


« Reply #78 on: December 01, 2011, 09:32:58 PM »

yeah, can we get an ETA on that too ? lol beer
Logged

webber04
Extreme Botter
*****

Karma: 74
Posts: 3760


« Reply #79 on: December 13, 2011, 09:40:05 PM »

this would be a great variable to add
and not only for Cake but for all sites

would StackSizeinChips be that difficult to add since u are already getting the stacksize and dividing it by BBs??
Logged
Pages: 1 2 3 [4] 5 Print 
Shanky Technologies Poker Bot Support Forum  |  General Category  |  PPL Support (Moderator: sngbot)  |  Topic: cake
Jump to:  


Login with username, password and session length

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!