Open Side Menu Go to the Top
Register
3/6: Wake Up, San Francisco 3/6: Wake Up, San Francisco

11-07-2014 , 03:54 PM
Could someone link the PokerCruncher thread? Is the app available for Android?
3/6: Wake Up, San Francisco Quote
11-07-2014 , 03:56 PM
Quote:
Originally Posted by BigBadBabar
i still don't understand what's going on in this thread. it's not even a spot where sims/equities are relevant.
I agree yeah it's hard to let go of Kings full. I guess jdr0317 just wanted to verify the equities. But the app has to be right in even very edge-case setups, or have some reasonable explanation for the difference, I'm investigating.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 05:57 PM
Quote:
Originally Posted by BigBadBabar
welcome pokercruncher guy
+1

Quote:
Originally Posted by rj999
But the app has to be right in even very edge-case setups, or have some reasonable explanation for the difference, I'm investigating.
Hope your coding goes better than mine has this week. Freaking interrupt handlers.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 06:29 PM
Quote:
Originally Posted by chillrob
Could someone link the PokerCruncher thread? Is the app available for Android?
http://forumserver.twoplustwo.com/16...uncher-429410/
3/6: Wake Up, San Francisco Quote
11-07-2014 , 08:39 PM
Update:
Bug in PokerCruncher is fixed on my side. But I'll wait a week or so to submit the fix (to all versions of the app), due to my upcoming move, and to test around some more. Below is the result I get now on my side; a match (the app's Tie number is double because it's total odds of tie-ing, not the 1/2 share individual tie equity). I tested on some other basic scenarios as well and am seeing the same correct results as before (this bug didn't affect cases where there's only one range anyway). But need to verify more.

Programmers out there may find what I needed to do to fix this entertaining - I didn't add any new code; I just *deleted* some code. I had an extra nested "try again" loop for cases where there were multiple ranges and they overlapped a lot with each other and the board cards so as to cause greater than "normal" conflicts/failures when picking random hands from the ranges. It turns out this extra nested try again loop was causing the bias towards AA in this scenario. After removing the nested retry and just relying on the top-level retry loop, all looks good. Sorry, duh on me in the first place after looking at that old code again today.

I'll send another update when the app update goes live (2..3 weeks away, will take some sweet time on this delicate one). Thanks for your understanding and good words and support everyone.

My local app fix; not released yet:
==========
PokerCruncher-Advanced-iPhone V.7.3.1

(Equity, Win, Tie)
Player 1: 46.1% 42.9% 6.55% [AcKh]
Player 2: 15.9% 15.9% 0% {KK+, 88, AKs, K2s+, AKo, K7o+}
Player 3: 38.0% 34.7% 6.55% {KK+, AKs, K9s+, AKo, KTo+}

Board: [Ah Ks Kd 8d ?]
Deal To: River
Dead Cards: {}

Monte Carlo Simulation: 720000 trials
3/6: Wake Up, San Francisco Quote
11-07-2014 , 08:58 PM
Quote:
Originally Posted by rj999
Update:
Bug in PokerCruncher is fixed on my side. But I'll wait a week or so to submit the fix (to all versions of the app), due to my upcoming move, and to test around some more. Below is the result I get now on my side; a match (the app's Tie number is double because it's total odds of tie-ing, not the 1/2 share individual tie equity). I tested on some other basic scenarios as well and am seeing the same correct results as before (this bug didn't affect cases where there's only one range anyway). But need to verify more.

Programmers out there may find what I needed to do to fix this entertaining - I didn't add any new code; I just *deleted* some code. I had an extra nested "try again" loop for cases where there were multiple ranges and they overlapped a lot with each other and the board cards so as to cause greater than "normal" conflicts/failures when picking random hands from the ranges. It turns out this extra nested try again loop was causing the bias towards AA in this scenario. After removing the nested retry and just relying on the top-level retry loop, all looks good. Sorry, duh on me in the first place after looking at that old code again today.

I'll send another update when the app update goes live (2..3 weeks away, will take some sweet time on this delicate one). Thanks for your understanding and good words and support everyone.

My local app fix; not released yet:
==========
PokerCruncher-Advanced-iPhone V.7.3.1

(Equity, Win, Tie)
Player 1: 46.1% 42.9% 6.55% [AcKh]
Player 2: 15.9% 15.9% 0% {KK+, 88, AKs, K2s+, AKo, K7o+}
Player 3: 38.0% 34.7% 6.55% {KK+, AKs, K9s+, AKo, KTo+}

Board: [Ah Ks Kd 8d ?]
Deal To: River
Dead Cards: {}

Monte Carlo Simulation: 720000 trials
it's cool you got a chance to fix this bug. it may have never come up if it weren't for the operator error. i'm not sure why anyone who held AK on AKKx would put KK in their opponent's ranges, since it's an impossibility. i'll have whatever jdr is smoking. pass the dutchie on the left hand side.


Last edited by rodeo; 11-07-2014 at 09:05 PM.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:12 PM
Also interesting, although Player 2's range as given includes AKs and AKo, it is actually impossible for him any AK hand based on the known cards and the range given to Player 3.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:20 PM
How come CO can have K9s but not 88?
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:26 PM
Quote:
Originally Posted by thesilverbail
How come CO can have K9s but not 88?
I suspect jdr0317 was thinking the CO would fold the flop with 88--an action I am extremely skeptical of given the fact that CO is getting nearly 16-1 to peel closing the action in that spot.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:41 PM
Quote:
Originally Posted by chillrob
Also interesting, although Player 2's range as given includes AKs and AKo, it is actually impossible for him any AK hand based on the known cards and the range given to Player 3.
Yes. Player2 can't have AKs or AKo. Player3 can't have AKs (but can have AKo). And neither player can have KK. And if Player2 picks up any hand with a King in it, that's the last King, and then Player3 has to have AA and you're crushed.

It's an interesting setup. I owe jdr0317 a drink for creating it and finding this bug. Thanks.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:42 PM
Quote:
Originally Posted by rj999
Yes. Player2 can't have AKs or AKo. Player3 can't have AKs (but can have AKo). And neither player can have KK. And if Player2 picks up any hand with a King in it, that's the last King, and then Player3 has to have AA and you're crushed.

It's an interesting setup. I owe jdr0317 a drink for creating it and finding this bug. Thanks.
no one has to have anything, including AA - not sure where this is coming from
3/6: Wake Up, San Francisco Quote
11-07-2014 , 09:48 PM
Quote:
Originally Posted by BigBadBabar
no one has to have anything, including AA - not sure where this is coming from
You may have missed the "if" part of what I wrote: "if Player2 picks up any hand with a King in it ...". Player2 has a range; in the cases where he picks up a hand from his range that has a King in it, then the only hand left that Player3 can have is AA.

At the top level, without any if's, yes both players have ranges and don't have to have anything specific.
3/6: Wake Up, San Francisco Quote
11-07-2014 , 10:11 PM
Quote:
Originally Posted by BigBadBabar
no one has to have anything, including AA - not sure where this is coming from
This is all based on the ranges given by OP and put in the simulators, not our opinions of what they can have.
3/6: Wake Up, San Francisco Quote
11-18-2014 , 08:49 PM
Quote:
Originally Posted by rodeo
it's cool you got a chance to fix this bug. it may have never come up if it weren't for the operator error. i'm not sure why anyone who held AK on AKKx would put KK in their opponent's ranges, since it's an impossibility. i'll have whatever jdr is smoking. pass the dutchie on the left hand side.
The fix for this edge case hand ranges bug in PokerCruncher is now live in the iPhone version of the app (version 7.4.1). Updates to the iPad, Mac, and Android versions of the app to follow in the next few weeks. Thanks again to jdr0317 for finding this bug, and apologies for having the bug in there in the first place (for 5+ years before jdr0317 found it).
3/6: Wake Up, San Francisco Quote
11-19-2014 , 01:53 AM
Quote:
Originally Posted by rj999
The fix for this edge case hand ranges bug in PokerCruncher is now live in the iPhone version of the app (version 7.4.1). Updates to the iPad, Mac, and Android versions of the app to follow in the next few weeks. Thanks again to jdr0317 for finding this bug, and apologies for having the bug in there in the first place (for 5+ years before jdr0317 found it).
I'm not exactly proud of the question I asked in order to find it, but once again, your app is outstanding. I actually don't really drink, but you'd find me in northeast midstakes games, probably jamming some combo draw after spewcapping preflop 6 ways again .
3/6: Wake Up, San Francisco Quote
11-19-2014 , 03:53 AM
Never heard of the app until this thread. Enjoy my $5.99

Should be fun messing around with this at the table instead of reading reddit.
3/6: Wake Up, San Francisco Quote
11-25-2014 , 03:11 AM
Haven't checked the forum in over 6 months and this is the first thread I read, so with that said, Here's my recommendation:

1. Fold the turn, only because you have better things to do at this point with your time.
2. Delete all equity calculators from all electronic devices you own.
3. Visit your local library, read chapter zero of any poker book published after the civil war (chapter normally titled "how to read the board - what hands beat me")
4. Make sure you understand the content from that chapter
5. Revisit hand
6. Laugh out loud
7. Profit
3/6: Wake Up, San Francisco Quote
11-29-2014 , 02:37 PM
Just go 'woooooo!' like it's a big party while you put the fourth bet in.
3/6: Wake Up, San Francisco Quote

      
m