Thursday, June 24, 2010

Just Conjecturin', Vol. 19: The Real Problem with the UB Hand Histories

Sorry for the bloggus interruptus, folks, but between the part-time gig and an emergency family trip to Wisconsin for ten days in mid-June, this stuff had to take a back seat. I also wanted to give Chops at Wicked Chops Poker a chance to publish Part 1 of his planned series on the UltimateBet scandal, and that was interesting, if a bit of an empty tease. Here's hoping there's more to come in future posts.

As I expected would happen back in April or May, the interference factor over at 2+2 has increased due to a couple of nasty trolls, one of whom seems to have intentionally sabotaged some of Chops' efforts in trying to speak to former and current UB employees. It's a shame that some people don't realize that the proper way to investigate a story is to listen to everyone, whether that person is presumed, innocent, guilty, insane or whatever, then compare the stories one receives and explore the differences therein. Unfortunately the tale behind the scandals is almost too big; too many people have things to hide, and thus efforts by good writers to explore the tale honestly are being messed with by people with hidden agendas.

But please don't feed the trolls, folks. I've done a little bit of that but am swearing off everything in that area except the exposing of lies, which unfortunately has to be done. Think about why the trolls might be doing what they're doing, and keep an open mind to all possibilities. Don't exclude the chance that certain factions involved in the behind-the-scenes struggles involving Cereus and its earlier component companies might even be willing to self-administer a flesh wound or two to throw others off the correct path. It's all quite Machiavellian, and it pays to keep that in mind as one attempts to learn from the unfolding saga, but all that is grist for a future post.

This time, it's back to those confounded, messy hand histories that UB sent out over the past several months in response to ongoing requests and complaints from players. The problem was that the hand histories were all but incomprehensible, often broken down into files that contained individual hands, rendered in Notepad file format. Barry Greenstein, one of the better souls in poker, posted the HH's that he received here, and many months ago I did some digging around in those files to see what I could find, first deciphering the formats and then trying to find instances of shady play.

The first part I accomplished pretty well, the second part not so much. It was only after I made the attempt to find "cheating" hands that the poker world learned that the UB hand-history batches that were sent out were pre-edited; the powers that be at Cereus had only included hands where the already-identified hands gave action against the players requesting the hand histories. And in Barry's case, he had not received a refund, meaning that the possible cheating hands were nothing more than... oddly played hands.

This was probably the same conundrum that a former WSOP world champion not affiliated with UB encountered when he agreed to examine thousands of possible cheating hands in conjunction with settlement talks between the current and former owners of UB. Except for the most obvious forms of cheating, such as when a cheater's K-K is folded to another player's A-A pre-flop or when a ridiculous bluff or call of a bluff is made (such as what happened at the end of the infamous PotRipper tourney), a lot of supposed cheating is indistinguishable from the variance in playing styles that is the hallmark of poker. It takes long-term patterns and unusual win rates to determine more subtle forms of cheating, and it's possible that outside of an internal expose, the methods undertaken by Michael Josem and others might have been the only viable way to uncover the cheating seen in the UB scandal.

But about those UB hand histories: Why exactly are they so messed up, so useless for any real analysis?

Given that Cereus has shown no interest in sending out complete hand histories for the affected players, one has to dig further. Company officials have floated claims that the older hand histories are on a different server, in another format, and are just generally difficult to access. It's been reported to me that UltimateBet's first few years of hand histories were saved in MySQL format, and that they ported over to Oracle in 2005 or 2006, but for programmer Uri Kozai to be able to access all supposed cheating hands in the 2003-07 time frame and derive a financial solution to the supposed cheating based on net wins and losses by the known cheating accounts, the company was clearly able to combine all hand histories into a coherent whole for examination.

And yet that ability seems to have disappeared when the players want their complete hand histories? Sorry, that one doesn't fly, despite the general incompetence demonstrated by Cereus's technical and support staff in recent months.

Let's assume a different starting line. Let's begin with the wholly plausible argument that Cereus could send out complete hand histories to players, but they just don't want to. Is there any evidence to support this hypothesis?

Yes. In this video interview, no less an UltimateBet persona than Annie Duke stated that publicly-named-as-cheater Russ Hamilton had 88 accounts at UB, as pointed out by 2+2 supermod Kevmath in this thread.

If one assumes that Duke's "88 accounts" story regarding Hamilton is true -- and it jives with what's been alleged to me by other sources -- then there's a ready-made explanation for why the hand histories as sent out are fractured and incomplete. Only 23 accounts and over 100 screen names, including multiple name changes, were identified. Most of those, but not all, have been traced to Russ, but even if every single one of those was a Russ account, it still leaves as many as 65 Russ Hamilton-controlled accounts that are unnamed and unaccounted for.

While it's possible that none of these other accounts were involved in the cheating, it's more likely that at least some of them were. Cereus probably did locate the largest among the cheating accounts, the ones that would be most likely to be uncovered by cheated players' analysis, but Hamilton by all reports did not provide the account names, the same way his share holdings had to be forcibly seized. In addition, Cereus and the KGC have provided no evidence that they investigated any accounts beyond those named within player complaints, nor did they submit to a truly independent audit, claims to the contrary notwithstanding.

Therefore, it's likely that hundreds of thousands or a few millions more in cheating went undiscovered when compared to the numbers officially released. Don't forget to add in the fact that the methodology used in determining refunds accounted only for action given between players and cheating accounts, and had no way to account for action wrongfully denied, as in the made-up K-K versus A-A example above. As massive as the total cheating amount was, it was almost surely underestimated.

And that returns us back to the bollixed hand histories. The real reason they're messed up is because Cereus likely does not want players looking at additional hands where some of these other still-hidden accounts may have played. A settlement between the new and old ownership groups was already reached, additional shares belonging to three others besides Hamilton were forfeited, and there's no way for the cash-strapped "new" Cereus to go back for a second helping of settlement funding. Given the other financial struggles and massive debt load Cereus faces, there's no way they can afford the risk of additional millions in refunds themselves.

The corporate answer, then, is just to obfuscate. Joe Sebok's good-faith efforts notwithstanding -- and despite my disagreement with his choices, he still did what he did in signing with Cereus in good faith -- I expect we'll see a fire engulf the MIT server farm before full UB hand histories ever see the light of day. Joe's efforts in that regard were fruitless and frivolous; he technically achieved his goal of getting out the hand histories... but not really. That's the sad and the short of it.

Wednesday, June 02, 2010

Holloway Killing Suspect at LAPT Peru?

I just caught this on the Yahoo! newswire, via an AP feed:

[By FRANKLIN BRICENO, Associated Press Writer Franklin Briceno, Associated Press Writer]

LIMA, Peru – A young Dutchman previously arrested in the 2005 disappearance of Alabama teen Natalee Holloway is the prime suspect in a weekend murder of a Peruvian woman, police said Wednesday.

Joran van der Sloot is being sought in the Sunday killing of 21-year-old Stephany Flores in a Lima hotel, Criminal police chief Gen. Cesar Guardia told a news conference. He said the suspect fled the country the next day by land to Chile.

The Dutch government said Interpol has issued an international arrest warrant for Van der Sloot.

Guardia said the 22-year-old Dutchman, who was in the country for a poker tournament, appears with the young woman in a video taken at a Lima casino early Sunday....

I'm gonna guess the guy was a no-show for the tourney if this is true. Maybe Otis or Shamus can check it out?