Easystreet Sports confiscates $46,000 for alleged illegal software use, but provides no compelling evidence

Easystreet Sports confiscates $46,000 for alleged illegal software use, but provides no compelling evidence

This is another fairly long article, regarding sportsbook Easystreet's confiscation of $46,000 from a US player on the basis of unsubstantiated allegations of robot play; you can navigate using the contents links below.


• The case
• Evidence presented by casino
• Offer made to player
• Report from software engineer
• Response to report criticism
• Conclusions
• Useful links



EastStreet sportsbook and casino recently confiscated a $46,000 win from a video poker player on the basis that robot software was used, an accusation apparently corroborated by an improbably fast play rate of 18 hands per minute over extended periods of time and the fact that his behaviour after hitting his various $20,000 royal flushes was not consistent with normal gambler activity - he did not stop directly, but played on for another one or two hands.

'Bots (robot software) are deemed illegal in the casino terms and conditions:

Artificial Intelligence - Robots:

You are not allowed to use any software program which, in our opinion, is endowed with artificial intelligence ("AI Software") in connection with your use of the Service.

We constantly review the use of the Service in order to detect the use of AI Software and in the event that we deem it has been used we reserve the right to take any action we see fit, including immediately blocking access to the Service to the offending user, terminating such user's account and seizing all monies held in such account.

What constitutes "artificial intelligence" is a bit of an unnecessary complication and serves to substantially muddy the waters, but it's reasonable to infer that a software programme that performs the mechanics of dealing, selecting and discarding cards and whose decision-making to achieve those ends represents optimal play can be considered a violation of this rule.


The matter was initially commented on in a report at Sportsbook Review, and they also made a Youtube video on the subject.


The player has a history of fraud of one kind or another, but there is no evidence of such in this case, which revolves exclusively around the allegation of the use of illicit software. Additional useful information is that the player made a total of about nine deposits of $250, receiving a 100% bonus for each. The wagering requirements were 25 times the value of the bonus, which yields substantial profit value for the player. This is very relevant to the case.


Let's take a look at what the casino offers as indicative of illicit software use:


1) He played in a manner not humanly possible, around 18 hands per minute for extended periods of time.

This is an average of one hand per 3.3 seconds, which is certainly fast, but not at all impossible if the software has an autohold feature in which paying hands and other typical holds, such as low pairs, are automatically held. A skillful, experienced player will be able to play close to whatever speed could be achieved by a purpose-built machine, as the thinking time needed is extremely short. The speed of play here neither disproves nor proves 'bot use.

Another sticking point for the casino's allegation is that at no point was the player's skill level determined, as only the final hand, and not the play decisions which led to it, was examined. It's clearly more time-consuming to make the perfect play each time than to play in completely random fashion, as the former requires time for thinking and the latter does not. As it stands we do not know this player's skill level, which makes it that much harder to make any kind of a judgement against him.


2) He did not stop directly after hitting his $20,000 royal flushes. Surely a gambler would not behave like this?

Of the two royals for which play logs have been provided, we can see that on the first he stopped one hand later:




After the second, he stopped after two further hands:





As such, it's a tad misleading to say he "didn't stop", as he clearly did. However, why he did not stop immediately is at least an interesting question.

The player was playing fast and hard - noone would question this. It's also very likely that he was concerned primarily with getting through the wagering requirement of the bonus and not especially focussed on the outcome of each hand. In either event - disinterested bonus player or feverish gambler - it isn't unreasonable to fail to visually register the appearance of the royal until after the next hand has been clicked into - the process of dealing the initial hand, choosing the cards to hold, drawing the new cards and then dealing the next hand becomes almost hypnotically automatic at that speed of play. This is therefore reasonably consistent with normal, manual play.


How would this post-royal stop be explained if the player WERE using a 'bot?

Since play did stop, either the 'bot would have been programmed to stop by itself or would have been stopped manually by the player. It cannot rationally be suggested that it was programmed to stop, as it would be pointless to stop once the jackpot hand was gone. As such, it could only have been manually stopped by the player, in which case the scenario would be that he was watching the game being played by the 'bot, manually stopping it to take a break after hitting the jackpot. This would be very strange, as the whole purpose of using a 'bot in the first place is to avoid having to be involved with the play - you would simply leave it to click its way through the wagering requirements.


The behaviour on the royals seems therefore to suggest manaul play with no 'bot software - the opposite of what the casino alleges.


RX forum moderator Wilheim, acting as mediator between casino and player since Easysports is an advertiser at The RX, put together a resolution offer for the player:

First you will need to take a polygraph conducted by an accredited professional polygraph purveyor. The polygraph will only address your play at EasyStreet during the session in which you hit your Royals, no questions about your past will be used against as far as this polygraph goes, EasyStreet simply wants to get at the truth and nothing else whatever it is. I am sure they actually hope you pass so they can in good conscience pay you.

Should you pass the the polygraph then an appointment will be set up shortly after (maybe the next day) to proceed to the offices of DGS for the video poker demonstration. You will be asked to come relatively close to the approximately 18 hands of perfect strategy on a 5 card draw video poker game per minute for a little more than five hours. The game will be identical to the one you played on when you won the Royals. You can have representatives from SBR or elsewhere accompany you during the entire process. At no time will you be asked to be alone with anyone you feel uncomfortable with. During the polygraph I imagine spectators will be asked to watch from an adjoining room. Should you fail the polygraph there will be no need to take the DGS test.

Should you pass both tests you will immediately be paid your entire balance and be allowed to either fly back to the US or remain in Costa Rica and actually visit EasyStreet and probably SBR should you care to.

Polygraph tests are not foolproof, and there are potential issues of testing bias which add further complications; also, the request for five hours play of perfect strategy was unacceptable because at no point was either five hours' play demonstrated or perfect strategy established. The player refused the offer.


In passing, I'll add that a more reasonable offer might have been along the lines of:

1) A polygraph test, on a strictly neutral basis within a reasonable distance of the player's home in the USA, with at least one back up test to help achieve consistency of results.

2) A play test of no more than two and a quarter hours' duration, requiring a skill-level that has not as yet been established but which would need to be in order for this to go ahead.

This was, however, never proposed, and I suspect the player wouldn't have agreed to it if it had been.


The RX then apparently engaged a software engineer to examine the play for the purposes of making a final ruling on the matter. This can be read in the Final decision of the dispute at Easystreetsports Casino discussion thread. The report reads as follows:

My Background:

6 years as a Gaming Engineer and software architect in the US. Projects include Wynn, Encore, Caesars Palace, Native Games America and IGT/Acres platform support.

Research and Conclusion:

Based on my independent research into the issue I have come to the following conclusions.

1) A human did not play the 8762 hands of video poker that were examined. This conclusion is based on the fact that the "player" played an avg. of 17.6 hands of video poker per minute for 499 minutes without a single error. This is a statistical impossibility.

2) It's been stated that perhaps the auto-play feature was in use at the time and that the "player" was simply using the auto play feature to achieve his abnormally high rate of perfect play. This feature was verified to NOT be enabled and consequently unless the player somehow breached the platform security (nothing suggests this occurred), toggled the feature on, set a more advanced strategy than is currently available to that feature, toggled the feature off, and then wiped the logs; this as well is very unlikely.

3) The "player" had no apparent reaction to hitting the 3 royals (in fact playing straight through the royals at a continued rate of ~3 seconds per hand) and was unable to accurately answer whether he was dealt a royal (as he stated) or that he held 2 cards and then received a royal (which he did).

4) The odds of a player hitting a single royal flush is roughly 1 in 40,000. The odds of a dealt royal flush (the player stated he received a dealt royal flush) is 1 in 649,740. The odds of hitting 3 royal flushes in 8762 hands of poker is statistically impossible. In fact in all of the years I've been in gaming I've NEVER seen that happen (and I've reviewed millions of hands of poker).

5) Load tests on the system show an average screen draw time of approx 1.3 seconds, this leaves only 1.7 seconds for the "player" to recognize all of the cards on screen, compute optimal strategy, physically issue whatever action he wanted, and the system to receive that action and begin a new hand. While possible (though incredibly unlikely), it's even less likely that the "player" could keep the rate of play up with no discernible alteration in strategy, timing, etc for 136 minutes (which was the longest non interrupted play period).

6) Based on my review of the play logs, research of the EasyStreet system, and discussions with other industry professionals; it is my professional opinion that the player used a bot or some other form of machine augmented assistance to play the hands at a rate fast enough to attempt to overwhelm the RNG and provide favorable odds to the "player".

A software engineer is neither a mathematician nor a psychologist, yet he claims just such expertise in this report.

• He states that it is "statistically impossible" to play many hands without an error, yet this is not his area of expertise, and irrespective of this, perfect play was never determined.

• He states that the player did not stop on the royals, which is a misrepresentation of the situation as the player stopped after one or two additional hands. It seems odd that a neutral third party would misrepresent thus.

• The comment "the odds of hitting 3 royal flushes in 8762 hands of poker is statistically impossible" is patently absurd and displays ignorance on several levels. In the first place, such an occurance is entirely statistically possible; in the second, the number of hands given includes only those sessions which contained a royal and ignores all the other sessions the player played; in the third, using selective data to "prove" your contention is an entirely invalid way to conduct statistical tests to begin with; and finally, the probability was never questioned as the software provider, DGS Systems, had already discounted the possibility of the player manipulating the software.

Is this software engineer accusing the software of being non-random? And why is he choosing to comment on mathematics, an area in which he is not qualified to offer an opinion?

• The report descends into farce at the end:

The player used a bot...to play the hands at a rate fast enough to attempt to overwhelm the RNG and provide favorable odds to the "player".

How do you "overwhelm" an RNG, and how does this happen by simply playing fast?

Why would such "overwhelming" cause the mathematics of the game - the odds - to change favourably for the player?


There is nothing in this report that proves bot use, and it is shot through with inaccuracies, ignorance and misrepresentations.


Following the release of this farcical report, Easystreet announced a video poker contest. The prize?

Total cash awards could exceed $46,000.

How convenient.


A day after the RX moderator's final report was issued, SBR published a video of a phone conversation with the player. Among the matters covered were his previous allegedly fraudulent activity at other sportsbooks and his reason for not immediately stopping on the royal flushes. No answer was forthcoming to the fraud allegations, but useful information was provided about the play after the royals.


Following the general criticism that followed the inaccurate and ill-informed report from the unnamed expert, he passed on his rebuttal to the feedback a few days later:

A few points, which I'll try to answer in the order that they are generally posted:

A) It's come into question if I reviewed the purported 22k total hands that the player played. It's clearly stated that I reviewed 8762. I reviewed the actual hands the player was dealt, what he kept, what was drawn, the avg screen draw time, the actual execution time, player response time, etc. The spreadsheet that was posted is obviously missing this information, but my review was on-site, in front of a management terminal, and with access to other pertinent information to base my decision on.

B) The word impossible should really be improbable (or highly unlikely). The probability of a player playing at that rate of play for that amount of time with perfect strategy (or at the least no discernible deviation in the strategy being used, whether it's perfect or not is, of course, purely an opinion) is very unlikely, improbable, etc. I apologize for the use of the word impossible, once again, it should have been highly unlikely, improbable, etc (I've incidentally never seen it happen on any video poker machine in any casino or casino bar in las vegas, laughlin, pechanga, or anywhere else I've worked, advised, contracted, etc.)

C) I play 20+ hours of poker a week on avg, video and table based poker for my job AND personal recreation, and in ~10 years of playing, I've never seen it happen. It's not impossible that it happened, but it is highly unlikely.

D) The 3 royals worked out to roughly a 1 in ~700 chance (I believe this has been stated elsewhere as well). Once again, highly unlikely (or improbable) is a better way to say this and I apologize for the use of the word impossible.

E) In reviewing the logs of the player, and the logs of other players, I still believe that the player used a bot to obtain his rate of play for the period of play in question. If it was in fact a human that played at that rate of play, with that strategy and no discernible signs of fatigue, my hats off to him or her. I've never seen anyone play that fast, for that long, that well. I was asked to determine if the player was human. It is still my opinion, after reviewing all of the information available to me at the time, that the player of the hands in question was not human.

This rebuttal is as shot through with holes as the original report. In no particular order:


• The expert has still failed to identify himself. This is procedurally bizarre, but maybe not surprising given the incompetence and ignorance he or she displays.

• The word "impossible" has been downgraded to "unlikely". Disregarding the fact that odds of one in 700 are no more "unlikely" than they are "impossible", this demonstrates an extraordinarily cavalier attitude on the part of the "expert" in question. Did he put such little thought into his original wording?

• He claims that perfect strategy is "an opinion". This is patently false. Perfect strategy, or the optimal way to play any given hand, is a matter of fact. The expert evidently does not understand this.

• He claims "no discernible deviation" in the player's strategy. With no established benchmark - he does not recognise optimal play - how do you establish deviation? You cannot deviate from a random strategy.

• He claims that odds of one in 700 are "highly unlikely". This is irrelevant. It lies within reasonable probability. Additionally, it ignores the fact that these odds are themselves incorrect as they are based on only the sessions containing royals, and statistical analysis cannot be made with selective data. He appears unaware of this.

• The expert never addressed the matter of what was meant, in his original report, by "overwhelming the RNG", nor how such overwhelming, if it could happen, would cause the odds to shift in the player's favour. This is not entirely surprising, as although statistical analysis is not necessarily the forte of a software engineer, matters of RNG behaviour are very much his department, and to make such inaccurate statements within his area of expertise is inexcusable.


Both the initial report and the follow up remarks demonstarte substantial ignorance of the relevant mathematics and even the specific area of expertise of software operation. It has little credibility as a professional appraisal, and does not prove or disprove 'bot use.


In conclusion: casinos are right to forbid the use of software robots, as these devices are typically used by profit-orientated players who seek to gain multiple bonuses via the use of accounts other than their own, be they genuine accounts managed by the player on behalf of family members and friends / colleagues, or fraudulent accounts based on stolen or faked IDs. However, the fundemental fault here lies with the casino in offering large bonuses with lax wagering requirements which are ripe for easy exploitation. Easystreet is a sports-focussed sportsbook with a casino sideline, which explains but does not excuse their poor casino management. Not for the first time, the player is being asked to pay for the casino's mistakes in offering easy bonuses.


The correct way to handle a matter like this is to pay the player if he is genuinely owed (the lack of any compelling evidence to the contrary here suggests that he is) and solve the problem - in this case, tighten up the bonus rules. The wrong way to go about it is to incorrectly blame the player and claw back the money which the casino lost essentially through its own inefficiency.


Other useful links:

Sportsbook Review: EasyStreet casino winner accused of using robot software

Sportsbook Review: EZStreet theft, deception and TheRx

RX forum: Easystreet

Wizard Of Odds: Easy Street Sports

Casinomeister: EasyStreetSports Casino steals $46K


(Page top)

Post a Comment

May 2005 | June 2005 | July 2005 | September 2005 | October 2005 | November 2005 | December 2005 | January 2006 | February 2006 | March 2006 | April 2006 | May 2006 | August 2006 | October 2006 | January 2007 | February 2007 | March 2007 | May 2007 | June 2007 | July 2007 | January 2008 | February 2008 | March 2008 | April 2008 | June 2008 | July 2008 | September 2008 | October 2008 | December 2008 | January 2009 | February 2009 | March 2009 | May 2009 | June 2009 | July 2009 | August 2009 | September 2009 | October 2009 | November 2009 | December 2009 | January 2010 | February 2010 | March 2010 | April 2010 | May 2010 | June 2010 | July 2010 | August 2010 | October 2010 | November 2010 | December 2010 | January 2011 | February 2011 | March 2011 | April 2011 | May 2011 | June 2011 | July 2011 | August 2011 | September 2011 | December 2011 | February 2012 | May 2012 | July 2012 | August 2012 | March 2016 | April 2016 | June 2016 | November 2016 | December 2016 | March 2017 | May 2017 | June 2017 | August 2017 | August 2021 | October 2021 | May 2022 | December 2023 | May 2024 | Atom feed