Turbo Texas Hold-Em: Real Deal?????

[Brooks]

I've had the TTHE software for about 3 weeks, and have so far played at least 30K hands with it. I've been playing a "normal" game against a mixed lineup of 9 other computer opponents. From my experiences so far, I have to ask: Does anyone else have doubts about the "randomness" of the dealt cards? Too many of these hands I've played seemed staged or rigged....I'm convinced that the program has predetermined who wins and when.

I've had streaks where I'd win practically every playable hand; suddenly the cards shift and I can't win with anything. During these losing streaks, I'll consistently see my pocket Aces or Kings bettered by small sets, my straights topped with larger ones or flushes, and boats beaten with larger boats or even 4 of a kind....I've seen more disgusting suckouts on the river than I can realistically believe. I'd try and adjust my play, play tighter, waiting for premium hands even in later position...nothing would work...I'm sure I must have been in the "losing loop" of the program.

My biggest nemesis in this game has been "Jane Doe". She's evidently an expert player; she's the only one in the money...about $17,000+ with every one else in minus figures (including Yours Truly, the Human player). She's also amazingly lucky; habitually administering bad beats on the rest of the table.

I was in the depths of one of these losing streaks that I had one very memorable beat. I held pocket kings and hit a set on the flop, raising from the start with Jane calling. On Fourth street a third club hits the board, now I'm worried she's hit a flush--sure enough, she reraises and I call. On the river comes a fourth club, the Ace. Now at last I've got her, since I hold the Kc, I've got the Nut flush....I triumphantly bet; Jane reraises! Huh...I automatically hit the R key to reraise, thinking Jane should 'realize' I've got her flush bettered, but again she reraises. It's then I scan the board again and notice that besides the Ace, the board has the J-9-10 of clubs (yeah, you got it!) "No f*****g way!", I yell and reraise. Yup, Jane reraises. I decide to end the torture and call...sure enough, Jane shows the 7-8c for a st. flush. Strange that a high-caliber player like Ms. Doe would be calling raises from mid-position with a 7-8 suited....she must be the resident psychic among the AI players! Again, the frequency of these kinds of showdowns makes me wonder about the true randomness of the programs card distribution. I've witnessed too many of these miracle draws....it simply defies the odds.

Anyway, enough of my rant about TTHE. In my net search of other software, I found the "AceSpade" line....they even quote Mr. Sklansky and Mr. Malmuth in recommending their software over TTHE. Anyone else have or know about AceSpade's Hold-Em software?

[Mason Malmuth]

Hi Brooks:

David and I recently reviewed the TTH software. We were appalled by how incredibly bad it played and sent a letter to this effect to Mr. Wilson.

The statement on the Ace Spades site is a little misleading. While we concluded that it plays better than TTH, we also told them that their program was making too many errors to meet our standards.

Best wishes,
Mason

[Ragnar]

Brooks,

I've played over 50,000 hands of TTH. I'm doing this because I had almost no poker experience of any kind before I took up the game less than a year ago. So I need seat time and it is cheap. I intersperse this with B & M play and study. I am well aware of Mason's current and previous criticisms of TTH, but it is helping my game.

I'd suggest changing the lineup about every 1000 hands. You can learn quite a bit about adjusting your play to different lineups that way. You've probably played way too many hands against one lineup. I'm guessing you are too invested in beating Jane.

Ragnar

[Brooks]

Ragnor,

I'm sure you're right about my playing too many hands against one lineup. I understand that the AI players are programmed to supposedly adjust to a person's style of play. In the case of 'Jane Doe' they can also program a player to be as "wise" as it takes to do most of the winning.

My main criticism concerns the randomness of the cards; just seems like they could do a much better job. If there's one thing computers are good at, it's processing numbers. Surely they can come up with a program that can generate a more random deal sequence of every possbile combination with a 52 card deck.

Let me give you a brief update on some more of my play with TTHE. I decided to start a new $20/40 game with the same lineup; I now vow to closely follow Jane's play and take notes. These are the facts: Jane jumps to a significant money lead, winning 3 of the first 10 hands (on hand #6, she beats a full house with 4 Kings!!) In the first 20 hands, Jane has a lead of over $600, the nearest other player is at $274. I've already seen enough when on hand #26 Jane tops my two pairs by hitting an Ace-Hi straight on 5th street....no surprise to me anymore, I was expecting it after the turn!

To put it plainly, it's obvious to me the program is "dealing" to her. Certainly these hands do happen in real play, but the prevalence of these hands that I've seen in TTHE (and the ratio of who gets them) is why I deem this software a less than realistic hold-em poker simulation.

[Sheriff Fatman]

I think you're being too paranoid about the program. After all you can control the hands to some extent by using the fixed deal numbers (which are designed to let you run the exact same hands in different simulations).

I've done simulations on TTHE and seen short term effects still present after 100,000 hands. You really need to run 1,000,000 hands to see any reliable figures.

You should find that the strongest players are Conan the Librarian and Smiling Mack. Jane Doe plays 38% pre-flop and will not be a long-term winner.

Also, one good play to test your play is to not display the names of your opponents (which tend to give away their playing style) and to have frequent seat changes. This will undoubtedly help improve your reading skills.

[Lunamondo]

You can run simulations and see their strengths. Jane is an average player and a loser accordingly. I have that player currently at one of my ten player lineup right at my left where she's possibly hardest to handle and when she is running good she gives as many frustrating bad beats as other average players, perhaps more so at your (immediate) left.

You have got a good experience of fluctuations, and about that type of average players; that's how it's in real life, they can have weeks of bad beats in store for you.

In many lineups especially there are long sessions of break even to losing mega sessions, but that's when one has not been running lucky, while when one runs lucky it "suddenly" pulls one's win amount up, but all that from some part also depend of the type of lineup one is playing against.

There are times when players like Jane together with some other more or less average players are beating you and beating you and beating you, it becoming such a streak it becomes a hell. But it will not last forever and after that one wonders why that lineup is not giving such hell anymore for a long time.

Different lineups create different type of fluctuations, and even just one good player becoming more or less average player or the other way around can change the type of game and its fluctuations.

[Ragnar]

Brooks,

I was only responding to one portion of your post and not to the hands that Jane is getting. From your account they do seem weird.

I often turn the function that allows the computer players to adjust to the play of the others and me. I did it for a different reason. I'm trying to learn to play against particular kinds of players, especially those I see in my brick and mortar games. Since I play low limit, they don't adjust very well. The second reson I turn it off is that they would start to make very strange plays after awhile.

I'd construct a lineup without Jane and forget about her. When I first started playing I played against the low limit lineups. One of the players is Elsworth Tooey, which is a misspelling of the Ayn Rand character Ellsworth Toohey (probably for length reasons). As Toohey is one of the villains in Rand's literature and I am a Rand devotee I started trying to beat him, instead of win money. Of course that can often happen with B & M opponents when you dislike them as well. (Although I'll admit that focusing on a computer lineup with the name of a character is a little bizzare.) It is not conducive to good play. I learned to play against the lineup in general and not focus on him.


Ragnar

[Brooks]

Thank you all for the feedback, I appreciate it!

I'm not completely negative on TTHE; I like many of it's features; think the graphics are great. What I was looking for with this program was to get some volume exposure to Hold-Em Poker, and get a better sense of the game's statistics; I was a little bothered that I might be getting a distorted view of it's actual probability of card play.

I want to report that another new game with my now too familiar lineup yielded pretty much the same results; this time Jane Doe surged to an even more substantial lead; at the end of 50 hands she had second place more than doubled with $1440. However, the start of a fourth game finally had poor J.D. struggling near the bottom after 50 hands, while I just managed to stay in the black with $200+ (heeeeeey ... maybe this program ain't so bad after all!) /images/graemlins/laugh.gif /images/graemlins/grin.gif Anyway I'll be doing a lot more playing and testing (and learning) with this software; will definitely try some different lineups (give Jane a rest; she needs to regain her touch!). By the way, I wonder, on average, how many years of playing would it take for 1 million hands of real Hold'em .... that's a little scary to me. /images/graemlins/tongue.gif

[toots]

I find the program adequate as an adjunct to real B&M experience and any studying I want to do.

I do think it's important to get fluent at modifying the profiles and/or changing the advisor to test your style of play, and running simulations to see how well it does. Tedious, but useful.

But, the program really honks me off. It's a Windows 3.1 program trying to run under XP. Of particular annoyance is that when I have my second monitor enabled, it completely screws its table size so that I can't even see my own hand (at position 10). I've had to tweak the startup link on it to disable the second monitor, start the program, wait 15 seconds, then re-enable the second monitor after the program's had a chance to start.

[Sheriff Fatman]

The point is that you cannot make any judgement on playing styles after 50 or 100 hands. For a start, the quality of cards dealt is still a massive factor here. If you take this view into a real cash game you will end up tilting big style when the suckout artists have good runs whilst playing badly.

I now have 12,000 logged hands on real cash games (still nothing compared to others on this forum). The 3 opponents who have made most money are seeing flops as follows:

1) 90% of flops (42 hands)
2) 51% of flops (86 hands)
3) 50% of flops (48 hands)

Does this mean they are good players? Absolutely not - they will all lose money in the long run because of the way they play. The top player got there by capping every street of every hand he played, yet he still won money over a sample size similar to yours because he won 60% of the hands that he was showing down (with random cards).

You simply cannot make judgements over a small number of hands. If you do, be prepared to be upset frequently.

Here's a test for you. Run your line up in simulation mode for 100 hands, then 1,000 hands, then 10,000, then 100,000 hands, then 200,000. If they're not already in there, add Conan or Smiling Mack to the lineup. Let me know whether Jane Doe is still a winning player at the end. The rankings after 10,000 hands onwards will be very different to the earlier ones. However, you should see a pattern developing by then.

Run the same test as many times as you like - I'm confident that the same players will come out on top by the end but that your short term results will differ every time.

Believe me, you need to understand this point before you take your game any further. The 'short-term view' is very dangerous to any poker player as, based on what you see others doing, it can encourage you to lose a lot of money.

Let me know how you get on.

[EVIL]

Your skepticism about the "randomness" of the cards in TTHE is probably rooted in the fact that random data often appears less random than people expect.

I doubt there is a serious problem with the randomness of the cards in TTHE, since it is not much of a mystery how to write a RNG that would be good enough for the purposes of that software, and probably much easier than trying to "stage" or "rig" the hands in some way. Also, Wilson Software has no incentive to have anything but randomly dealt cards in TTHE.