BBO Discussion Forums: Robot race strategies - BBO Discussion Forums

Jump to content

  • 8 Pages +
  • « First
  • 4
  • 5
  • 6
  • 7
  • 8
  • You cannot start a new topic
  • You cannot reply to this topic

Robot race strategies

#101 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2010-November-12, 21:35

I've been wishing that the Robot Reward tourneys you should be given the option of finishing the current hand when the time runs out. It's really frustrating to bid a good slam and then have time run out.

But there needs to be some limit, so you can't prevent the tourney from ending. As you point out, sometimes it's the bots delaying the game, so perhaps it should allow 5 seconds per trick, but not count the time that the bots spend thinking.

But it's not really a racket, because BBO can't gain from this. They always pay out the same amount of winnings, it just goes to different players. And for every time you don't catch up because time runs out, there are dozens of times where someone doesn't overtake you for the same reason. So as long as the rules are the same for everyone, it's fair.

#102 User is offline   free0022 

  • Pip
  • Group: Members
  • Posts: 2
  • Joined: 2008-September-03

Posted 2010-November-21, 15:08

I play 20-24 boards when I have good connection.usually i play only 19+ hands(but vul may play a little less) or good shape 16+. Never open 1N.
0

#103 User is offline   free0022 

  • Pip
  • Group: Members
  • Posts: 2
  • Joined: 2008-September-03

Posted 2010-November-21, 15:14

View PostScoti, on 2010-July-06, 00:17, said:

Wow really?? Can my Robo's get some lessons from yours!? Seriously, they are not only slower (which I am gathering may be my connection speed) but Robo-P will always manage to overbid one trick. If it happens to find a correct bid, finds a way to lose a trick.
My last hand the bidding went something like: 1!s P 2!h P, 3!c P 3!h P, 4!d P 4!h P, P(I give up) P
I had a big hand for anything but my heart void. P reveals his hand, 5 hearts with 10 high & the rest under 8. Lol & cry . . .

that's your fault,first,you should not try 3c,that's a bad bid,just simple 2S and see what partner had(they will always bid their 2nd suit...
2nd, after partner 3h rebid,you should absolutely bid 3NT... 4d just looking for trouble
0

#104 User is offline   mike777 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 16,739
  • Joined: 2003-October-07
  • Gender:Male

Posted 2010-November-22, 15:39

FWIW I have never played these games but here is a strategy I was told. This is for the 12 bd games to be played in 15 minutes. 4 games per hour, hour after hour. The goal is to average 60%.



1C – balanced hand with 11-12 points. Pass the response when no game is likely.

1♥,1S six card suit, 11-16 points.

1NT opener -- Any distribution with less than a 6 card major and 13-16 high cards, plus any UN-balanced hands (with less than 6 card major) and 11-12 points.

2NT opener --- Same as 1NT, except more than 16 high cards.

4♥, 4S 6+ card suit, 17+ points.



Otherwise don't open 1d or at the two level.
0

#105 User is offline   pretender 

  • PipPipPip
  • Group: Full Members
  • Posts: 78
  • Joined: 2010-February-08

Posted 2011-January-23, 19:07

I would say that I'm fairly good at these best hand robot races.

Here's one strategy I'll share although I think most of you probably have come up with it yourselves already:

When partner bids something and RHO doubles, REDOUBLE!

It's amazing the number of times you get to play it there, often with overtricks. And if not, you just place the contract as you originally would. GIB N won't do anything stupid, and it's not like you needed its cooperation in the first place.
0

#106 User is offline   cvcherry 

  • PipPip
  • Group: Members
  • Posts: 13
  • Joined: 2010-October-25

Posted 2011-January-24, 19:36

In the new bridge bingo game, the contracts in the bingo board play a very important role. When your hand seems not match with the bingo board, I almost always pass with a minimal opening hand
0

#107 User is offline   Downtown13 

  • Pip
  • Group: Members
  • Posts: 1
  • Joined: 2009-August-31

Posted 2011-February-01, 18:17

I have played a bunch of these tournaments over the past three months, so I will share some tips.

I generally get between 20-25 hands a round (not including passouts). The most I ever got through was 29-30 (not including passouts, though that is rare.

Your robot partner generally plays pretty well, and you should obviously try to get it to play the hand. It can make some very difficult contracts, but also occasionally makes mistakes, but overall it plays better than I do. It doesn't do stupid stuff like misclick, or forget what the contract is, which is a benefit.

GIB will sometimes promise one more card in a suit that it actually has, so be careful of that. GIB sometimes acts as if its suit is better than it actually is, so be weary when it says strong and rebiddable.

If you promise a minimum of 4 cards GIB will sometimes raise you with 3, leaving you in a seven card fit. GIB doesn't seem to differentiate the difference and importance between 7 and 8 card fits at times.

East-West can be inconsistent on defense. Sometimes they defend very well, and sometimes they defend very poorly. East-West has some difficulty defending 3NT at times. Sometimes they will block their long suit, not take or lead the tricks they have in their long suit, or lead away from an ace or duck on the first round of their long suit when leading it.

Use Fourth Suit Forcing. I like to use this to force to game. Often times in these situations you will want to be in No Trump, so doing this can try to get GIB to bid 2NT or 3NT. Sometimes it will rebid a minor so you will have to bid 3NT. This conventions also is helpful in showing secondary support.

Remember if your partner opens 1H, you respond 1S, and partner rebids 2H it promises 6. If your partners opens 1 of a major, you bid 2 of a minor, and they rebid 2 of the same major, it only promises 5.

Sometimes you can get two suits of the same color confused, so be careful about that.

When I started playing these I opened No Trump, but I stopped opening No Tump. If I have 16-20 hcp, I generally open one of a minor, partner responds one of a major, and I jump to 2NT. Partner can use New Minor forcing to show 5, Rebid the suit to show 6 or more, and raise to 3NT if they only have four.

Use New Minor Forcing after your partner opens at the one level and responds 1NT, if you have a 5 or 6 card major in the suit that you responded. This is a generally a good sequence.

I generally don't bid grand slams. It takes too much time to explore them, and I don't want to go down with that many points, so I usually just settle at small slam.

I usually just ignore control showing cue bids and put the contract where I want it. I sometimes ignore the Blackwood response and jump to slam if I am rushed for time.

Sometimes when your partner bids Blackwood, you should pass if you think responding will get you too high, or if 4NT is an okay contract.

It is okay to bid Jacoby 2NT with only 3 card support.

I am pretty aggressive about making takeout doubles, and it has worked for me. Occasionally I will get a big negative score, but that does not happen too often. I will make takeout doubles with only two cards in a suit, even a major. If you are in a part score battle, sometimes it is best for your partner to play a contract even if it will go down to save time.

Don't open one of a major if you won't jump to four or higher if partner raises to 2 of a major.

If you have a 4 card major and open one of minor, don't open unless you will raise partner's response of one of a major to at least the 3 level. I never raise partner's major response to only 2, because I wouldn't have opened that hand.

Sometimes GIB will take a contract out of 3NT and put you in 4/5 of a minor or 4 of a major on a seven card fit. Sometimes this is bad, but it also works out well, because it avoids 3NT without a stopper sometimes.

If you have say AJ1098 in spades and partner has KQ in spades, it is usually better to be in 4 spades that than in 3NT when you have 4 or 5 losers right off the top.

I am also aggressive at bidding Michaels and Unusual 2NT. It is important not to get carried away when partner responds. If you raise to game, or double an opponent's game, your partner might think you have 25-30 points, and will overbid.

If there are a few minutes or less left, and the opponents bid a game, you might as well double. If you can set it you can get a bonus. If you can't set it don't finish the hand.

You can try to trick GIB sometimes, such as making weak bids and weak jump overcalls with strong hands. Occasionally GIB will overbid you, and you can double them. GIB might also double you, and then you can redouble that.

If there are a few minutes or less left, and you are declarer and the opponents double, you can redouble and try to make the contract. If you can't make it, don't finish.

Sometimes you should turn autoplay singletons off when time is almost expiring if opponents are in game or a doubled contract etc, so you can get to the last trick without playing to see if they will make it or not. Make sure to turn autoplay singletons on when you play robot reward again.

I don't have a specific requirement to open hands. It partially depends on if partner is a passed hand, whether we are vulnerable or not, etc. If I have a balanced hand, the minimum to open is generally 16-17 hcp. The more unbalanced the hand, the fewer high card points I will open with. If I have length in the majors, I am more likely to open than if I have length in the minors. I am generally pretty picky about opening hands.

If the opponents make a takeout double of a contract you think you can make, you can redouble and sometimes it will be left in, and you can get a nice bonus. Two posters above talked about this.

If you are a passed hand and your partner opens one of a suit, bidding two of a lower ranking suit might be passed. I like to bid 3NT in these situations unless my hand is really unbalanced. If partners opens 1 of a major, I like to bid 3NT with two card support.

When declaring make sure to keep track of how much trump are out. Obviously you should always do this, but in timed tournaments you can't try to focus on too many things.

Try to anticipate what the opponents will do, but be careful not to react too fast so you don't play the wrong card if the opponents play an unexpected card.

Sometimes you will get really good hands in a tournament, and sometimes you will get really bad hands. Also, you can have a bad tournament, but get a good/ok score with a few good hands, and vice versa.

If partner opens one of a major, you can jump to 6 to save time. Sometimes you will be off two cashing aces, but by jumping to slam, the defense might not take them. GIB will not raise you to seven, although it used to do that.

GIB doesn't fully understand responding to your takeout double of a preempt or a weak two, and will leave it in and they will make the doubled contract when they should really takeout. They might do this with a hand with 5 or less hcp and 3 or 4 cards in the suit preempted. I don't know how to alleviate this.

If you try to make a penalty double of a game after earlier making takeout double, GIB might interpret it as a takeout double, and you might get into a bad contract.

Be careful when raising partners response to a takeout double. A free raise by your partner can be pretty light.

Be careful overcalling 1NT when vulnerable.

Another tip is if you are in No Trump and have Ax and your partner has xxx or xx in the same suit, hold up the first round, and sometimes the opponents will switch to a more favorable suit. Sometimes East-West has trouble identifying a strong suit in No Trump.

One hard thing to decide is whether to raise a preempt or weak 2 to game, pass, or bid 3NT. If anyone has any tips on this, feel free to share.

Slams really help, but the key to doing well is to make a bunch of games.

You can go look at your hands and results after the tournaments, so that can help.

The 1 dollar robot reward hands generally don't have as stiff of competition as the 5 dollar robot reward hands, so if you are new to the robot games, you might want to start with the 1 dollar games. The robot games are better when there are more people. More masterpoints and more money is available.

You do not have to be a great player to be successful at the robot tournaments. I am an intermediate player with about 85 ACBL masterpoints, and I haven't played an in person ACBL club game/tournament since 2008. The key is to master the format of the game, and don't quit if you think you can have success.

Back in November I would buy 10 dollars at a time and play tournaments, mostly robot tournaments until I ran out of money, or got to a dollar or less. I eventually got better and went from 10-11 dollars to a little over 750 dollars. I would not have envisioned myself doing this back in Mid-November. I made 7 different ten dollar deposits in November before I was able to get enough money to not have to buy anymore bridge base dollars. I haven't bought any more bridge base dollars since.

One suggestion to get better is to just keep playing, be patient, and practice and try things to get better. That worked for me.
0

#108 User is offline   Phil 

  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 9,903
  • Joined: 2008-December-11
  • Gender:Male
  • Location:North Texas, USA
  • Interests:Getting back into golf after a very long layoff.

Posted 2011-February-01, 19:55

View Postmike777, on 2010-November-22, 15:39, said:

FWIW I have never played these games but here is a strategy I was told. This is for the 12 bd games to be played in 15 minutes. 4 games per hour, hour after hour. The goal is to average 60%.



1C – balanced hand with 11-12 points. Pass the response when no game is likely.

1♥,1S six card suit, 11-16 points.

1NT opener -- Any distribution with less than a 6 card major and 13-16 high cards, plus any UN-balanced hands (with less than 6 card major) and 11-12 points.

2NT opener --- Same as 1NT, except more than 16 high cards.

4♥, 4S 6+ card suit, 17+ points.



Otherwise don't open 1d or at the two level.


Re: 12 board robodoops

I've played a lot of these over the last year. 60% isn't a hard score to get if you aren't distracted or tired, but a reasonable player should get 65% regularly. I've been over 70 a lot and cracked 80 once (but it included a very lucky board where I mismoused and passed stayman, which was about the only making bid).

The knowledge that another hand has no more hcp than you is huge. Treat a 10 or 11 point hand in 1/2 like a 3rd seat opener. You can open a 4 card major and pass the response if you like. You can make a heavy weak 2 if you like it won't go overboard.

GIB's signaling is weird if you aren't used to it. It seems like it starts count and then switches to reverse SP mid hand.

My experience with GIBs declarer play has not been positive. I will try to hog the hands when I see an edge, but I don't go overboard with it. Think of it like playing with a weak, but not hopeless client. I have seen people do crazy things to avoid let GIB play the hand, like opening 2N with a void.

I do stretch the NT ranges some. My 1N is a fair 14 to bad 18, however if I have spades, I won't upgrade 14's as much since I can rebid and play the hand. Gavin responded to a comment I made on bridgewinners.com that he wouldn't open 1N with a hand like AKx KQ AJx J98xx, since he likes to stay with the field in these things, but I think thats wrong. A lot of people are opening 1N, and my expectation when GIB plays the hand is about 40-45%, which means you are losing ground in these tourneys.

I never stretch to open 2N. I saw jdonn's comment earlier in this thread and I agree that its bad policy with GIB. It does push to 3N with some crappy 4 counts. I will open 22's 2N for the same reason.

I haven't noticed that GIB adjusts its game for matchpoints at all. It doesn't compete enough and never makes profitable doubles that I have seen. However, it seems to take a lot of low-level doubles as competitive so you can made cooperative doubles and not expect to get hung.

Bidding after a 2 opening is very problematic. I have seen it bid random 4 card suits for no real reason.

GIB evaluates well after a strong jump shift. You them.
GIB loves to make short suit leads against NT, like Gnome. Sometimes they don't work out, but if you see a medium spot lead, there's a good chance its a short suit. I haven't figured out a way to exploit this, but maybe there is.
Winner - BBO Challenge bracket #6 - February, 2017.
0

#109 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2011-February-01, 22:59

Don't bother bidding Jacoby 2NT unless you have at least 16 or 17 HCP. If you have 13-15 HCP in a Best Hand tourney slam is unlikely. Just jump straight to game.

In 3rd/4th seat, if you have opening values and a good 6-card major, open 4 of the major.

If GIB opens a weak 2 in a major, try to find an excuse to raise it to game.

#110 User is offline   newbag 

  • Pip
  • Group: Members
  • Posts: 2
  • Joined: 2011-March-30

Posted 2011-March-30, 23:16

"If you are leading, and you have a big minus score coming, just stop playing (Provided that there is less than 4 min left...)"

Do we know for sure that the robot does not evaluate/adjudicate unfinished hands? I have followed the above practice, before I ever read this post, because it seemed like the self-interested thing to do. I'm just wondering if the robot estimates how the hand would/should have ended up. It seems to post the leaderboard very quickly after the tourney is over, which makes me think it doesn't evaluate unfinished hands. But then again, computers are fast. Anyone know?
0

#111 User is offline   chasetb 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 851
  • Joined: 2009-December-20
  • Gender:Male
  • Location:Podunk, backwater USA

Posted 2011-March-30, 23:48

I have heard if there are 7 or less tricks played, the hand will not be simmed, if there are 8 or more tricks played, it will be. I know from experience that after 9 tricks are finished it will finish with robots.
"It's not enough to win the tricks that belong to you. Try also for some that belong to the opponents."

"Learn from the mistakes of others. You won't live long enough to make them all yourself."

"One advantage of bad bidding is that you get practice at playing atrocious contracts."

-Alfred Sheinwold
0

#112 User is offline   diana_eva 

  • PipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 3,934
  • Joined: 2009-July-26
  • Gender:Female
  • Location:bucharest / romania

Posted 2011-March-31, 02:00

View Postchasetb, on 2011-March-30, 23:48, said:

I have heard if there are 7 or less tricks played, the hand will not be simmed, if there are 8 or more tricks played, it will be. I know from experience that after 9 tricks are finished it will finish with robots.


This does not apply in robot race. Only in Express some of the hands are auto-adjusted if played enough. In robot race tourney if you stop playing nothing happens. You have 25 minutes to play as many hands as you can and get as many points as you can (total points, not duplicate).

#113 User is offline   palabreur 

  • PipPip
  • Group: Members
  • Posts: 30
  • Joined: 2011-March-09

Posted 2011-April-01, 03:10

View Postpretender, on 2011-January-23, 19:07, said:

I would say that I'm fairly good at these best hand robot races.

Here's one strategy I'll share although I think most of you probably have come up with it yourselves already:

When partner bids something and RHO doubles, REDOUBLE!

It's amazing the number of times you get to play it there, often with overtricks. And if not, you just place the contract as you originally would. GIB N won't do anything stupid, and it's not like you needed its cooperation in the first place.


Hmm... I wonder if the logic has changed, or if I'm just using it in a different situation to you. I've tried this probably ten times after center-hand opponent preempts, but never have I had them pass out the redouble.

[EDIT] And of course, the VERY next time I try it, it works! :) Different situation to what I had tried before:

(1) 2 (dbl) redbl all pass

with a single overtrick - game wasn't making, but +840 helped propel me to the win. :)

[EDIT 2] Ditto the NEXT try: (1) 2 (dbl) redbl all pass

So I guess maybe using it after partner makes a weak jump overcall is the trick.
0

#114 User is offline   croquetfan 

  • PipPip
  • Group: Members
  • Posts: 28
  • Joined: 2010-October-29

Posted 2011-April-13, 22:25

I think it's unfair that internet connection speed has an influence. Even though I am using ADSL, I have seen players posting a gams or slam result before my GIB has made a bid.
Surely there is some way of evening out the speed to level the playing field e.g. a timer that makes all bids take 5 seconds minimum and all plays
1 second. Maybe you won't get in 23 hands anymore and there won't be scores of 10000 but it would be fun.
0

#115 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2011-April-14, 19:28

I don't think much Internet bandwidth is used when playing a hand, so connection speed shouldn't be that significant.

#116 User is offline   palabreur 

  • PipPip
  • Group: Members
  • Posts: 30
  • Joined: 2011-March-09

Posted 2011-April-15, 10:15

View Postcroquetfan, on 2011-April-13, 22:25, said:

I think it's unfair that internet connection speed has an influence. Even though I am using ADSL, I have seen players posting a gams or slam result before my GIB has made a bid.
Surely there is some way of evening out the speed to level the playing field e.g. a timer that makes all bids take 5 seconds minimum and all plays
1 second. Maybe you won't get in 23 hands anymore and there won't be scores of 10000 but it would be fun.


As barmar points out, it's not internet connection speed that's the problem; it's lag. I've had bad lag nights where I'll sit there waiting for 30 seconds, and it's frustrating, but slowing everyone else down doesn't solve that frustration. Just don't play if you're lagging. There's so little information being sent, that "connection speed" can't be the problem.

LOL @ "23 hands"... a normal good round for me is 50 with 15 passouts.
0

#117 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2011-April-16, 21:31

I've played plenty of robot races, I still don't see how you can play 35 hands in 25 minutes -- I've never had more than high 20's including passouts. 35 hands is 42 seconds per board. That's 3 seconds to bid the hand and then 3 seconds for each trick. Do you actually have time to think about what you're doing? Do you actually make most of your contracts?

For a little while I tried a strategy where I never opened, I just waited for North to open and then tried to make it declarer as much as possible, figuring that it could declare faster than me. But it didn't seem to be any more successful than my previous strategy, and it certainly wasn't as much fun.

#118 User is offline   palabreur 

  • PipPip
  • Group: Members
  • Posts: 30
  • Joined: 2011-March-09

Posted 2011-April-17, 06:49

View Postbarmar, on 2011-April-16, 21:31, said:

I've played plenty of robot races, I still don't see how you can play 35 hands in 25 minutes -- I've never had more than high 20's including passouts. 35 hands is 42 seconds per board. That's 3 seconds to bid the hand and then 3 seconds for each trick. Do you actually have time to think about what you're doing? Do you actually make most of your contracts?


I looked back at my last 30 tournaments; over those, my average is 6630 with a standard deviation of 2667; so 95% of the time, my scores are between... 1300 and 12000, or so.

My average for completed tournaments is a little higher. Occasionally when things are not going well, I withdraw from the tournament and start another, if I don't expect to have a shot at winning.

Do I "make most of my contracts"? That's a red herring, since making contracts is not the name of the game, it's achieving the highest TP score possible.

I'm not sure why 42 seconds per completed hand is unbelievable. Many of the hands essentially "play themselves"; anything that takes effort, I just guess. Should I play for the finesse or the drop? If it's close, who cares? Plus, the part scores are the hands that take the longest, either for me to play or (especially) the computer. The more hands you pass out, the more higher level contracts you'll play, and the faster things will go.
0

#119 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2011-April-17, 21:03

There's no arguing with success. My average is 3510, and my highest recent score is 8530; I don't know my std.dev., but just throwing out the extremes has me between -820 and 7410 95% of the time. Your average is about the same as my maximu.

Yet I've come in 1st with scores ranging from 4760 to 8530. But I've only made a profit on 15 of them, and overall I'm in the hole (I really shouldn't bother with the $5 games I occasionally enter -- I've never won anything in any of them).

#120 User is offline   barmar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Admin
  • Posts: 16,442
  • Joined: 2004-August-21
  • Gender:Male

Posted 2011-April-18, 19:10

I played a couple of races as fast as I could after posting that, and I still only managed about 30 hands (including passouts).

I wonder: are the people who are good at these races also good at video games? It seems like you need lightning fast reflexes just to be able to play the cards that quickly, and that's similar to video game skills. Even at the end of a hand, when I'm cashing top tricks, cross-ruffing, or giving up all my remaining losers (i.e. situations where I would have claimed or conceded IRL) it takes 1-2 second just to click on the cards.

I wish I could watch one of you play a hand, so I could see your mouse flying back and forth.

I know about enabling "auto-play singletons", but that only helps on a few tricks; I try to arrange to make this apply as much as possible, but that often requires a little extra thinking to play tricks in the right order, so it may just break even.

It obviously hurts that I still feel the need to check the robot's bid descriptions occasionally. Too many times I've been screwed by forgetting that 1M-1NT-2NT-3X is a transfer.

When you're declaring or defending a part-score, do you just play without hardly thinking, since it makes so little difference to the total score?

Share this topic:


  • 8 Pages +
  • « First
  • 4
  • 5
  • 6
  • 7
  • 8
  • You cannot start a new topic
  • You cannot reply to this topic

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users