My Robot Strategy What to do and not to do
#1
Posted 2020-October-18, 11:24
I watch the score like a hawk. When playing IMPs I can make up the score if doing poorly or make a big mistake. I find that IMPs are the easiest to win.
The key to the robots is to get to decent contract and hope that you play the cards well. And oh, I deceive the robots often and make them change their strategy.
MY FINDINGS:
If I open a weak 2, the robots will almost always pass.
Don't open 1NT fourth seat; I find the robot devalues your points. I have seen it pass with 8 points and a good 4-card spade suit: Game was cold.
UGH: I can't figure out how to get robots to bid those long minor suits. They always go for majors even with a strong rebiddable minor or even the other major???? Which hints here please would be welcome.
Don't raise from 2 c to 5 c or 3c to 5c, the robot will almost always bid 1 higher.
I am finding that double for take out is a mistake. If the robot bids Hearts for example, it will think that's the suit for ever an eternity. Even if I bid and rebid and rebid my long minor, he takes me to 7 Hearts with just 5 hearts!!! This happens to be about once in 10 sessions.
I agree if the robot opens say 1 Spades and I have 3 Spades and 12 points, I am now jumping to game.
I find that the ACE of the opponents is always to the left of the shorter part of the suit, say Q10987 vs K2, lead towards the KING and the ACE is almost always to the left.
I find robots don't often declare well except for NT hands. I am working on taking that privilege away from the robots.
#3
Posted 2020-October-18, 14:38
stephenpag, on 2020-October-18, 11:24, said:
The ACBL has held online regional/national events where you can earn color master points. Check their website for the next dates
2020 Online Schedule
#4
Posted 2020-October-18, 19:54
stephenpag, on 2020-October-18, 11:24, said:
What seat you're in is totally irrelevant here. GIB will pass flat 8 counts in any seat, but analysis tends to suggest passing is the percentage option, even if you miss a game here or there.
stephenpag, on 2020-October-18, 11:24, said:
I'm not sure why you'd jump to game, but of course you'd force to game; who wouldn't?
stephenpag, on 2020-October-18, 11:24, said:
This is just nonsense and has been disproven numerous times.
#6
Posted 2020-October-18, 21:08
johnu, on 2020-October-18, 20:33, said:
I have some spare Aces that I can lend you for an afternoon.
#7
Posted 2020-October-21, 14:43
Do double their cold 3N, if they have anywhere at all to run. They *will* run.
#8
Posted 2020-October-21, 14:46
Bid game far more aggressively than normal when you hold a long suit.
Their insanely passive leads tend to make this a winning action.
The robots almost always switch suits. If you let them hold
the first trick, they seem to switch about 80% of the time.
#10
Posted 2020-October-22, 15:51
wbartley, on 2020-October-22, 14:34, said:
1NT is the best bid in the world. But what was your score? How many times have you done it? What is your average result doing it? For the record, I agree with you!
#11
Posted 2020-October-22, 17:00
pilowsky, on 2020-October-22, 15:51, said:
I got a zero on this hand. Everyone else was in 3NT making an overtrick. So, this was a little bit better result than average for me. Still, I forge on. Fools rush in, as they say.
I've taken to opening 1NT with any hand with 15 points or less in the best hand format, practically without regard to distribution (haven't been doing it with a void yet). When the robot bids Stayman, I invariably reply with 2D, even if I've opened 1NT with a six card major. I accept major suit transfers even if robot is transferring to my singleton. With 18-21 points I open 2NT, as do many other robot gamers. I used to open a minor and jump to 2NT with 16-17, but lately I've taken to opening a minor and jumping to 3NT and anecdotally it's working out much better. I make no claims about the effectiveness of this strategy compared with a more mainstream approach but it's amusing and I would guess I average above 60% playing with three robots using this strategy. I'm a halfway decent player so it might just be a wash.
I can confirm some of what other people have said on this thread
- Allowing the robot to hold the trick is just about the best thing you can do. They invariably shift for the worse. I would have loved to let the robot hold trick one on my example hand. Alas, I couldn't get under RHO's third hand low.
- Robots are only good declarers in comparison to their ability to defend.
- Robots run when you double them in 3NT where their side has bid any suit.
- Don't worry about having stoppers in unbid suits when bidding NT. Robots almost always make a stupid passive lead and if they do happen to lead your unstopped suit see item 1.
- Doubling for penalty with a robot partner is fraught with danger. The opponents may have bid four suits on their way to a tenuous major suit game and woe to the player who tries to collect a nice penalty.
Here are some of my other observations
- When you do make these off-shape notrump openings and the opponent bots enter the auction, bid your suit, even at the three level. I find the bots bid again in these situations when they should defend and they defend when they should bid on.
- It only makes sense that since the bots defend so badly, you should avoid defending with a robot partner. I probably bid one more twice as frequently with a robot partner as I would with a decent human partner and it's increasing by the day.
- There's no point in making a splinter bid with a robot partner. It seems to have no influence over its hand evaluation.
- A corollary to jumping to game with support for robot partner's opening 1M with an opening hand: In best hand tournaments, opening four of a major with normal, minimum range opening hands with a six or more card suit pays off. You'll rarely miss a slam and the robots will often surrender multiple tricks by defending too actively.
- No offense to goulash aficionados but goulash is the fast track to dementia.
I guess that last observation was a bit of a non-sequitur. Anyway, no regrets! I stand by it.
#12
Posted 2020-October-22, 18:41
I've been doing it on anything from 15-18 with any distribution with excellent results - check myhands! Doesn't work every time - often due to my poor play but I sure get a lot of makeable contracts.
#13
Posted 2020-October-22, 19:28
smerriman, on 2020-October-18, 19:54, said:
If it's "best hand" format and you hold 12 points, partner won't have more than 12, so maybe just leaping to game to avoid information leakage is reasonable.
#14
Posted 2020-October-22, 19:47