They strive to be sensible, not brilliant. (Part 12)
"I PLAY THE HOFFMEISTER NO TRUMP"
So proclaimed the plastic badge sported by P.J. Hoffmeister, a character in Richard Powell’s 1968 novel Tickets to the Devil, set at a national tournament in Miami Beach. P.J., who has all of 53 masterpoints, believes he’ll achieve bridge immortality if he can convince the masses to play his system.
Many players enjoy making up their own systems and conventions. Unlike P.J., few are hoping for fame. Most just want to improve their systems and experiment with new ideas.
Some “mad scientists” end up inventing new systems that look perfect on paper, but don’t translate into good results at the table. The pitfalls are demonstrated by some of the failed ideas I’ve tried over the years.
Brilliant Idea #1: “Unusual Michaels”
The proposal: After the opponents bid two suits (1C – Pass – 1H, for example), the fourth-seat bidder has five ways to show two-suited hands. Double and 1NT are sound and light takeouts (at least 4-4 in the unbid suits). Longer suits are described by cuebidding opener’s suit (5-5), cuebidding responder’s suit (5-6) or jumping to 2NT (6-5).
The most valuable system additions solve problems or correct deficiencies, rather than merely enhance your current methods. The ideal target for a new use is an “idle” bid that doesn’t already have a meaning in the auction you’re trying to improve.
Not all of these bids are idle, and that’s the downfall of this gadget. It offers pinpoint accuracy if you hold a two-suiter, but muzzles you when you’re dealt a strong notrump or a natural spade overcall. This fine-tuning for two-suited hands isn’t worth the sacrifice of two natural bids.
Brilliant Idea #2: Four-way Bergen Raises
The proposal: After partner opens 1 of a major, you retain the original Bergen meanings for 3D (4-card limit raise) and the jump raise (preemptive). The 3C response is modified to show either a 4-card constructive raise (8-10 support points) or a 3-card limit raise. Opener bids 3D to ask, and responder retreats to 3 of the major with the weaker raise. Any other bid shows the limit raise.
The 3-card limit raise is normally shown by responding a forcing 1NT, then jumping in opener’s major. It’s difficult to describe these values if opener jump-shifts over 1NT, so the two-way 3C was devised to allow responder to show support with his first bid.
That’s a worthwhile benefit, but there are risks in assigning multiple meanings to one bid. Many ingenious new conventions have been tunneled by the pesky opponents, who won’t always allow you to complete the asking sequence. After 1H by you, 3C by partner, how do you determine partner’s hand type if your opponent overcalls? You can dream up ways to try, but they involve tradeoffs (giving up a penalty double, for example) and the potential to get too high when partner has the weaker hand.
Brilliant idea #3: “The Nothing 2D”
The proposal: In a 2-over-1 system, opener’s rebid of 2D over a 2C response is an artificial catchall that denies extra length in his major and adequate stoppers for a 2NT rebid.
In most 2-over-1 systems, opener shows this hand by rebidding his major, which doesn’t promise extra length. Follow-up bids will usually clarify opener’s “real” length, but these auctions can be awkward. The benefit of the Nothing 2D is that opener guarantees at least six cards when he rebids his major. If he rebids 2D, you have maximum room to investigate other games.
This seems perfect, but at-the-table testing has uncovered its weaknesses. Using a bid to deny a holding means it’s not available to confirm one, and giving up the natural 2D has proved costly.
We’ve had three Nothing 2D auctions since adding it to our system. On the first, opener held four diamonds, but since his 2D rebid was artificial, we had to search for stoppers in both “unbid” suits. It was a painful auction that ended in 3NT from the wrong side. On the second, we missed a good 6D because opener couldn’t show 5-card diamond length. Our third use of the 2D rebid elicited a lead-directing double and we scored one fewer trick in 3NT than the field.
At best, the Nothing 2D needs further evaluation. These results suggest that it trades one problem for another (or three), and perhaps the original “problem” wasn’t serious enough to warrant a fix. We’ll look at ways to evaluate and test bidding inventions in the next installment.
© 2008 Karen Walker