To what extent to in­form ?

PAD

At any moment of the auction part­ner makes provi­sions for many allowable (more or less devi­ating from the model) holdings you can have. Thus, he knows that as­sum­ing an aver­age (model) hand he makes an error – the greater, the more your hand devi­ates from the model.

A rough estimate of this uncer­tainty may be ex­pressed by an av­er­age er­ror.

The greater the value of this error, the greater un­cer­tainty.

A simpli­fied ver­sion of cal­culating this aver­age er­ror has been given by me in Pik­ier #1, in Distribu­tional Uni­formity Principle. he prob­lem is how to cal­cu­late the aver­age er­ror. The is­sue is com­plex, and I will have to return to it later.  Any­way, it seems neces­sary to treat the hand as a point in a four–di­mensional space (each suit is equiva­lent to one dimen­sion), and intro­duce a met­ric, taking into ac­count both length and honors (plus prior­ity of the ma­jor suits). How­ever, in­tuitive „feeling” of the av­er­age error is fully satis­fac­tory for our pre­sent needs.

The aver­age error is, in fact, equivalent to AVERAGE DEVIATION from the model. The higher this de­via­tion, the larger the disper­sion of possi­ble hands around the aver­age. There­fore, from now on, we will replace (for sim­plicity but without consider­able dam­age) „aver­age error” with DISPERSION.

 

It is obvi­ous that the general pur­pose of transmit­ting informa­tion is to mini­mize dis­persion (which is a measure of lack of in­formation). At the same time it is clear that dis­per­sion can­not be totally elimi­nated (eg be­cause of SAD), which leads us to the fol­low­ing prob­lem:

 

What level of dis­per­sion may be consid­ered (in a given situation) ade­quately low, and how is part­ner to signal  his desires in this re­spect ?

Look at a gener­ally true for­mula:     loss = (strength) x (dis­per­sion)

Therefore, to mini­mize the loss (in points) it is neces­sary to mini­mize dis­per­sion, and – the more so, the greater the strength.  Hence:

       PAD        Pre­ci­sion Ad­just­ment Di­rec­tive

 

The more intensive the in­forma­tion ex­change –

 the greater the strength or dis­persion

with the word „or” not ex­cluding „and” (obvi­ous).

 

How can you rec­ognize that part­ner is seek­ing wider in­for­mation ex­change  ?

There are three symptoms, named: Firstly, Sec­ondly, Thirdly...

FIRSTLY

Partner tries to keep the bidding as low as pos­si­ble.

 

Thus:

 

PAD-1

The lower the bid – the greater the strength or dis­per­sion

 

and, in a more prac­tical for­mula­tion:

 

PAD-1a

The more your hand dif­fers from the model 

– the lower should the bidding be kept

 

The objec­tion that this simpli­fies mat­ters for oppo­nents can be counterat­tacked with a state­ment that: The more they bid – the greater the knowl­edge ob­tained and therefore the lower the dis­per­sion. Thus, sometimes (es­pecially with strong hands) it might be profit­able to allow op­ponents to bid (and to listen what they bid!).

 

SECONDLY

Partner se­lects bids af­ter which the chance of your pass

(ie stop­ping ex­change of informa­tion) is smaller.

 

Let's con­sider, from that view­point, three types of bids:

new suit    no trump     old suit

where:

new suit = a suit not bid by the part­nership

old suit = a suit al­ready bid by the partnership

Note that „old suit” is the most likely bid to be­come a final con­tract (and to stop ex­change of in­for­ma­tion), and „no trump” is more likely to end the bidding than „new suit”.

Thus, the second general subdirec­tive:

 

PAD-2

The newer the suit – the greater the strength

 

which may be di­vided into the fol­lowing state­ments:

 

PAD-2a

New suit is a positive sug­gestion (strongness)

 

 

PAD -2b

Old suit is a negative sug­gestion (weakness)

 

PAD -2c

No Trump is a neutral suggestion

This ex­plains why in natural systems new suit is usu­ally forcing.

Until now, I have been justifying Sub­direc­tives #2 stating that „new suit” car­ries more infor­ma­tion (ie signals less prob­able event) than „no trump” which, in turn, car­ries more in­for­mation than „old suit”. Intui­tively, this ex­pla­nation seems to hold but I am un­able to give a com­plete rea­soning (maybe because I don't know how to meas­ure devia­tions).

 

To illus­trate PAD –  the same ex­emplary situation:

W

E

 

1¨

1

 

          ?

The strength of bids from the view­point of:

HT = Honor Tricks

 

 

Therefore, the final ( SAD & PAD ) se­quence of bids ac­cording to their strength is:

1NT  2¨  2§  2  2©  2NT

and – as you see – is just the same as com­monly used !

 

 

 

SAD

PAD

SAD + PAD

1NT

=

3.5  HT

 

3.5

2§

=

3.9

+ 0.4

4.3

2¨

=

4.3

– 0.4

3.9

2©

=

4.7

+ 0.4

5.1

2

=

5.1

– 0.4

4.7

2NT

=

5.5  HT

 

5.5

 

THIRDLY

Partner uses special ways of bidding, fa­vor­ing the ex­change of in­formation

( forcing bids, cue-bids, con­ventions, gadg­ets), or stresses his strength

 

Now, as we know how partner ex­presses his de­sire to ex­change more in­formation, we have to think of how we will meet his wish.

Partner is signal­ing that, from his point of view, dis­persion is still too great ! 

Thus, it should be reduced, ie:

 

PAD-3

The more partner stresses
his need to exchange informa­tion, or his over­all strength –
– the smaller Positive Devia­tions are al­lowed
to be con­sidered  nonzero de­viations

 

Thus, we move gradually from typi­cal stan­dard PoDs in the initial phase of the auc­tion to showing smaller and smaller PoDs, go­ing down into such de­tails as confirming the king in a suit in which the model value has been a queen.

Precision Adjust­ment Directive ( PAD ), to­gether with its SubDi­rec­tives, controls this proc­ess.

Here I end the first part of NASA and invite the reader not to waste time

All that was said (as it stands now) is al­ready fully applica­ble !

 

 

 

 

Next

 

Pikier Writings

 

Content