General Report on Tunny
25B Page 158
sign incorrectly adjusted, so that it actually increases the score.
Accordingly a table is used for the ratio
| q = |  |
so that decibans per pip = 
The table, whose construction is explained in 25X (e), is
 |
.798 | .9 | 1.0 | 1.1 | 1.2 | 1.3 | 1.4 | 1.5 | 1.6 | 1.7 | 1.8 | 1.9 | 2.0 | 2.1 |
| q | 0 | .57 | .72 | .82 | .87 | .91 | .93 | .95 | .96 | .97 | .98 | .98 | .99 | 1.00 |
Crude decibanning is the result of taking q = 1.
A different rule is required for wheels obtained from a rectangle. (24X(e) IV)
(d) Decibanning a run on the wheel from another run.
Suppose, now, that a wheel has been obtained from a single run, say ΔΧ3 from /'s and that additional evidence is desired, say from 5's. The crude decibanning formula can be use x being the score on the wheel obtained from /'s, for although this wheel is not necessarily the correct wheel, its wrong signs are unrelated to the scores for 5's, and there is no reason to suppose that the wrong signs will enhance the score.
The same remark applies to runs on another message set on these wheels, even runs for the letters used, on the first message, to make the wheel.
Moreover it is unnecessary to apply a significance test, the significance of the original run being sufficient.
The statements in this paragraph are not strictly exact (25W(f)).
(e) Decibanning from a letter count.
When a previous approximation to a wheel is available, the usual method of decibanning for the next runs is to make a letter count (4, 8, 16 or 32) on the previous approximation. This is compact and helps one to see beforehand which runs are likely to be likely worth while (allowance being made for general fish characteristics): moreover runs having approximately the same decibanage per pip can be run together.
These runs, which may be numerous, will have contributed in varying degrees to the previous wheel. A table generalising that in para(c), would be unwieldy: moreover it would necessarily use only the evidence of such messages as were set on these wheels. In fact general fish evidence cannot
Back to General Report on Tunny. Contents.