CP - Past Form - Find this one of the most important factors, and essential at all distances, but moreso down at the shorter race end.

CF - Current Form - Often a disaster from 1000m-1200m, and perhaps out as far as 1400m. If used, in a race with first uppers, the neural system tends to give them no hope. Yet first uppers win between 19% and 33% of all races, depending on the distance.

(Both CP and CF, need consideration race by race. As an example, where there are lots of first uppers in the race, you may use something like CP3/CF0, perhaps CP3/CF1, then as distances increase, and first uppers go missing, to CP2/CF2, and perhaps from 1500m or 1600m onwards, to CP2/CF3).

TIM - Time - I personally find TIM provides more distortions, than it is of help. It will find the occasion long priced winner for sure, but I leave TIM on 0 at all distances.

(I may be wrong, but I do not understand how the neurals can differentiate between times from different tracks; generally there is about two seconds difference between track records at Flemington and Moonee Valley. Hawkesbury is one second faster than Canterbury, and so on....)
I also set my neural factors down the lower end of the points scale, and TIM1 often tends to over-power other points allocations.

JA - Jockey - One of the most important. Go back over past results, and you will find JA either on top, or near top for most winners, all distances.

TA - Trainer - No among the winners nealy as often as JA (or JT) but important enough to give it more than a zero.

JT - Jockey/Trainer combo - seems to fit somewhere between JA and TA - definitely has its moments.

BP - A medium range performer. Go back over past results and you will find it valuable one race, not worth a bumper the next. , don't set too high.

(And beware of BP on bush tracks. Watch what is actually coming up on the BP points allocations. The neural computer takes winning barrier statistics far too literally. You may find barrier 8 with 20 points, barrier 10 with 20 points, yet barrier 9 with no points - because barrier 9 has not won a race at that distance at that track. Totally illogical - then what happens when the runner in barrier 7 is scratched!) So beware of BP. Sometimes you may be better without it.

CRS - Course - Have found this to be of more importance in races of 1400m and over, and use it accordingly. Not sure why, but when I have checked past results, it has panned out that way. 1400m/1600m seem particularly relevant distances to apply CRS.

WT - Wet Track - I tend to use WT very sparingly, and generally only when tracks have been on the good side for a some time, then become rain affected. In mid winter in Melbourne for instance, doubt WT adds anything to neural accuracy. But a WT1 or WT2 probably doesn't do much harm either.

Think about how you will use WT, on the day the wet track happens.

DIS - distance - Checking past results, distance has never stood out as a major neural factor. I use DIS1, but at around 1400m, I tend to increase it to DIS2. There is a feeling that 1400m, is perhaps more of a specialist distance than some others. Feel that DIS, set low, doesn't impact overall results, one way or the other.

$ - prizemoney - In the early days, I always set $ very high, but have now dropped it back somewhat. Definitely doesn't feature in the profile of winners, as often as CP or JA, for instance, but remains important.

DLR - Days Since Last Raced - this one is important - to leave out of contention in all races up to around 1450m. DLR anything, tends to send first uppers to the sin bin. From 1500m and beyond, yes, it becomes important. However, you will find the DLR points allocations, across the scale 1-5, are so tightly grouped, that you need to actually look at the figures in isolation, to get value from them. (for example, should I dismiss all runners, which have a DLR allocation, of say 14 points, on my setting).

The neurals are obviously applied differently by everyone - there are so many combinations - then to use the information to advantage.

The neurals do find winners at longer prices, and from that point alone, perhaps have their advantage and place, over the pre-post or starting price markets.

In fields of 8-13, for instance, my neural setting can find the winners of 2 races in every 3, in the top four selections. (At certain distances, it is slightly better). The pre-post market can do much the same thing. But rarely will any market order, snare that $20, $30 or $40 winner, in the top two or three selections.

Taking the neural selection order to the next stage, and using it to a profitable advantage, is another matter altogether.

The neurals, are not the be all and end all to selecting winners, but they are different, they are fun, and they are enormously accurate for a computerised selection method.

They are an interesting challenge. And they are FREE

Wesmip1 has provided some very good information. I also like Trained at Track and only horse in the race to have won, Speed Maps and horse settling important and days last start