It's a brave individual that lays out system rules for me to pick apart
Ok, so let's break this one down...
Rule 1: Not bad. It implies that you are looking for value beyond the statistical probability of a horse winning. Not sure if you have any facts and figures for how effective that one is but I would suggest that you would need to be exceedingly selective because the plain fact is that most of the time a horse is at longer odds because it deserves to be.
Rule 2: If I were to suggest that this rule was extraordinarily vague I might be understating matters a little. What criteria are you using to determine a horse's preferred "conditions" (whatever that means)? For a start you might as well add in a rule that says eliminate all horses younger than 4 because it is highly unlikely that any horse younger than that and of any quality would have exposed enough form to make it possible to make that judgement. In all likelihood, you are referring to distance, track and weather conditions but presumably you do not have a clause for class or any such subjective assessment. Regardless, this rule is subjective at best and untenable at worst and needs to be significantly embellished.
Rule 3: As if rule 2 wasn't ambiguous enough, this one is basically impossible to determine unless you are The Horse Whisperer. Seriously, how on Earth are you even going to pin down criteria for this assessment, much less make it in a reasonable amount of time. The point of a racing system is that it streamlines the selection process so that you do not spend hour upon hour pawing over endless reems of data. This rule in and of itself is likely to kill off that particular outcome right off the bat.
Rule 4: "It's last few starts". How many is a few? Do any of the starts prior to a spell count? What if the previous start preceeds a let up but not a spell? Does a tightening of the margin from a half head to a short half head count as discriminantly closer?
Rule 5: What is a favourable barrier? Down the Flemington straight that can vary from meeting to meeting. On certain rain affected days at Caulfield there are sometimes lanes that favour horses drawn out wide, even though they will cover significantly more ground. The Sandown chute can often favour the wide barriers when there is significant speed on the outside. The Melbourne Cup start at Flemington heavily favours far outside barriers in big fields. There are so many vagaries that you need to be far more explicit for this rule to carry any kind of usefulness. May I suggest "no wider than barrier 11 after scratchings" and leave it at that. Ignore the oddities in favour of something more generic and practicable.
I have seen more systems over the years than I care to remember. Nearly all of them claimed to have a great testing period before they were "unleashed" and failed. The reason systems mostly fail is because they are often designed to capture a particular winner or set of winners ie. they are written with the benefit of hindsight. Quality systems are designed and the rules put in place prior to any results testing. They are dry tested after the development and once they can sustain a period of success beyond 12 months only then should they be employed with real cash.
I have several systems in my possession that have met all of the criteria above. Each of them has had losing years but have mostly stood the test of time. There is no pot of gold at the end of the rainbow. Steady profits will always outstrip flash in the pan systems that eventually fall on their systematic butts.
Having said all that, I hope you prove me wrong and make your fortune. It's no skin off my nose if you do.
Cheers
CS