The system, located at a, was created by hermes on 10/9/2002. The rules are as follows:

Rule 1: Races of 14 or less runners only.
Rule 2: Country races weekdays preferred.
Rule 3: Good or fast tracks only.
Rule 4: Only consider TABs 2-3, 4-5 and 6-7. Don't replace scratchings.
Rule 5: Look at the QTAB rating of the favourite (or the runner you believe will start favourite at jump). Now search the TAB 2-7 range for two runners with ratings directly below the favourite. For example, if the fave is rated 100 look for 99, then 98, then 97 and so on.
Rule 6: Maximum of two selections.
Rule 7: If all runners 2-7 are rated above the favourite, no bet.

hermes comments: "This system is a direct spin-off from some recent posts, mainly that of Bhagwan. It works OK on paper.

"Often, where the favourite is not the highest rated runner, there is only one selection.

"I've tested this over 120-plus races for a profit on turnover of 7.8 per cent. Promising. What is interesting is that it works better than looking at runners rated better than the favourite.

"As a general observation, with these sort of systems you need filters that pull both ways - one up and one down. By only looking at TAB numbers two to seven you already have a filter that places you in the statistically rich band of starters. To add to this another filter that looks for quality (those rated above the favourite) won't do as well as one that filters out some quality (those rated below the favourite). Thus in Bhagwan's original post, he filtered out last start placegetters - ostensibly good runners. I think that is usually the right type of move - one filter up (only TAB numbers two to seven), one filter down (only last start failures)."

The staking plan he recommends is as per Bhagwan's sensational bad horse method. That is, if TAB number two or three, bet four units; if TAB number four or five, bet three units; if TAB number six or seven, bet two units.