I have tried the 1/4 approach and used it for some time until I discovered the following.
Based on a horse paying $5 win $2 place

Horse wins horse places
1 win = $5 return = 400% POT $0 100% loss
1 e/w = $7 = 350% $2 0%
1/2 = $9 200% $4 33%POT
1/3 = $11 175% $6 50%
1/4 = $13 160% $8 60%
1/5 = $15 150% $10 66%
5 plce= $10 100% $10 100%

Based on the above versus risk on outlay the only bets that provide a decent POT are the win bet or the place bet. Any e/w bet cuts into the POT too much with the exception of the $1 e/w which provides a get out should the horse place and doesn't erode the POT too much should it win. I prefer the dutch bet for the saver rather than a $e/w.

ANY ONE WITH A METHOD DESIGNED ON AN E/W BET IS ERODING THEIR POT.

I now play 3 methods.

1/ Place only. The bet is 1/4 of the bank at all times up to my limit. Method of selection is simple. As 70% of winners pay under $12 I wait til the market is up and select the runner I feel has the best credentials of those in the market. Runner must be paying $1.70 or more.

2/ Win only. The bet is 10% of the bank. Method of selection is again simple. Wait til the market is up, see who the favourate is and if I feel he has winning credentials bet him. Must pay $3 or more. Theory is that 40% of favourites win.

3/ Dutch betting. No more than 4 runners (preferably 3) Study required here to reduce the chances and I prefer to stick to certain regions on this one. One of the selections is the main bet with the others savers. Hard to use if a selection is at odds under $3.

Comments appreciated, I await to be shot down in flames.