he uses 100+ charts to explain and demonstrate how he does the position sizing.
in <Reminiscences of a stock Operator>, Livermore explains his sizing operation also in details as follows:
Let us suppose, for example, that I am buying some stock. I'll buy two thousand shares at 110. If the stock goes
up to 111 after I buy it I am, at least temporarily, right in my operation, because it is a point higher; it shows me a profit. Well, because I am right I go in and buy another two thousand shares. If the market is still rising I buy a third lot of two thousand shares. Say the price goes up to 114. I think it is enough for the time being. I now have a trading basis to work from. I am long six thousand shares at an average of 111.75, and the stock is selling at 114. I won't buy any more just then. I wait and see. I figure that at some stage of the rise there is going to be a reaction. I want to see how the market takes care of itself after that reaction. It will probably react to where I got my third lot. Say that after going higher it falls back to 112.25, and then rallies. Well, just as it goes back to 113.75 I shoot an order to buy four thousand--at the market of course. Well, if I get that four thousand at 113.75 I know something is wrong and I’ll give a testing order--that is, I'll sell one thousand shares to see how the market takes it. But suppose that of the order to buy the four thousand shares that I put in when the price was 113.75 I get two thousand at 114 and five hundred at 114.25 and the rest on the way up so that for the last five hundred I pay 115.5. Then I know I am right. It is the way I get the four thousand shares that tells me whether I am right in buying that particular stock at that particular time--_for of course I am working on the assumption that I have checked up general conditions pretty well and they are bullish. I never want to buy stocks too cheap or too easily.
Both of them suggested a similar sound mathematical algorithm to size up the position which I 100% agree. what confuses me is "under what circumstances a trader should cut loss if he finally accumulates a full size swing", both of them didn't explain it and I kind of lost there.... At the moment I can only assume to get out at the same price if wrong after loading the full size (basically just scalping because one still has a tight room for error since it is a averaging up operation). I believe the above mentioned method is a true logical operation to stretch "risk-reward" further in a smarter way.