I’m with you there and I never average losers but look at positional traders who might hold for a year or longer at a time...dollar cost averaging can work on a fundamentally strong play...are you generating crazy alpha? Doubtful... but then again how many of us actually are?
Personally, there are very limited circumstances where I’ll add to a loser. For example, my position’s stop is close to being tested, when I consider my stop placement to be a significant “Line in the sand” that I expect to hold technically (As is usually the case), I have not reached my maximum allowable capital commitment on a single trade, and if my stop were hit, the loss in both positions would not exceed my daily allowable loss in a single position. Further, I would never add on to a setup I originally considered to be “Momentum” based. On the additional and original position, if I had not done so earlier, I will set a timer on expected price progression that can be effective immediately or effective at a specific time of the trading session, like an trailing stop that has both a price and time component. Whatever reason I use to add on to a losing position, if it has not moved significantly away from my stop area by the end of the day, I’ll either reduce my position size or exit the entire position by the close.
Everything, nothing, or some. Depends on your system. If you don't know, your system is not complete, let alone tested. Scaling, and size management is an integral part of a systems risk management, which in turns is part of tactical execution, entry and exit. It is not an afterthought. It is not an "out of plan" rescue method. Either embrace it with a well though out and tested plan, or don't. Going back and forth to avoid loss management, is like changing time frames after a trade goes wrong, or making a trade into an investment once it goes sour. It is another tool. Use it wisely and prudently and it can be good. Don't respect it and it can close accounts.
Sometimes you know that the price has gone to find the point where all the retailers have the stop. Usually at this point, the price turns around to go in the desired direction. It is the only place where I see logical adding a new position. I don't see the logic of adding more than once because if the day turns into a big loser then consistency cannot be achieved.
Up until the maximum amount of the capital that you want to risk if you truly believe that the price will eventually go in the direction that you originally intended. Otherwise it's not worth it to add to the loser because you just don't know how much more you are going to lose. Better to just cut the loss and start a new position when there are clear signs of reversal displayed.
It is simple: when you stop believing it can 'turn' your way, you stop adding to your loser and get out. When you know that point, you can calculate how many scale ins you do and how much your total max risk can be. If you do not 'know' the uncle point, you need to do research about that first, before you consider adding to 'loser'. I do not do it. I like to be green from the first get go.
Maybe your answer is to not trade during the opening, then? Generally, it's not unusual that you'll see some directionless volatility in the initial 15/30 minutes of the market. There's a difference between averaging down on a loss and scaling into a trade where you have a plan in advance about how you'll reach your full position, i.e., adding both as price dips below your initial entry and as price continues in your anticipated direction.
If you need to add to a loser that means your entry is not optimal and you need to work on that. If your initial stop is hit get out, end of story. Plus you can always re-enter later, if the market comes back to your buy/sell entry level (assuming the trading signal is still valid). On the other hand do add to winners, it is still the best way to turn a relatively modest capital into a small fortune, just one mega move in your favor and you are set for life, especially if you are trading Futures or Forex...