My gosh, backtesting is critical and amazing, but obviously it does not always directly correlate to being profitable in live trading. If you have like 70-80% win rate in back testing going for smaller points and tighter stops, it's probably more like 51-60% in actual trading. If you have wider stops and wider target or if your stop loss is based on structure and not an arbitrary set amount of points, than you're backtesting will be a lot closer to live, but obviously still not as good. Yes, I am aware you can have a 20-40% win rate strategy and still be extremely profitable, but for intra-day trading I don't understand how you can have an edge and be much below 50% win rate. But I am sure some people run them, it's just not what I found or would be interested in running, unless it's all I had of course, than I would if it made money, obviously. While not always easy, results can indeed be faked, if you can trade at a high level there is no reason you cannot trade in front of someone live and if that messes you up than you don't believe in what you're doing enough or you're not prepared to execute it at a high enough level to work with someone anyways. It's still amazing to this day all the excuses people use as to why they can't trade real time with someone. It's almost mind blowing.
Just a technical comment here. There are many ways to run back tests. How i arrive at mine, the system is "taught" how to behave on historical data between 2012 and 2021, including 2021. First step validation data is 2022, weeks 3-13, this is used routinely just to be sure the system is working correct. Second step validation data is 2022, weeks 14-24, this is used rarely on major milestones, like finding an extra +0.6% predication rate, basically a new revision of the system. Third and final step validation data is 2022, weeks 25-38, is used against fully packaged and ready for deployment system. YES, the implication here is that you can use this data ONLY few times a year to prevent result poisoning. Haven't bothered to update the database from week 39 yet, it's an involved process in my workflow, i do it every 2-3 months in large batches. If you keep a strict and disciplined data separation, you will never see 20% difference you mention, more like up to 3% due to market conditions and live slippage.
That's fine seems like you're using a much more technical and advance way of back testing than the average trader, so that makes total sense to me. My point is a lot of people back test a strategy with a slight edge and than try it on live. You (GENERALLY) need way more than that. A 70% win rate strategy via normal backtesting (that doesn't use structure as a stop and goes for fewer points) could easily drop in the 50% range. If a worst entry makes you miss just one trade in a day due to failure to reach profit target, or a worst entry makes you hit your stop loss just one time. That's now two trades negative that you counted as positive. So, let's say back testing showed out of 6 trades you had 5 winners 1 loser. Now all the sudden you have 3 winners and 3 losers. Your 83.33% win rate, just got demoted to 50%. See how easy that is? We didn't even factor in human error yet, if these are manual trades.
I am sure people do it, but teaching a system to trade, particularly one that relies on past data is one of the most difficult paths to make consistent profits. The market literal tells you everyday where likely liquidity is and where the highest probability is to place your stop on every trade you take (Every valid trade, not taking some random entry obviously). No automated system at a retail level can come close to a strong manual retail trader. The only edge would be if the trader is emotional and if he is emotional he probably isn't a strong trader. The only other advantage would be speed, which again would be a difficult way to go because than you are fighting against others more well funded and likely more technically advanced users trying to use speed as the edge too. You're not fighting against anyone when you manually day trade if taking proper set ups. In fact if anything you're being helped and supported by larger money if you're trading against people who are trapped. The only negative would be if you're trading such large size that you start to change the dynamics and effecting the individual trading setup you're in.
Correct, that's why i don't mix my prediction capability with trade results. Raw prediction capability is considered only an intermediate metric. There are other systems on top of the one that do the trading with accounting of all the conditions you mentioned.
Correct, i don't claim to beat strong retail in elegance human trader can offer. My "edge" is 24/7 automated swing trading system. It's all a statistical game after all. Yes. I choose automation route just because i enjoy programming. Sparring myself the psych side of trading is a blessing. Forgive me, i keep preaching the benefits of prop trading, but getting a $25.000 account and engaging up to 0.3 lots is a fantastic level of comfort, nowhere near your self funded regular brokerage account.
Agree. I have two funded accounts and my trading partner has 6, more to come soon. So I agree with you. The reason he has more is because he's better at execution and much less likely to make an emotional mistake. I am extremely strong technically, but have been known to lose discipline, which is really ignorant and disrespectful to both myself and him. Now that I see it in that light I haven't been making any more emotional trades. Just viewing it a different way like that as helped me tremendously. At any rate with more accounts in his hands we still both get the benefit of my technical analysis, while removing the risk of me doing something ignorant. It's possible I am finally past that phase though. Time will tell. I'd like at least 3 months of no emotional trades and honoring my stop loss, before I would add any accounts on my end.
HA! Keep at it. From my rough estimate cycling three and more accounts should be a rock solid platform, barring the underlying mechanism, be it either a system or a trader, doesn't get the yips. I've only discovered these special purpose hedge funds, in the beginning i though it was an elaborate scam too. ;] Turns out talking to so called "investors" is a scam and willingly walking yourself into a bondage over promises.
Yeah, it's pretty crazy the negative reputation they get. The older more established ones certainly pay. I've been paid by them multiple times. Part of people's complaints are true. I mean they do make most of their money off of people failing, resetting and than the data and or one time activation fee for the accounts. However, it's pretty clear that is there business model with even a small amount of investigation. They of course run a business, not a charity. If you can trade, you will get paid (outside of something extreme happening like company closes down or whatever, obviously). Also, it's like brokers advertise special deals and lower commissions all the time, how many people do you hear bitching about them being a scam? I've seen very few post like that. It's virtually the same thing though. Both involve trading, both involve high failure rate. In fact if anything the prop firms are the more fair and honest way. If you completely blow the account you only lose a portion of it. Make a mistake in live trading and 100% of your account could be gone, you get no discount or reset.