Strategy Inspired by Perry Kaufman in 2 Minutes

A strategy with no indicators, just one entry rule, and 98% win rate. Sounds like a holy grail? It isn't. But, the reason why is exactly what makes this test so interesting. This strategy doesn't use a single indicator. No RSI, no moving averages, no settings to tweak. All you have to do is look at the last three candles. And one trader whose video I came across claims the legendary Perry Kaufman has been trading it for over 50 years. Sounds almost too simple, doesn't it? So, let's build it in Algo Cloud and test it. Not on Forex, the way he did it, but in US stocks. And we will take an honest look at whether the edge is really there, or whether it's just a nice story. The rule is simple enough to fit into one sentence. If a stock closes lower three days in a row, I buy it. Three declining closes in a row, and I'm [music] betting the price bounces back up. No complicated candlestick patterns with wicks and colors like the three white soldiers you read about in the books. I only care about one number, the closing price. Has it fallen three days in a row? I buy. That's all there is to it. It's plain mean reversion. After a short sell-off, I'm expecting [music] a bounce. You could mirror it the other way around. Three rising days as a signal to go short. But, shorting stocks is expensive and impractical. So, we'll skip that today and stick to long trades only. And let's be straight about it. The trader tested this on a 14 currency pairs, and it only made real money on two of them. We are going to run it on stocks. And let's see if it fits better there. Let's build it. As always, we'll open a new strategy in Algo Cloud's stock picker engine. First, we set the triggers, so we will buying and selling on bar close. Next, we set the universe to the S&P 500. We run the test from the year 2000. We set slippage to zero and commission to a tenth of a percent. For money management, we will use a fixed amount of $10,000 per position, so the equity curves are easy to compare. And that's it. Now for the famous entry. I simply say today's close is lower than [music] yesterday's one. And yesterday's is lower than the day before. So we set today's close to be lower than yesterday's close. And we duplicate that. Here's the first candle, and the second candle is here. We just change it so yesterday's close is lower than the close from 2 days ago. That gives us two. We duplicate it again and set the close from 2 days ago to be lower than the close from the day before that. That gives us three. First day, a declining [music] candle. Second day, declining candle. Third day, a declining candle. And there we have the basic Kaufman exit. Two days and out. And when several stocks meet a condition at the same time, I need to know which one to take first. [music] That's why I set up a position score. We trade 10 stocks long and zero short. We set the score as the percentage decline over the last 3 days. The further a stock has fallen, the higher its score, and the sooner I buy it. So, I'm taking the most beaten down names. We express that percentage score with the ROC indicator, rate of change, but we need the lowest value, so we have to start with the opposite sign. So, I add the minus function here and apply it to the percentage ROC over the last 3 days. The score is simply minus ROC. For the first exit, I use Paris original. Holds 2 days and get out. We'll be changing it shortly, but for now it gives us something to push off from. Let's run the test. And straight away, we can see what we are working with. It trades like crazy. Over 20,000 trades, a win rate right off the 50% line, and a drawdown of almost 40%. It [music] makes 5% a year. But, it's a roller coaster, and with that many trades, commissions really, really bite. Now, let's hold the position for 3 days. Change the exit after from two to three, and run the back test. The return barely moved, and the win rate is still around 50%. But the drawdown should up to 62%. And let's test 4 days as well. This is the highest return of all the time based exits. 6 and 1/2% year and the best return to risk ratio. But don't let it fool you. The drawdown is still almost 60%. Holding longer lifted the return, but the brutal drawdown stayed. And now, the exit that won in the video. It's called bailout. The rule goes, hold as long as it takes to get into profit for the first time and the moment you are there, jump out. Not a day longer. And this is where we run into a limit in Algo Cloud. Let me be straight with you. You literally cannot write close on the first profitable close because the exit condition can't find out what you paid for the position. But, there's an elegant way around it. I set a small profit target of 1%. Still small enough to be in practice the first profitable close and comfortably above that what [music] commissions eat up. So, I'm saying the same thing. The moment you are a little bit in green, take it and go. So, I set the profit target to 1%. Now, the profit target is the only exit. Holds until you're in profit. Let's [music] run the backtest. In the Forex test, this logic pushed the win rate above 80% on two pairs. And on stocks, 98%. Almost every single trade closes in the green. The profit factor is almost twice as high. The drawdown dropped from 60% to 33 and the curve is by far the smoothest of the lot. There's only one catch. The overall return is lower. 4% a year instead of six on the four-day hold because we could aggregate short at 1%. We've traded return for peace of mind and near total certainty. But watch out for one thing, that smooth equity curve is hiding. The bailout never books a loss. Losing positions just stay open until they come back into profit. So, that 33% drawdown is almost entirely floating. It's hanging in open trades that are waiting for a recovery. That high win rate doesn't mean the risk has disappeared. It's just moved off the equity curve and into the open positions. And if the price doesn't come back, you can sit on a loss for months with your slots blocked. When you look at the numbers side by side, you see the same thing you see with every mean reversion strategy. All four exits are in profit. It's just that the win rate swings from 50 to almost 100% and the drawdown from 33 to 62. The exit dramatically changes the character of the strategy. The win rate, the size of the wins and losses, and the drawdown. But the real edge comes from the entry, from buying weakness. So, what do you think? A 50-year-old strategy, one simple rule, no indicators, and we built it in 2 minutes. On four exit, only worked on a couple of pairs, but on stocks, the logic really holds. All four exits finished in profit and with the bailout, the win rate jumped to almost 100%. Not bad for such a trivial rule. Just don't go thinking it's the Holy Grail. It's not. The drawdown can exceed 50%. So, this is more of a demonstration of the method than a finished system. So, please keep that on your mind. But most of all, you now know exactly what it does and why. And that's the whole point. Do not copy. Do understand. Thank you for watching, guys. And if you enjoyed this video and you want to watch some more, don't forget to hit like and follow, and see you next time. >> [music]