Catching the falling knife?

I analyzed the MATIC chart and noticed that during each bull cycle, there are strong corrections of 15-20% within a few hours, and then the growth continues as if nothing happened. We all know why these corrections occur: the market liquidates long positions.

I sat down to “chat” with ChatGPT about this phenomenon. It confirmed that I had correctly observed this pattern and listed exactly when these events occurred. After each of these events, the price increased by 50% or more.

I am thinking and planning how to develop this strategy.

Everyone is welcome to share their ideas.

During MATIC’s past uptrends, the cryptocurrency has indeed experienced several significant corrections, typically ranging between 15-20%. These corrections are common in volatile markets and often serve as consolidation phases before the continuation of the trend.

Historical Corrections:

  1. May 2021: In May 2021, MATIC experienced a sharp correction of about 20% after a rapid ascent. This was part of a broader market correction that affected many cryptocurrencies. MATIC’s price dropped from its peak of around $2.68 to approximately $2.14​
  2. September 2021: Another significant correction occurred in September 2021, where MATIC fell approximately 18% from around $1.75 to $1.43. This correction happened amid overall market uncertainty and regulatory news impacting the broader cryptocurrency market​
  3. November 2021: In November, MATIC experienced a correction of about 15% as part of a broader market pullback after a prolonged uptrend. The price fell from around $2.20 to $1.87 before resuming its upward trajectory​
  4. April 2022: There was a notable correction of about 17% in April 2022, where MATIC’s price dropped from $1.65 to around $1.37. This correction coincided with market-wide declines in the crypto sector​

This one happened in 6 bars (6 hrs).

ChatGPT even helped me in creating an indicator for this purpose:

indicator(“Customizable Price Drop Alert”, overlay=true)

// Input settings for the user
dropPercent = input.float(15.0, title=“Drop Percentage”, minval=0.0)
length =, title=“Number of Bars”, minval=1)
useClose = input.bool(true, title=“Use Candle Close Price”, tooltip=“If false, uses the real-time price”)

// Calculate the drop percentage from the highest price in the last ‘length’ candles to the chosen price
highPrice = ta.highest(high, length)
currentPrice = useClose ? close : close

drop = (highPrice - currentPrice) / highPrice * 100

// Define a constant message
alertMessage = “Price dropped by the specified percentage from the highest price in the specified number of bars”

alertcondition(drop >= dropPercent, title=“Price Drop Alert”, message=alertMessage)

plotshape(series=drop >= dropPercent, location=location.belowbar,, style=shape.labeldown, text=“Drop Alert”)

1 Like

Did you test this indicator to see if it works in tradingview?
For me it’s full of errors.

Thanks for these insights. We already have the price change indicator, which would probably work for the use case you mention. Just keep in mind that catching falling knives is risky. Luckily you can easily backtest.



I already tried other LLMs to try to create an indicator, the most often is not working and I am not capable to correct the mistakes (LLMs cannot either) and the main reason I think is because I’m not good enough with coding skills. I suggest you to use LLMs in another way, and this can be by looking for other indicators that already exist and may be good doing this, when I do this exercise with Gemini (connected to Internet) it proposes me this:

  • Relative Strength Index (RSI): When the RSI gets above 70, it suggests the market might be overbought and due for a correction.
  • Moving Average Convergence Divergence (MACD): A bearish crossover (when the MACD line falls below the signal line) can indicate weakening momentum and a potential pullback.
  • Stochastic Oscillator: Similar to RSI, readings above 80 suggest overbought conditions.
  • Volume: A sudden drop in volume during a price decline might indicate the correction lacks conviction and could be short-lived. Conversely, high volume on a price decline suggests stronger selling pressure.

Can be a good start, but I was also thinking that QFL could be doing this nicely Qfl — Indicators and Signals — TradingView. Luckily you can try all of these indicators on Gainium to see if they can be helpful in your situation.

Also be careful because MATIC can behave differently this time as it may not be as trending as in the last bull run and also be careful because if the trend does not continue you may cut yourself with the knife! If you can backtest things by having all of this in your mind, I will be happy to know what you found! Good baktesting :wink:

1 Like

You can also avoid spikes and falling knifes by adding a timeout so if anything happened you don’t get liquidated.
QFL is a great strategy too and it works on timeframes of 1-4h or 1D

1 Like

Actualy, QFL was the thing I’ve been searcing for. Thank You guys!
Let’s see what can we do. :slight_smile:

1 Like

I don’t understand this “timeout” thing.

Start saying that I thought this feature was also available for SL but it’s not actually :sweat_smile: maybe a feature suggestion for @aressanch for the future.

When there is a spike or a falling knife you can set the bot to trigger the SL only after x seconds so if there is a quick spike or a temporary downtrend you can set it so the SL is triggered after a certain time - this way you can potentially avoid liquidation and losses

This simple indicator can be even better

But I’m not sure how fast would this order be executer via webhook.

Webhook alerts are almost the same as built in Gainium conditions in terms or speed - it may have some little delay because a third party application is added to the game but I’m talking of a few milliseconds

Could be useful indeed! It will not work every time, as the price sometimes hits the line and fall even more as it is showed in this image, but looks quite interesting.

You can consider using the risk:reward window to define a proper SL and TP for this :wink:

yesterday during the dump a webhook started the deal after 3 minutes, lost almost all volatility.

This is because of the way exchanges process requests - it’s easy to check, look at the alerts history and the bot log (at the bottom of your bot) look at the time there should be a small delay - during high volatility exchanges are overloaded that’s because there is a delay

1 Like

alert fired in tradingview- 20:00:43
alert received at Gainium- 20:02:51
deal started at Binance-20:03:09

1 Like

Yep as mentioned having a third party app connected via API may create delays - as you can see the alert actually ARRIVED late and then Binance executed later because of the congestion

1 Like