Make a Contribution to Remove NSFW Ads

Stock Market heads, peep ya boy

ScorpDiesel

ScorpDiesel

Rising Star
Registered
With AI, I created a new number system (My ideas, AI just formulated the math). This thread is for the historical record (on a porn board LMBAO!). I'll be submitting a scientific paper for publishing. Lotus is my algorithm.

MetricLotusSPY
CAGR20.38%10.49%
Sharpe Ratio0.7130.620
Max Drawdown-54.97%-55.19%
Final Equity$50.42$8.24

Target: 66% CAGR (Renaissance Medallion level) - yes, I know the caveats and hurdles of this goal, but fuck it.
Current CAGR: 22.35%

I don't know shit about stocks, so I'm relying on the AI to guide me. I'll periodically update this thread on progress.

Note: these numbers are built on historical stock market data to test the algorithm. Once optimized, I'll test it on paper trading to see how it does for a few months, then I'll use real money.
 
Last edited:
Update: I discovered the fucking AI stopped including transaction fees, so the numbers I posted are fake news, but we were able to claw back some alpha after the correction. So with transaction fees included, these are the corrected numbers as of now:

- CAGR: 3.01% → 9.14% (3x improvement)
- Sharpe: 0.436 → 0.680 (beats SPY's 0.620)
- Max DD: -25.95% → -35.77% (still 19pp better than SPY)

In parity with SPY, but we're still pushing for more.
 
ScorpDiesel said:
Update: I discovered the fucking AI stopped including transaction fees, so the numbers I posted are fake news, but we were able to claw back some alpha after the correction. So with transaction fees included, these are the corrected numbers as of now:

- CAGR: 3.01% → 9.14% (3x improvement)
- Sharpe: 0.436 → 0.680 (beats SPY's 0.620)
- Max DD: -25.95% → -35.77% (still 19pp better than SPY)

In parity with SPY, but we're still pushing for more.
Click to expand...
Why are you paying transaction fees on trades?
 
xfactor said:
Why are you paying transaction fees on trades?
Click to expand...
From what I understand, which ain't much, is that brokerages charge transaction fees on trades, so I have to account for that in the algorithm.

from gemini:

When testing a trading algorithm on historical data (backtesting), the #1 mistake people make is ignoring transaction fees and 'slippage.'
Slippage is the hidden cost of trading. Even if your broker offers 'zero commission' trades, you still lose a tiny fraction of a percent every time you buy or sell because of the bid-ask spread (the difference between what buyers are willing to pay and what sellers are asking for).
If an algorithm is highly active—meaning it buys and sells constantly throughout the day or week—those tiny fractions of a percent start to add up massively.
For example, an algorithm might show a theoretical 25% annual return on paper. But if it had to make 1,000 trades to get that return, the accumulated cost of crossing the bid-ask spread 1,000 times will often completely wipe out the profit, turning a 25% gain into a net loss in the real world.
By forcing the model to account for a strict transaction fee in the backtest (like 0.05% per trade), you prove that the algorithm has actually found a strong, reliable market trend, rather than just 'twitching' and over-trading on random daily market noise.
Click to expand...
 
Last edited:
ScorpDiesel said:
From what I understand, which ain't much, is that brokerages charge transaction fees on trades, so I have to account for that in the algorithm.

from gemini:
Click to expand...
Slippage occurs when you are putting in market orders so the best way to minimize it is putting in limit orders.

Slipping mainly occurs with crypto trading but can if there is low liquidity in the order book.
 
xfactor said:
Slippage occurs when you are putting in market orders so the best way to minimize it is putting in limit orders.

Slipping mainly occurs with crypto trading but can if there is low liquidity in the order book.
Click to expand...
Good looking out. I'll share this with the AI and see what it says.
 
Update:
Breakthru. I tried some different strategies and what worked was more diversification on individual stocks. I created 2 teams, Old Guard which is the previous algorithm (numbers shown earlier), then the New Guard investing in individual stocks:

Code: 
| Metric       | Old Guard | New Guard |     SPY |
| ------------ | --------: | --------: | ------: |
| CAGR         |    10.51% |    17.62% |  10.53% |
| Sharpe       |     0.695 |     0.998 |   0.622 |
| Max Drawdown |   -36.40% |   -34.43% | -55.19% |
| Final Equity |     $8.27 |    $30.91 |   $8.30 |


New Guard delivered the strongest overall performance, with the highest compound annual growth rate, the best Sharpe ratio, and the highest final equity, while also maintaining a smaller maximum drawdown than SPY.

I'm coming for the alpha!
 
Update. Trying some different strategies:

| # | CAGR | Sharpe | Max DD |
| 7 | 17.76% | 0.973 | -34.77% |
| 1 | 17.43% | 0.989 | -34.84% |
| 4 | 17.05% | 0.948 | -34.94% |
| 2 | 16.99% | 0.953 | -34.77% |
| 3 | 16.73% | 0.978 | -33.82% |
| 5 | 16.65% | 1.009 | -32.27% |
| 6 | 16.57% | 0.970 | -33.99% |
| 8 | 16.47% | 0.974 | -33.54% |
 
xfactor said:
Slippage occurs when you are putting in market orders so the best way to minimize it is putting in limit orders.

Slipping mainly occurs with crypto trading but can if there is low liquidity in the order book.
Click to expand...
AI's response:
Regarding the transaction cost question — good distinction. Our engine models a flat turnover * 0.0005 (5 bps) cost per rebalance. That's a reasonable proxy for:

Slippage — yes, mainly a market-order problem. Limit orders avoid it but risk non-fills on fast moves. - Spread — the bid/ask spread on large-cap equities is typically 1-3 bps. Our 5 bps is conservative. - Commission — effectively zero on most brokers now.

For the assets we trade (SPY, large-cap stocks, GLD, TLT), liquidity is deep — slippage is negligible with limit orders. The 0.0005 is more of a conservative spread+friction estimate than actual slippage. Crypto is a different animal entirely with thin order books.

The model's daily-close rebalancing is naturally limit-order friendly — you know your target allocation EOD and can place limits for next open or MOC orders.
 
Update:
| Strategy | CAGR | Sharpe | Max DD |
| Baseline | 17.43% | 0.989 | -34.84% |
| Gen 1 Best | 18.09% | 0.943 | -34.71% |
 
Update:

| Strategy | CAGR | Sharpe | Max DD |
| Baseline | 17.43% | 0.989 | -34.84% |
| Gen 1 | 18.09% | 0.943 | -34.71% |
| Gen 2 | 18.94% | 0.919 | -34.31% |
| Gen 3 | 18.96% | 0.920 | -34.24% |

Only optimized for CAGR, which is why the other values aren't improving. That was a dumb move. I'll fix it next time.
 
update:
| Strategy | CAGR | Sharpe | Max DD |
| Baseline | 17.43% | 0.989 | -34.84% |
| Best (Gen 7) | 20.11% | 0.941 | -35.59% |
| Delta | +2.68pp | -0.048 | -0.75pp |

Goal: That juicy Medallion Fund 66% CAGR. Long way to go, but I'm on the right path and I have several more strategies to try.
 
Update:
| Strategy | CAGR | Sharpe | Max DD |
| Baseline | 17.43% | 0.989 | -34.84% |
| Best | 20.27% | 0.945 | -35.66% |
| Delta | +2.84pp | -0.044 | -0.82pp |
 
Update:
| Strategy | Train (2005-2018) | Validation (2019-2026) |
| Baseline | CAGR 16.22%, Sharpe 0.996, DD -20.81% | CAGR 18.34%, Sharpe 1.116, DD -26.63% |
| group1 Winner | CAGR 17.07%, Sharpe 0.843, DD -33.04% | CAGR 24.49%, Sharpe 1.136, DD -34.82% |

Niggas, my algorithm does better on CURRENT market data than historical data. My concern was that the algorithm might be overfitting to the historical data, but it's doing the opposite. This is good fucking news!
 
ScorpDiesel said:
Update:
| Strategy | Train (2005-2018) | Validation (2019-2026) |
| Baseline | CAGR 16.22%, Sharpe 0.996, DD -20.81% | CAGR 18.34%, Sharpe 1.116, DD -26.63% |
| group1 Winner | CAGR 17.07%, Sharpe 0.843, DD -33.04% | CAGR 24.49%, Sharpe 1.136, DD -34.82% |

Niggas, my algorithm does better on CURRENT market data than historical data. My concern was that the algorithm might be overfitting to the historical data, but it's doing the opposite. This is good fucking news!
Click to expand...
good drop, everything is upside down but if u know which way to look u should be fine, just trying to stay clear of falling knives over here
 
ScorpDiesel said:
Wait, what do you mean "everything is upside down..."? You're referring to the markets or something else?
Click to expand...
yeah,im pretty much just sitting still for another 10days before I start buying again, im just watching the market for stocks ive had my eyes on for a minute to see how much lower & how much longer they go b4 I pounce, the cycle never fails
 
Update:
| Strategy | Train (2005-2018) | Validation (2019-2026) |
| Baseline | 16.22% CAGR, 0.996 Sharpe | 18.34% CAGR, 1.116 Sharpe |
| v1 Winner | 17.07%, 0.843 | 24.49%, 1.136 |
| v2 Best | 17.90%, 0.851 | 24.52%, 1.086 |
 
You must log in or register to reply here.

Similar threads

PLASMA CANNON
Top 25 Censored News Of 2007, The Shit You Didnt Know
Replies
28
Views
3K
Dr. Truth
Dr. Truth
Back
Top