@Stone141319: Hermes Evolution Diary · Latest Battle Report! Hello everyone, before I know it, I've been studying Hermes for over a month! Many people when they first start playing with AI quantitative trading think the key is to find high win-rate strategies, but what's most likely to blow up your account is actually not knowing how to set take-profit and stop-loss. Today I'm sharing the current take-profit and stop-loss setup of my Hermes semi-automated trading system...

Hermes Evolution Diary · Latest Battle Report! Hello everyone, before I know it, I've been studying Hermes for over a month! Many people when they first start playing with AI quantitative trading think the key is to find high win-rate strategies, but what's most likely to blow up your account is actually not knowing how to set take-profit and stop-loss. Today I'm fully sharing the current take-profit and stop-loss settings of my Hermes semi-automated trading system, as a reference for those studying AI automated trading～ If it helps, feel free to follow me, I'll continue sharing more details later. Thanks for the support! In many cases, when we're in profit, we always want to hold a bit more, and when we're at a loss, we think it might bounce back... The AI strictly executes the strategy, but humans tend to be emotional. Fortunately, we can hardcode the rules in advance—take profit when it's time to take profit, stop loss when it's time to stop loss—letting the AI execute in a "counter-human" way. —————————————————— Currently, the take-profit and stop-loss of my Hermes doesn't rely on just one method, but adds three layers of protection. **Layer 1: Attach native take-profit and stop-loss directly when opening a position.** Whether for spot or futures, Hermes will first try to attach take-profit and stop-loss together when placing an order, so that the position is protected from the moment it's filled. **Layer 2: If the exchange doesn't accept the attached protection parameters, first open the position, then supplement with native protection orders.** This is a common scenario in real trading—the exchange may reject the attached take-profit/stop-loss parameters due to price rules, trigger price rules, precision, or mode restrictions. If we simply abandon the entire trade because of this, the cost could be high. So Hermes will first open the position, then separately place a native OCO/conditional protection order to avoid giving up the whole trade outright. **Layer 3: If the native protection order still doesn't get placed in the end, the software layer takes over the take-profit and stop-loss.** The main loop continues to execute according to the strategy: partial take-profit, fixed stop-loss, trailing stop-loss, timeout exit—never let a position run naked. So the real structure is: Attach on open → Supplement after opening → Software layer fallback —————————————————— **Why design it this way** The reason is simple: top traders never bet risk control on a single point of success. If you only rely on native protection, and the exchange rejects due to parameter rules, the position could be naked for a short time. If you only rely on software take-profit and stop-loss, in case of script interruption, network latency, or API timeout, the protection may not be fast enough. So the most rational approach is: - If native protection can be attached, prefer native attachment; - If it can't be attached, supplement it; - If that still fails, software layer acts as a fallback. Do you think I'm being too idealistic? Of course, I won't deny that there is no absolutely perfect strategy in trading, but the goal of this design is not just to look elegant—it's to ensure that positions are kept under control as much as possible under any circumstances. —————————————————— Of course, how to specifically set take-profit and stop-loss is a matter of opinion. Let me share it: my current Hermes take-profit and stop-loss is designed as two sets of rhythms, one for spot and one for futures, each divided into three tiers. **Spot** fluctuates relatively slower and has no leverage, so it allows more room: - Tier 1: +5% close 20% - Tier 2: +10% close 20% - Tier 3: +18% close 60% - Base stop-loss: 4.5%, automatically widened by ATR if the coin's volatility is larger. **Futures** because of leverage, needs a faster pace: - Tier 1: +3.5% close 25% - Tier 2: +7.5% close 25% - Tier 3: +12% close 50% - Base stop-loss: 2.4%, max widened to 5.5%. The logic is simple: Tier 1 first reduces risk, Tier 2 locks in some profit, and Tier 3 is left for trends that truly break out. Spot gives room, futures controls drawdown. —————————————————— **In a nutshell:** Hermes' take-profit and stop-loss now features three layers of protection + two sets of rhythms, with one core purpose: let small profits be taken first, let big trends keep positions, never let losses drag into big wounds, and never bet risk control on any single link. The self-evolving Agent is still iterating～ For those also playing with Hermes / AI quant / take-profit stop-loss strategies, feel free to discuss your parameter settings, share lessons learned in the comments, or DM me to discuss optimization ideas! Looking forward to sharing real-time battle reports with you again next time. #Hermes #AI trading #TakeProfitStopLoss #AutomatedTrading #OKXAgent #Web3