A friend sent me his chart layout a while back because he kept missing entries and wanted a second opinion. I counted nine indicators on it. RSI, MACD, two moving averages, Bollinger Bands, a stochastic, volume profile, an Ichimoku cloud, and something custom he found on a forum. Every trade he described followed the same script. Seven inputs agree, two disagree, and he waits for the holdouts to flip. By the time they do, the move is half over, so he passes, and then watches the rest of it play out without him.
The part that took me embarrassingly long to accept in my own trading is that most indicators are computed from the same price series. RSI, stochastics, MACD, rate of change, they are all transformations of the same closes. When four of them agree, you have not gathered four independent pieces of evidence. You have measured one thing four times and let it vote four times. Confidence climbs with every confirmation. Accuracy stays where it was, because no new information entered the system.
More inputs, more confidence, same accuracy
There is an old line of research on expert judgment, the best known version involving horse-race handicappers, where researchers kept handing experts more data points and tracked two things separately, how accurate the predictions were and how confident the experts felt. Accuracy flattened out early, after roughly a handful of inputs. Confidence kept climbing the whole way up. I have never seen a cleaner description of what a heavily decorated chart does to a trader.
The mechanism matters because it explains why paralysis feels responsible while it is happening. Each additional input gives you a small hit of justification. You are doing homework. You are being thorough. But past the first two or three genuinely independent inputs, the marginal indicator mostly restates what you already have, and the disagreements between correlated inputs are mostly noise. So you end up in the worst possible spot, waiting for noise to resolve while your confidence in the eventual decision inflates for no statistical reason.
There is also a mechanical cost that is easy to miss. Waiting for one more confirmation almost always means a later, worse entry. If the setup was a long at support, the confirmation candle you waited for just carried price away from support. Your stop distance grew, your target distance shrank, and the risk-reward you originally liked is gone. Paralysis rarely feels like losing money in the moment. It shows up later in your results, as a string of trades that were right on direction and mediocre on price.
You were never going to have complete information
The other half of this is accepting what a market actually is. Someone in your market knows more than you do, always. The insider filing that explains a move publishes days after the trade happened. The whale wallet you are watching was positioned before you found it. The order flow that matters sits on books and in venues you cannot see. Even well-connected desks act on partial information, and they know it. If your process implicitly waits until the picture is complete, it waits until the information is priced in, which is another way of saying it waits until the opportunity is over.
Once you accept that, the job of pre-trade analysis changes shape. You are checking whether the bet is structured well. Is the risk defined, is the size survivable if you are wrong, is there asymmetry between what you stand to make and what you stand to lose. Certainty was never on the menu, so a process that quietly demands certainty is a process that never terminates.
Three rules that force a decision
What fixed most of this for me was constraining the decision process the same way you would constrain a position.
The first rule is a decision deadline matched to your timeframe. If the setup is on the daily chart, the decision is due by that day's close. If it is a four-hour setup, you get one candle. When the deadline hits there are exactly three legal outputs, enter, pass, or set a two-sided trigger and walk away. Keep researching is not on the list. The quiet benefit is that a pass gets recorded as a finished decision, so your brain stops treating the setup as unfinished business and you stop reopening the chart every twenty minutes.
The second rule is a hard cap of three inputs per trade decision, chosen before the chart is open. Mine map to three questions. One structural input for whether the trade makes sense at all, usually trend or a level that has mattered before. One trigger input for timing, usually price action at that level. One risk input, which is really just where the idea is invalidated and what position size that implies. The specific choices matter less than the fact that they were picked in advance, because choosing indicators after you already have a bias is how people shop for confirmation. Everything else on the chart is decoration.
The third rule is that any setup you do not act on immediately gets a two-sided trigger. A one-sided alert, buy if it breaks the level, leaves the pass case undefined, and undefined pass cases are exactly where paralysis hides. A two-sided trigger specifies both outcomes up front. If price reclaims the level and holds it through the next candle close, I am in, stop below the reclaim. If price gets rejected at the level, or neither side triggers within a fixed number of sessions, the idea is dead and the alert gets deleted. Every setup now terminates, one way or the other. It becomes a position or it becomes an explicit pass with a reason attached.
The compact version, if you want something to keep next to your screen, looks like this.
- Pick three inputs per strategy, in writing, before you look at any chart.
- When a setup appears, set the deadline immediately, one candle on your trading timeframe.
- At the deadline, produce one of three outputs, enter, pass, or a two-sided trigger with an expiry.
- Log passes with a one-line reason, the same way you log trades.
The logging step sounds bureaucratic and it does most of the work. Reviewing a month of passes tells you things a win-rate number cannot. If most of your passes would have been winners, your trigger input is too strict. If your third input never changed a single decision, it is dead weight and you can test a replacement. You are debugging the decision process itself instead of arguing with individual outcomes.
None of this makes the decisions correct, to be clear. Being decisive with a bad strategy just reaches the bad outcome faster, so the three inputs deserve backtesting before they deserve trust. What the rules fix is narrower and still worth having. They convert the slow bleed of missed entries and late fills into a process that ends on time with a decision you can review. Most days, that is about the most you can ask from a process.