Because the decision about drink six is not made by the person who decided about drink one. Every plan you make sober is handed over, after the first drink, to somebody with less interest in it. That is why "I'll just have one" fails repeatedly for people who are perfectly capable of keeping other promises, and it is why the plan has to sit in front of drink one rather than in front of drink six.
This is a described feature, not a private failing
Losing control once you have started is one of the recognised features clinicians look for when they assess drinking. It appears in the AUDIT, the standard ten-question tool used across UK services, which asks directly whether you have found yourself unable to stop drinking once you had started. ([GOV.UK, clinical guidelines for alcohol treatment](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/3-identification-and-brief-interventions))
That matters for one reason. If this were a character problem, the answer would be to want it more, and you have already run that experiment. It is a described pattern with a shape, and patterns with shapes can be planned around.
Four things happening between one and six
None of this requires brain scans. It is all observable from the outside, which is why it is worth naming.
**The plan is made by a different person.** Sober you at four in the afternoon has a full view of tomorrow: the meeting, the money, the way last Sunday went. Two drinks in, tomorrow gets quieter and tonight gets louder. Nothing has been overridden by force. The person holding the plan has simply changed their weighting.
**The first drink is the strongest cue there is.** Everything else you have removed from the environment, the bottle in the fridge, the route past the shop, the app with the card saved, does its work by keeping alcohol out of reach. Drink one puts it in reach, in your hand, with the taste and the glass and the setting all present at once. It is the most reliable prompt for another drink that exists.
**The gap between drinks shortens without a decision.** Almost nobody consciously decides to go faster. What happens is that the interval between finishing and pouring quietly contracts, and by the fourth nobody is timing anything.
**The evening reorganises itself around continuing.** You go to the shop before it shuts. You stay rather than leave. You choose the pub for the next thing. Each of those is a small logistical choice that removes an exit, and by nine o'clock the exits have gone.
Read that list back. Not one line of it is about weakness. All four are about conditions, and conditions are the part you can change.
So the plan has to sit before drink one
The practical consequence is unglamorous and quite freeing. Stop making rules about how many, and start making rules about whether, when and where.
Rules that hold, because they are decided in advance by the sober version of you:
- **A drink-free evening is decided in the morning**, not at seven. Write it down or tell someone. The decision made at seven is made by somebody who is already tired and hungry.
- **Change the environment rather than the target.** Nothing in the house, no card saved in the delivery apps, a different route home. [Clearing alcohol out of the house](/guides/alcohol/clearing-alcohol-out-of-the-house) is a fifteen-minute job that does more than a month of resolve.
- **Set the exit before you go out.** The time you are leaving, how you are getting home, and who knows. An exit you have to invent at ten o'clock will not get invented.
- **Eat first, always.** Hunger and tiredness are the two conditions that make everything above happen faster, and they are the easiest to remove.
- **Never test yourself on purpose.** The evening you go out specifically to prove you can stop at two is the evening the whole design is against you.
If you already know which evenings this happens on, [mapping the times you drink](/guides/alcohol/mapping-your-drinking-triggers) turns that into a one-page list of the hours and places to change.
The moderation question, answered honestly
The obvious next question is whether you could learn to have two and stop. This is where a lot of writing gets dishonest in one direction or the other, so here is the state of it.
The UK clinical guidelines ["recommend abstinence as the appropriate goal for most people with moderate or severe alcohol dependence"](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/5-psychosocial-interventions), and in the same passage say that if someone makes an informed choice to aim at low risk drinking instead, they should be supported in that goal. That second half is not a throwaway. It means no service is entitled to refuse you help because you will not say "never".
The research behind it has been argued over bitterly for fifty years. Controlled drinking studies from the 1970s were attacked and defended for decades, and the modern reading is that the evidence "did not unequivocally favour either abstinence-based or reduced-drinking goals" ([Findings](https://findings.org.uk/PHP/dl.php?f=cont_drink.hot)).
The part that matters to you tonight is the second finding in that review. No indicator reliably identifies in advance which individual can drink in a controlled way. Nobody can tell you which group you are in, and neither can you, which is why the honest advice is to judge by your own record rather than by theory. If the last ten attempts at "just two" produced six, that is your data.
One boundary is not disputed. Moderation Management, the peer group built around moderation rather than abstinence, says its own approach ["is best suited for people who are concerned about their drinking and are not experiencing signs of physical withdrawal"](https://moderation.org/). If you get shakes or sickness in the mornings, the moderation conversation is not the one you are in. [How to stop drinking](/guides/alcohol/how-to-stop-drinking) is, and stopping suddenly can be dangerous when you are dependent.
What to say to yourself at the moment of the first one
The internal argument is always the same shape, and it is much easier to answer when you have seen it written down.
"One won't hurt." True in isolation, and irrelevant, because the question was never about one. It is about what one does to the next four.
"I've been good all week." Rewards work fine when the reward is not the thing you are trying to stop.
"Everyone else is drinking." They are, and their sixth drink costs them a headache rather than a fortnight.
"I'll stop at two tonight, I feel different." Feeling different is not evidence. Check what the last five times produced.
None of that has to be argued down in the moment. It only has to be recognised as the same script arriving on cue, which is a much smaller task.
If it went to six last night
Then the useful thing is the next few hours, not the next month. Stop where you are, eat, sleep, and tell one person today rather than next week. [You drank, what now](/guides/alcohol/you-drank-what-now) is the four-step version, and it carries the safety point that matters most: after a break, your tolerance has dropped.
If you want a phone line rather than a page, Drinkline is 0300 123 1110, weekdays 9am to 8pm and weekends 11am to 4pm. Samaritans is 116 123, any hour.
Common questions
**Why can I stop at two sometimes and not others?** Because the conditions differ, not because your resolve does. Look at what was different: whether you had eaten, how tired you were, who you were with, whether there was an agreed leaving time, and whether the drink was in the house. The variable is almost always in that list.
**Is losing control after one drink a sign of dependence?** It is one of the features clinicians ask about, and it appears in the AUDIT questions used across UK services. On its own it is not a diagnosis. Combined with morning symptoms or repeated failed attempts to cut down, it is worth a GP conversation.
**Can I learn to drink normally?** Some people do, and there is no reliable way to identify in advance who can. The UK guidelines recommend abstinence for most people with moderate or severe dependence and say an informed choice to aim at low risk drinking should still be supported. Judge by your own record rather than by hope.
**Does eating first really make a difference?** It removes one of the conditions that reliably makes the evening go faster. It is not a technique for drinking more safely, it is a way of not starting the evening in the state where the pattern runs hardest.
**Why does this happen more when I am on my own?** Because nothing external is marking the time. Other people create natural stopping points: last orders, someone leaving, the meal ending. Alone, there is no boundary that you have not set yourself in advance.
Sources
- [GOV.UK, UK clinical guidelines for alcohol treatment, identification and brief interventions](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/3-identification-and-brief-interventions), for the AUDIT and its question on being unable to stop once started.
- [GOV.UK, UK clinical guidelines for alcohol treatment, psychosocial interventions](https://www.gov.uk/guidance/clinical-guidelines-for-alcohol-treatment/5-psychosocial-interventions), for abstinence as the recommended goal for most people with moderate or severe dependence, and supported informed choice otherwise.
- [Findings, controlled drinking review](https://findings.org.uk/PHP/dl.php?f=cont_drink.hot), for the evidence not unequivocally favouring either goal, and for no indicator identifying in advance who can drink in a controlled way.
- [Moderation Management](https://moderation.org/), for its own statement that it suits people not experiencing signs of physical withdrawal.
