This guide covers what to measure, a weekly process that works with PokerTracker 4, and how CoolerTracker turns that review into a luck score, a calendar, and hand-level labels.
Connect PT4 on Windows, review one week with labels, and use a Luck Score from 0 to 100 instead of guessing.
Key takeaways
- A results graph shows the outcome, not whether the week was coolers, lost flips, or leaks.
- Log four buckets after every session: results, high-leverage all-ins, cooler vs bad beat labels, and card quality in the spots that mattered.
- Review the tournaments with the largest luck swing; study decisions, not inevitable coolers.
- CoolerTracker reads local PT4 data on Windows 10/11 for PokerStars, Winamax, and iPoker. 7-day trial, then $99/year.
What tournament variance actually is
Variance is the gap between expected value and the result you cashed. In tournaments it is worse than in cash because:
- You put in many buy-ins for a few large scores.
- Most chips go in as a favorite, a dog, or a flip. The sample of those spots is small.
- ICM and bubble play make "correct" folds look like you ran bad if the table does not pay you.
A graph of profit is the outcome. It is not the explanation. Two players can finish −8 buy-ins in a week for opposite reasons: one lost set-over-set three times, the other called light in spots they should fold.
Why a results graph is not enough
PT4 (and any HUD) already answers "what happened to my bankroll." It does not, by itself, answer:
- How many chips went in as a coin flip versus a cooler versus a suckout.
- Whether the week's ROI would still be negative if those spots had run at equity.
- Whether a single tournament's "disaster" was one inevitable cooler or a cluster of bad beats.
Without that split, you review the wrong hands. You study a set-over-set cooler (nothing to fix) and skip a series of 60/40s you keep taking in the wrong ICM spots.
If you want the definitions CoolerTracker uses for those labels, read Coolers vs Bad Beats: Complete Comparative Analysis. Short version: a cooler is two strong made hands colliding (the loss is expected given the hands). A bad beat is the favorite losing against the odds. They are not the same event.
The four things to log after every session
You do not need a 40-column spreadsheet. You need four buckets, every session, in the same place.
1. Results (the noisy number)
Buy-ins, cashes, ITM, and ROI. This is your PT4 baseline. Keep it. Just do not treat it as the review.
2. High-leverage all-ins
Flips and other spots where a large part of your stack went in close to 50/50 (or a clear favorite/dog). Count them. Note direction: with you or against you. A week with eight lost flips and two won is a variance week, not a "I forgot how to play" week.
3. Coolers vs bad beats
Label the big ones. Cooler: both of you already have a strong made hand. Bad beat: you (or they) were a real favorite and the runout changed it. Mixing these two is how tilt starts. You cannot "play around" a cooler the way you can review a call as a 30% dog.
4. Card quality in the spots that matter
Not "I never get aces." The question is whether, in the tournaments you actually played, the hands that decided the result ran above or below expectation. That is what a luck score is for: one number so you are not eyeballing 200 hands at 1 a.m.
Write the four buckets the same night. Memory rewrites coolers as bad beats by morning.
A weekly process that actually gets done
Use this as a 20-minute Friday review, not a second job.
- 1
Export the week from PT4. CoolerTracker reads the local PT4 database. Rooms it supports today: PokerStars, Winamax, and iPoker. Analysis stays on your PC.
- 2
Sort tournaments by swing, not by date. The −4 buy-in satellite and the deep run that lost on a cooler deserve more time than five minis that finished on the bubble with no decisions left.
- 3
Open every tournament with a large luck swing. In CoolerTracker that is the tournament detail view: result, luck score for that event, and hands tagged as normal, cooler, flip, or bad beat.
- 4
Study only the labeled hands that were decisions. A cooler against you is context. A call as a 35% dog on the bubble is homework.
- 5
Check the calendar, not just the graph. A week of green results with a low luck score is the week to trust. A week of red with a high "against you" score is the week to protect your study time (and your tilt). CoolerTracker's calendar shows result and luck per day so that pattern is visible without a spreadsheet.
If you only do steps 3 and 4, you are already tracking variance instead of staring at ROI.
How CoolerTracker measures it
CoolerTracker is a Windows app (10 and 11) that sits on top of PT4. It does not replace the HUD. It classifies the variance inside the tournaments you already imported.
Luck Score (0–100)
One number per tournament (and in session view) from flips, coolers, bad beats, and the quality of cards you received, compared with expected equity. The landing copy also shows "ROI without luck": what the result looks like if you strip the run.
Automatic labels
Hands get tagged as normal, cooler, flip, or bad beat. You review the pile that matters, not every 2-bet fold.
Calendar view
Weekly or monthly. Each day shows result and luck, so you can see a cold streak as a streak, not as a personality trait.
Local data
It reads your PT4 database on the machine. Hand histories are not uploaded for analysis.
Setup is: install from the Microsoft Store, connect PT4, import. There is a 7-day trial, no card required. Annual plan is $99/year if you keep it.
That is the product. It is not a magic ROI printer. It is a way to stop arguing with a graph.
How to read a "bad" week without tilting
Use three questions, in order:
- 1
Did I take a lot of flips, and did they go against me? If yes, the graph is lying about your week. Log it and move on.
- 2
Were the big losses coolers or bad beats? Coolers: nothing to fix in the hand. Bad beats: still usually nothing to fix in the hand, but worth checking you were actually the favorite (sometimes the "suckout" was a thin value shove).
- 3
Is there a cluster of the same decision? Same ICM call, same 3-bet jam range, same blind-vs-blind spot. That cluster is the leak. Variance is the noise around it.
If you cannot answer 1 and 2 from memory, you are not tracking variance yet. You are journaling mood.
FAQ
Do I need PokerTracker 4 to track this?
For CoolerTracker, yes. It connects to the PT4 database. You can approximate the four buckets by hand, but you will miss labels and you will not get a luck score.
Is a luck score the same as "I ran bad"?
No. A score is a summary of labeled spots (flips, coolers, bad beats, card quality) against expected equity. It is not a license to ignore a leak. Read it next to the hands, not instead of them.
Cash games vs tournaments?
The same labels exist, but MTT variance is louder because of payouts and smaller samples of all-ins. This article is for tournament logs.
Which rooms?
PokerStars, Winamax, and iPoker today. Windows only.
Will tracking variance improve my ROI?
It improves what you study. ROI still needs a real sample and better decisions. The point is not to feel better about a downswing. It is to stop studying the wrong downswing.
Track the next week properly
Download CoolerTracker, connect PT4, and review one week with labels instead of a graph. If you only read one more piece, make it the cooler vs bad beat breakdown so the tags mean something when they show up.