Investigation Tips

I am going to assume that you are reasonably proficient with Excel, or other spreadsheet program.

From the original spreadsheet, look at the columns on the far right, labeled, "Tray URL, "Opening lead URL". These are hyperlinks to the time that the cards come out the tray, and to the opening lead. There are different spreadsheets with different data, but these columns should be AB, AC or something similar. Some spreadsheets have an additional column which is a link to the BBO file, "BBO URL". This assumes that you are using Excel, I do not use OpenOffice or other spreadsheet tools, but should be similar. If there is no data for "BBO URL", enter the following in second cell in that column and copy to the rest of the column:

If the first column is the BBO (LIN #), replace B2 with A2 in the line above.


The general rule is to look at a number of hands, see if you spot a pattern, determine a hypothesis, and then use the remainder of the hands to validate the hypothesis. Never look at all the videos before creating a hypothesis.

A hypothesis should be reasonable, and have bridge logic associated with it.

To help with the research, use additional columns of the spreadsheet to record what you see or follow the tips below.

If there are too few hands, drop the hypothesis. For example, every time he makes a loud yawn he has a singleton in diamonds. If there is only one yawn, you are not going to convince anyone that it anything but co-incidental. But, if the hypothesis is he scratches his right ear with a singleton spade, right side of nose with singleton heart, left side of nose with a singleton diamond and left ear with a singleton in clubs even though these are four separate individual incidents there is bridge logic associated with the hypothesis, and if the only time these scratches occur is with these hands, then the case is a lot stronger.

There are some columns which are useful to sort on before you look at the data. For example, if I am looking at opening leads I will sort by Column ("We bid"). I believe that I am more likely to signal length/attitude in suits when we have not bid. If I am looking at the location of the opening lead, I will sort by the opening leader, so that I see the same person lead.

Additional Links

If you want to record what you see so that others can then see the same, I suggestion creating links. Look at the provided spreadsheet and see columns X, Y, Z. It should be "Opening lead (HH)" for column X. If this is not the case, adjust the columns in the formula below. These are the time in hours (HH), minutes (MM) and seconds (SS) that the opening lead was made. See column AC, this is a link to the YouTube video starting at the time of the opening lead. For example, cell ABC is


Cell C2 contains the Youtube URL, X2/Y2/Z2 are the HH/MM/SS.

You can do similar with additional columns if you want to record a time - it helps to be able to look back at a specific time in a video.


Screen shots of the video can help with an investigation. You can look at multiple pictures, next to each other, of a particular event. For example, where is the right hand of the partner of the opening leader when the screen is opened. Compare 10 pictures, is the hand always in the same place, is it sometimes vertical/horizontal/angled/fingers showing etc. Trying to find a hypothesis and record all possibilities can be time consuming. Look at 10 pictures rather than videos can be much quicker.

Google "screen shots" and your operating system, e.g. "Windows" or "Mac" and you can find instructions on how to take a screen shot.

Opening Leads

A Dummy lead is 45 degrees from horizontal/vertical pointing in the direction of dummy. A Declarer lead is 45 degrees from horizontal/vertical pointing in the direction of declarer. Note that in both cases it is not pointing at dummy, or at declarer, it is just the easiest way of describing a diagonal lead. Nearly all players make one of the five leads: horizontal, vertical, dummy, declarer, and a lead which is in between horizontal/dummy. For classification we use horizontal/vertical/dummy/declarer.

Check the orientation of the opening leads - use column AC. The data in column AA were crowdsourced, they may be wrong. Make corrections if needed. If you have a spreadsheet with an empty AA it means that the crowd sourced data was not provided.

Compare the orientations to the "Lead suit" (usually column T). To make it easier to work, hide columns U-Z (the HH/MM/SS columns). See if there is a pattern. Easiest way is to sort by Orientation column, then look at the suit led (Column T).

Look for correlation with honors. Look for correlations with length of suit.

Look at the depth of the opening lead, left/middle/right, near/middle/far from the table. If looking at placement of cards, taking screen shots can help so you can easily compare.

If trying to find out signals after the opening lead, sort by column L (We Bid). You are far more likely to signal if you do not bid than if you do.

Suggestion is to first sort by Column M (we bid). Then, for all for the auctions where we did not bid, sub-sort by the person on opening lead (column Q). You can now look at all the leads by one person one after another.


For bids looking at the placement on the bidding tray - gaps/orientations. To help, you can create an Excel column for each time a bid is placed. I recommend a column for Hours (HH), Minutes (MM), Seconds (SS). See how the formulae for AB and AC take the HH/MM/SS and create a hyperlink, do the same if looking at bids. The gaps are both the gaps between each bid (i.e horizontal as viewed by a player), and also vertical (as viewed by a player). Some players make angled bids - I realize that this may seen strange with a bidding tray but it does happen.

Record what you see, decide if there is a pattern, try a hypothesis and find a correlation.

Look at the first bid - is it always flush against the side of the bidding tray? If not, record the different placements.


All players have quirks. Just because you see one does not mean this is a signal. Create a hypothesis and test it before claiming it is a signal.

False Positives

I strongly believe that if you cheat, you do not necessarily cheat on every board.

This may vary based on the cheating method. For example, if the cheating method is based on bidding, then I am likely to cheat on more boards - I have less information to go from. If the signal is based on opening lead, I am also likely to cheat more. However if the signal is during the play of the hand, either I may not want to give the signal, or not have enough time to give a signl.

Example: supposing I am the partner of the person on opening lead and I have the ability to signal my length in that suit. I have a signal for 4 cards, and a signal for 5 cards. You lead. I may not have enough time to make the signal, or choose not to make the signal. From the opening leader's perspective:

1. If you signal 4 cards, you have 4 cards.

2. If you signal 5 cards, you have 5 cards.

3. If you do not signal 4 or 5 cards, you may have 4 or 5 cards.

Number 3 is very important. If you signal all the time, the correct inference is "If you do not signal 4 or 5 cards, you do not have 4 or 5 cards". If you do not signal all the time, the inference is very different.

If you signal all the time, then a false positive would be failing to signal a 4 or 5 card suit.

If you do not signal all the time, then failing to signal a 4 or 5 card suit is not a false positive. In this case a false positive would be a signal for 4 cards when you do not have 4 cards.

Statistical Verification

If you have a valid hypothesis, then it must pass a reasonableness test. This is not yet defined. Finding a hypothesis which turns out to be valid but could occur randomly 1 in 100 times is not sufficient. 1 in 10^50 is presumably sufficient. This is a complex area. Recommendation is that you get others to check the numbers.

We do have a group that can crowd source anonymously and privately a reasonably valid hypothesis.

Copyright © 2015 Hammond Software.