.png)
Stat Checking in DataVolley
Why Stat Accuracy Checking is Important
We use performance analysis so we can make informed decisions to enhance performance. The informed decisions are based on objective and reliable measures of performance, i.e. volleyball stats. So if the stats are inaccurate then there will be inaccuracies in our “informed” decision making. For this reason, I always stat check every volleyball match I scout. I found even after a clean scout, where you didn’t encounter any issues whatsoever, I still found errors after going through the stat checking process.
So below is the stat-checking procedure I go through when I’m using DataVolley. (I have previously written about stat checking but this was when I was using StatEasy).
Stat Checking Procedure on DataVolley
1. Obtain the scoresheet via the scoresheet pink/ blue copy and/or take a photo
I have written previously on the importance of the scoresheet so this is just good performance analysis habit. You can do without it but it does make life a lot harder. This is particularly useful when you know you made a mistake during a set but you don’t have time to correct it. One example is when I notice the score in DataVolley is not the same as the one on the official scorer’s desk or the team(s) are in the wrong rotation because I have the wrong server. It is likely I either awarded the rally point to the wrong team or I accidentally awarded a point mid-rally. There was one time I completely missed a substitution and I didn’t know when they were substituted. In both cases, I need the scoresheet to refer to.
2. Have the match video as one video file
If the camcorder records the whole match as one video file so I can proceed to step 3 straight away. If the camcorder recorded the match in two video files (usually if the match video is longer than 1h40m) then I use a video editing software program (usually Pinnacle Studio 18) to export the two video files as one video file. This step will take an hour or two to complete.
3. Sync the match video to the scout file
I always check the stats against the match video. It makes life easier because once you find an error, you can just check the video to definitely see what happened. To sync the match video to the scout file, I have to do this in DataVideo. The shorten procedure is as follows: New match> Import video> Copy the video in the folder of the match> Import an existing scouting file> Synchronise the scout> save scout. At this point, the synchronised scout is saved in DataVideo but not in DataVolley. This means I have to use Windows Explorer to copy and paste the DataVideo’s synchronised scout file (C:> Dvs_Video> *Match Name*) into DataVolley’s scout folder (C:> Data Project> Data Volley 2007> Seasons> *Season Name*> Scout), overwriting the unsychronised scout file. It is at this point when I think why couldn’t DataVolley and DataVideo just be one program. If I statted the match via match video, the video should already be synced up as long as you “Open a video…” within DataVolley itself.
4. Verify Codes (“Ctrl+V”) in DataVolley
This is the actual first level of stat checking. This function will find some of the obvious errors.
This above is an example of an obvious error. There is no serve in the rally. It could be because I was updating the scout and I missed the first part of the rally and I only caught the point ending action. Whatever was missed, you just stat the actions from match video. In the above example, the beginning of the set was not filmed so there is nothing to stat. The function will also catch some other obvious errors.
3. Check each player’s attack combination codes and block attempts by using the “Search” function
Three common errors I find is coding mistyping the player’s shirt number (3 instead of 13), the right shirt number but to the wrong team (13 instead of a13) and mistyping the wrong attack combination codes (a13G61 instead of a13V61). So by looking at a list of an individual attacker’s attack combination codes, via the Search function, you can spot these mistakes.
In the above example, #10 Marta Campi is a middle blocker so I should expect all of her attacks to start with a W, C or G, as they denote middle attack types (see link for full list). I shouldn’t see any V attacks as they denote attack attempts for a high ball set. The only exception is the VC code, which denotes first touch/ overpass attacks. If there is anything that looks unusual then I would just see the associated match video. It looks like in the above example looks clean.
It is the same for wing players – I should only see V codes and not any W, C or G codes. I also look at setter as they should only be VC (overpass attacks) or VV (setter attacks) codes. I do click on random attempts just to make sure they are coded to the right player, although I should really do this for all of them.
Another fairly common error is coding the block attempt to the wrong player. Sometimes the ball goes between the two blockers and you’re not entirely sure who got it. Also, block errors via net contact can be tricky to know who touch the net. So I watch every block attempt and make sure I got the right player.
4. Look at the Play by Play Report
Very occasionally, I stat two unconnected point ending actions in one rally. For example, you may have statted a kill mid rally and then you stat the real point ending rally. These type of errors will show up on the Play by Play Report, where the point bar stays the same (circled green). It should only go left or right.
5. Check who received
You would expect the outside hitters and liberos to have reception attempts. If middle players, opposites or setters have reception attempts charged to them then this should be investigated. In the below example, the setter (#8 Canevari) had a reception attempt. The video showed the setter was out of rotation so the reception error was assigned to her.
6. Use Science Untangled’s DataVolley file validation application (https://apps.untan.gl/dvalidate/)
This is a great application because it finds further checks that DataVolley’s Verfiy Codes doesn’t find.
In the above example, I tend to ignore “Consecutive actions by the same player” because it usually means a player passing then attacking straight after. There is also “Unexpected sequence: attack is followed by play from same team” that I tend to ignore. This usually means an player attacking once, an opposition dig overpass (which is not statted) then the player attacking again. All the other errors I would check. The video time is in seconds so you have to divide by 60 to get the minute.
*** Update 19 January 2021: I actually got a shout out from the guys at Science Untangled after reading this blog post (look in the About section of the link). They added an additional check of whether middle players received ***
7. Other checks?
Polonia uses two liberos so the final thing I have to do is go to the Players List and asteriks the defending libero that they have played in the set. Since I don’t collect defence stats, DataVolley won’t recognise if the defending libero has played. You may also check the substitutions in the scoresheet just to make sure you haven’t missed any substitutions.
Final Thoughts
The actual stat-checking (step 4 onwards) takes about 30 minutes to complete, depending on how clean the scout was. I feel this extra work is worth it. This is so I feel confident that any analyses I make in my match report is based on highly accurate data. It also builds up the trust with the players and coach that the stats you produce are accurate and any analyses based on it are informed. If the players or coach consistently spot errors in the stats, they’ll start to lose that trust in the data. If they don’t trust the data then they wouldn’t trust any analyses based on the data. That is something you want to avoid.
Comments