How do you analyze player stats from tournaments?

[MrDov]

As title says, I'm wondering how everyone analyzes the data they get from their various tournaments? Is there something more efficient than going through each tournament individually to see your player stats?

[DotDash]

Originally Posted by MrDov

As title says, I'm wondering how everyone analyzes the data they get from their various tournaments? Is there something more efficient than going through each tournament individually to see your player stats?

It's been a while since I had the time to do this, but what I used to do (and there may be better / more efficient methods), is go into the Player Search screen for a number of completed tournaments, and filter it so that only players with a relevant number of at bats (or IP) are visible. Then use Report -> Write Report to Disk. Then once you have generated such reports for a number of tournaments, combine them in Excel, sort them by player name, add up the combined stats across tournaments and teams per player, and then recalculate the statistics you think are most important in evaluating a player's performance. I liked to calculate batters' Runs Created per Plate Appearance, and then multiply it by 580 to get an idea of a player's average full-season performance. I don't recall what I used to evaluate pitchers, but the method was the same.

The key here is to not just look at your players' performance, but at the performance of all players in multiple tournaments so that you have enough data to draw conclusions on.

A shortcut to building an average winning tournament team is to look at a number of teams that ended 1st or 2nd, see which players are most common on those teams, and assess whether it is feasible for you to obtain such cards in order to compete at whatever level you are looking at. (Gold tournaments and upwards can get very expensive if you want to field a competitive team.)

Hope this helps. I'm sure some players have streamlined / automated this process, but I'm not sure if they'd want to share their secrets to becoming competitive

[TwanLX2000]

My method is not incredibly sophisticated, but I track the rosters of each team that places (1st-4th in a 32 team tournament) and use this data to construct a team based on position flexibility and handedness. I usually want at least 100 data points (approximately 25 tournaments), and I repeat this process about a half a week after each Live update / low historical drop. It’s not super in depth from a statistical standpoint, but I don’t want to just know how players perform; in Bo7 where entire rosters need to contribute, I want to add in the element of what their roles will be in a series of fatiguing games.

This usually nets me 15-25 packs a day (sometimes upward to 30), and it will only improve once GO is available. I can’t imagine that I’ll perform any stronger if I put more work into it.

[daves]

There were some TWITCH shows analyzing tournaments. They were great, but I am not sure there is one at this time?

[MathBandit]

Originally Posted by DotDash

It's been a while since I had the time to do this, but what I used to do (and there may be better / more efficient methods), is go into the Player Search screen for a number of completed tournaments, and filter it so that only players with a relevant number of at bats (or IP) are visible. Then use Report -> Write Report to Disk. Then once you have generated such reports for a number of tournaments, combine them in Excel, sort them by player name, add up the combined stats across tournaments and teams per player, and then recalculate the statistics you think are most important in evaluating a player's performance. I liked to calculate batters' Runs Created per Plate Appearance, and then multiply it by 580 to get an idea of a player's average full-season performance. I don't recall what I used to evaluate pitchers, but the method was the same.

The key here is to not just look at your players' performance, but at the performance of all players in multiple tournaments so that you have enough data to draw conclusions on.

A shortcut to building an average winning tournament team is to look at a number of teams that ended 1st or 2nd, see which players are most common on those teams, and assess whether it is feasible for you to obtain such cards in order to compete at whatever level you are looking at. (Gold tournaments and upwards can get very expensive if you want to field a competitive team.)

Hope this helps. I'm sure some players have streamlined / automated this process, but I'm not sure if they'd want to share their secrets to becoming competitive

I haven't dove this far into it so could be way off, but from a strictly maths perspective I think you would also have to exclude any players on your team, as well as any players on a team that played against your team, to avoid biasing the sample. Ideally you'd pick tournaments you didn't enter at all but for a format you play (or are interested in playing).

For example, if you happen to have several low-MOV pitchers on your team (relative to the format), players who play against you will have higher HR totals than you would expect them to play against the field at large. You don't play against yourself so you don't want to know which hitters will do well against a field that includes your pitchers; you want to know which hitters do well against the field of pitchers from every team other than yours. Similarly for opposing pitchers, any differences between your lineup and a more representative sampling of lineups would lead to difference in pitcher performance against you specifically as opposed to against the broader field.

tl;dr don't use data from players on your team or from players on teams that played against your team.

[Warfsu]

Originally Posted by TwanLX2000

My method is not incredibly sophisticated, but I track the rosters of each team that places (1st-4th in a 32 team tournament) and use this data to construct a team based on position flexibility and handedness. I usually want at least 100 data points (approximately 25 tournaments), and I repeat this process about a half a week after each Live update / low historical drop. It’s not super in depth from a statistical standpoint, but I don’t want to just know how players perform; in Bo7 where entire rosters need to contribute, I want to add in the element of what their roles will be in a series of fatiguing games.

This usually nets me 15-25 packs a day (sometimes upward to 30), and it will only improve once GO is available. I can’t imagine that I’ll perform any stronger if I put more work into it.

This is basically what I do. When I have more time, I will do more of a deep dive into which guys are better suited for my roster, but the initial research builds your player pool.

[MrDov]

Thanks for the responses! I was hoping for something a little more sophisticated, like an API you can use to access results for tournaments, but looks like the process would be a little more manual.

[professor ape]

Originally Posted by MrDov

Thanks for the responses! I was hoping for something a little more sophisticated, like an API you can use to access results for tournaments, but looks like the process would be a little more manual.

It’s pretty manual, but would you want it to be easy enough for everyone to do it?

[MrDov]

Originally Posted by professor ape

It’s pretty manual, but would you want it to be easy enough for everyone to do it?

I don't see the harm in easier access to data unless the fear is that there would be very obvious "best" players so every roster ends up looking the exact same. Assuming things are a bit more random than that, I think any sort of competitive advantage should be based on interpreting data rather than simply access to it.

[Torgonius]

Originally Posted by MrDov

I don't see the harm in easier access to data unless the fear is that there would be very obvious "best" players so every roster ends up looking the exact same. Assuming things are a bit more random than that, I think any sort of competitive advantage should be based on interpreting data rather than simply access to it.

Everyone has the same exact access to the game data, provided they want to do the work to export it.

Processing it and interpreting it are 100% up to the player.

And while there may not always be a single best player at a position, there are usually only 2 or 3 that outperform the competition, so most rosters will look similar.

[MrDov]

Oh maybe I’ve been looking in the wrong spots of the menus. I can look at results for tournaments I wasn’t apart of? Also are tournaments eventually inaccessible or can I still look at the results for one I took part in a month ago?

[Torgonius]

Last year (I haven't done the testing this year), you could access all the tournaments you entered up to 48 hours later.

If you can get into enough tournaments over a few days, you can get a nice chunk of data and then exclude your team from the results pretty easily.

From there, it's doing that Excel alchemy.

[professor ape]

Originally Posted by Torgonius

Last year (I haven't done the testing this year), you could access all the tournaments you entered up to 48 hours later.

If you can get into enough tournaments over a few days, you can get a nice chunk of data and then exclude your team from the results pretty easily.

From there, it's doing that Excel alchemy.

That’s pretty much the way to go. With the Live update this morning I’ll be starting a fresh sheet. I’ve also been tinkering with a projection model. I revamped my Ace of Bronze pitching staff this morning and will be looking more closely at the rest of my roster later today. We’ll see how it goes.

[revo]

Hey guys-
I go into a completed tourney, then Tournament >> Player Search & Compare >> Player Search and it gives me a list of all the players in the tourney.....but all of their stats are zeroed out.

What am I doing wrong here?

[Torgonius]

If you're looking for stats, go into the tournament - statistics - sortable stats screen.

Set up views to show the stats you want shown and save them as global views.

You have to decide what works best for you, but I do a view for batters and a view for pitchers and filters for PA>=1 (batters) and PI>=1 (pitchers) to filter out the blanks.

[DentalHijinx]

Anyone suggest some good metrics to evaluate pitching? I feel like I have a handle on the hitters, I just can't seem to find a stat or two to run that helps me judge which starters/relievers to use.

[Ozzie7227]

Originally Posted by DentalHijinx

Anyone suggest some good metrics to evaluate pitching? I feel like I have a handle on the hitters, I just can't seem to find a stat or two to run that helps me judge which starters/relievers to use.

I make my decisions based on (1) FIP, (2) BB rate, and (3) K rate. From the Bronze data I've analyzed so far, the pitchers that'll give you the best chance of consistent success (55% or so win rate) are those posting 4.50 FIP and below, putting up better than 4.5 BB/9, and giving you at least 8.0 K/9. With the randomness inherent to Bronze quick tourneys, I feel any additional effort into observed stat analysis gives greatly diminishing returns for the time effort involved.