Ping and Pong

[MiningInMySleep]

Hello,

Could we please have an exact definition of Ping and Pong because I clearly don't understand how it works.
This whole time I thought (assuming boomerang safety) that a ping is any filled order (buy or sell) that does not have a matching ping at least pongWidth away in profit. However, this is clearly not the case.

I've tried looking through trading-bot.data.h, but can't find anything relating to how boomerang operates.

Thanks!

[ctubio]

🐨 your description seems the case to me; the related code is all functions named matchPong

[MiningInMySleep]

The reason I'm asking is because I am seeing multiple pongs in K web UI but K log is showing them all as pings. For example on BTC/USD on Bitfinex I have a massive ping at 49k, pongWidth set to .1% and a K log full of pings at 50.7k which are limited by the trades/sec setting when they shouldn't be.

Edit: Images to illustrate the issue here.

These pings should have been closed forever ago, but the majority (90+%) of the sells were pings in my K log (despite being pongs on web UI).

Edit 2: Stats are also very, very strange.

Edit 3: More background info. I have been running with ewmaPrice and stdev protections off. K UI latency is reporting ~30-40ms. Lifetime was set to 3000ms, so K was definitely taking trades, but not classifying them as pongs which I believe lead to them being limited by trades/sec.

Edit 4: It eventually started ponging, but well beyond the set pongWidth currently at .08%. Just to clarify, with boomerang, a pong should be makerFee% + pongWidth? Or are we doing 2*makerFee% + pongWidth? Either way, it seems to be ponging at much larger width when using %.

[ctubio]

yea it was a bug, (Quote object is correctly never destroyed but isPong status was wrongly never removed, so once a quote is a pong, all next quotes become also pongs)

after 4eca855 each niew quote will have the correct status (and limitations) of pings and pongs without matter what was the previous quote

[eniware-org]

I think it is related to the issue I reported in #1047 (comment)
Also running a boomerang safety.

[ctubio]

🐨 many thanks @eniware-org for your help fixing this issue'¡

[Corfucinas]

Any updates on this?

[ctubio]

🐨 many thanks @Corfucinas for your help fixing this issue'¡